Home / BI Encyclopedia

BI Encyclopedia

Whether you’re new to our industry or a seasoned expert, our BI Encyclopedia can be a useful resource.

Learn anything and everything there is to know about business intelligence and analytics

Get insights and knowledge about big data and its principles. Our Business Intelligence Encyclopedia covers the technoloagy and concepts really well. Everything is well defined and examples are used. Synnect Analytics and its BI team has come together to bring clarity and understand with simple concepts.

Encylopedia

A

Actionable Reporting

Making a report or report features actionable can mean different things:

  • Setting up automated business alerts, to inform decision-makers of critical pieces of information so that immediate action can be taken
  • Scheduling and delivering reports automatically, so that users can always be armed with critical information without having to run the report themselves
  • Automated processes and integration with other business applications, so that, for instance, action taken within the BI application (e.g. placing an order) can be written directly into the database

In all cases, the goal of actionable reports is to set up processes to make information jump out to the user and to let the user act on it without leaving the application or having to record the action in another interface.

Benefits of Actionable Reporting

The main benefit of actionable reporting is to make information dynamically reach its purpose without additional action on the part of the user.

  • When the purpose is knowledge of critical items, automated report scheduling lets data and information reach users effortlessly, so that they can focus on decision-making rather than on running reports
  • When the purpose is acting on key information, automated business alerts let decision-makers know when a critical situation crosses a pre-defined threshold, so that action can be taken without need for any further analysis
  • When the purpose is acting on the data, automated processes and integration with other business applications lets users take action from within the report, saving time and effort

Active Directory

Active Directory is a Microsoft-developed directory service that stores information about Windows networks. Stored network components or objects may include organizations, sites, systems, users, and shares. When integrated with business intelligence platforms like Synnect Analytics, Active Directory and other security frameworks may be used to authenticate users via single-sign on.

Ad Hoc Reporting

Ad-hoc reporting is a model of business intelligence (BI) in which reports are built and distributed by nontechnical business intelligence users. In other words, with ad-hoc reporting, all the technical user does is set up the BI solution, connect it to the data-sources, establish security parameters and determine which objects end-users can see. From that point on, the actual reports are created by business end-users.

Ad-hoc is Latin for “as the occasion requires.” This means that with this BI model, users can use their reporting and analysis solution to answer their business questions “as the occasion requires,” without having to request queries from IT. Naturally, ad-hoc reports can be and look as simple as a one page data table or as complex and rich as interactive tabular or cross-tab reports with drill-down and visualization features–or present themselves in the form of dashboards, heat maps, or other more advanced forms.

The Goal of Ad-hoc Reporting

Ad-hoc reporting’s goal is to empower end-users to ask their own questions of company data, without burdening IT with the task of creating a myriad of reports to serve different functions and purposes. Ad-hoc reporting therefore makes the most sense when a large number of end-users need to see, understand, and act on data more or less independently, while still being on the same page as far as which set of numbers they look at.

For example, a company with a large outside-sales force would be the perfect fit for ad-hoc reporting. Each sales rep can set up his own report for his territory, showing performance against sales goals, orders taken, number of visits to each client, etc., in a format that makes the most sense to him. And just as importantly, the numbers used are pulled from the same data sources as the rest of the company, thereby promoting consistency and minimizing surprises at the end of the quarter.

A good-quality, Web-based ad-hoc reporting solution greatly enhances the benefits of the ad-hoc reporting model for the company adopting it.

The Benefits of Web-based Ad-hoc Reporting

  • Get critical information to the right people at the right time – Self-service results plus automatic scheduling/delivery of information let you facilitate timely decision making. Users get the information they need when they need it to answer critical, real-time questions.
  • Flexibility for constantly changing environments – Business needs to evolve. Answers to changing business questions become more critical. It’s impossible to predict what questions and answers users may need in the future.
  • Saves training costs and time – Streamlines users’ access to critical information. Easy-to-use wizards allow users to get up and running quickly, requiring less time to learn the application and providing clear guidance and saving time to build reports.
  • Encourages collaboration and information sharing – Users can easily create, organize, publish and make reports available to other users via the Web for on-demand viewing.
  • Reduces IT workload – The Web-based reporting application itself can be deployed quickly for widespread availability to end-users. Once deployed, it empowers users to build the reports themselves anytime they need the information. No waiting for IT report developers to build them.

What to Look For in a Good Ad-hoc Reporting Solution

A good ad-hoc reporting solution should–like all BI applications–be squarely aimed at the achievement of the company’s strategy. The key here is to identify what each end-user’s strategic function is within the organization, and ensure that the ad-hoc reporting solution is optimized to make that function easier and more effective, while not offsetting benefits by being too costly.To do so, a good reporting solution will offer the following characteristics:

  • Being easy to use. If it is or even appears to be complicated, many end-users will be turned off and user adoption will suffer. For this reason, some of the better ad-hoc reporting solutions available today offer a basic set of intuitive features that are wizard-driven and will look easy even to the proverbial “non-computer person,” while also offering more advanced sets of tools for the user who feels confident.
  • Being robust. Assuming that adoption is not an issue (see previous point), the ad-hoc solution should offer end-users what they need to see, understand and act upon their data. Far from being a more hi-tech version of Excel, it should offer interactive features like ad-hoc dashboards, drill-down and drill-through, advanced sorting and filtering, rich visualization tools like heat maps, charts and graphs, etc.
  • Being Web-based. For it to be truly useful, a BI solution (including ad-hoc reporting) should run on the Internet. Apart from offering the familiar navigability with which we are all familiar, a Web-based solution is available from virtually anywhere and on any device sporting Internet connection. Another benefit of a Web-based ad-hoc solution is that the system administrator won’t have to set it up individually on every user’s machine: installing it on the server is enough, and all the users need to access it is a simple URL.
  • Being easy to set up. Today’s better Web-based ad-hoc reporting solutions are data-source neutral, meaning that they can connect practically out of the box to most of today’s commonly-used data-sources, including databases, Web-services, flat files, etc. This saves the IT department the burden of creating complex metadata structures as the underlying layer, which is time-consuming, cumbersome and expensive.
  • Having a server-based licensing with no per-user fees. If the benefit of ad-hoc reporting is that of empowering end-users, it should not come with a “user-tax” in the form of per-seat licensing.

Advanced Analytics

Advanced analytics is a desirable feature that may be embedded in software applications to provide deeper insights into trends. Applications utilizing this feature provide a unique value proposition by developing advanced (and often proprietary) statistical models and making advanced analytics easily accessible in the users’ analysis.

In addition to delivering a base level of analytics in the application (typically referred to as table stakes), software providers may differentiate themselves by adding more sophisticated self-service functionality and advanced analytics.

Agile BI

Agile BI (business intelligence) is a flexible and scalable architecture that embraces rapid, iterative development and the commoditization of data storage. This agile architecture allows organizations to quickly adapt to changing business requirements while reducing total cost of ownership. As a result, companies are able to rapidly and affordably respond to changing market conditions. In short, agile BI delivers on the IT mantra of “do more with less.”

Agile BI’s flexibility, scalability, and affordability are revolutionizing the way companies use BI in the same way cloud computing transformed the way we look at data storage. An agile architecture is foundational to delivering next-generation BI capabilities like self-service, mobile, and big data. As companies increasingly demand BI solutions that deliver competitive insights in an accessible manner to a broader audience, the adoption of agile BI is a necessity, not an option.

Amazon Redshift

Amazon Redshift is a data warehouse product built by Amazon. Available on Amazon Web Services, Redshift is able to handle analytics workloads on large scale datasets and is searchable using traditional BI applications and SQL methods.

The product manages setting up, operating, and scaling data warehouses. It automatically monitors notes and drives to help recover from failures, provision the infrastructure capacity, and operate administrative tasks, such as back-ups and patching.

Amazon Redshift’s performance can reach at least ten times higher than other databases for data warehousing and analytics workloads. A few of its features include:

  • Columnar data storage: Unlike other databases which use a series of rows, this feature is ideal for data warehousing and analytics because the system requires fewer I/Os, leading to an improved query performance.
  • Advanced compression: With its columnar-based data storage, Amazon Redshift automatically samples data and selects the most appropriate compression scheme when loading data into an empty table.
  • Massively Parallel Processing (MPP): Amazon Redshift automatically distributes data and query across nodes, which allows for faster query performance as the data warehouse grows.

Advantages of Amazon Redshift:

  • The ability to access large databases in a low-cost and timely manner
  • The ability to handle databases larger than a petabyte
  • The ability to quickly return results with a variety of resources
  • The low cost can save a lot of money, compared to the cost of using your own hardware and software

Analytics ROI

Analytics ROI is the Return On Investment on embedded analytics. The components of the ROI formula are:

  • Timeframe – Quantitative analysis is performed over a specified timeframe for a technology investment, typically three to five years.
  • Benefits – The combination of the strategic benefits (e.g., revenue increase) and operational benefits (e.g., cost reduction).
  • Costs – The investment to develop and maintain the solution.
  • “-1” – The formula assures that a positive ROI is achieved only when benefits exceed the costs.

To calculate analytics ROI, use the following formula:

Analytics_ROI_Formula

As an example, let’s say a commercial SaaS provider brings in R2 million in revenue per year. They expect that new embedded analytics functionality can drive a 10 percent increase in sales (to keep this simple, we’ll ignore annual compounding). Over three years, that comes out to R600,000 in added revenue. Because the self-service functionality is expected to free up half the time of a developer (and based on a R100,000 internal cost per year per developer), you also have a R50,000 per year increase in developer efficiency, so the total benefit is R750,000 over three years.

The costs are expected to be R50,000 per year in software plus R25,000 in expert technical services. If a developer dedicates one quarter of their time to this project, your developer costs are R25,000 per year. That makes the total cost $250,000 over three years. The formula looks like this: (R750k / R250k) = 3, so the ROI is 200 percent.

As a second example, consider an internal manufacturing application that helps process R2 million worth of product a year. Embedded analytics helps to streamline the process, reduce waste, and improve yield, all to the tune of 10 percent per year of total production. This results in R600,000 in savings over three years. And just like the first example, with R600,000 in revenue – if we make the same assumptions for additional benefits and for cost – we also end up with 200 percent ROI.

API

API, short for Application Program Interface, is a set of routines, protocols, and tools for building software applications. In business intelligence, APIs may be used to enable end-users to directly update source systems. A user can start from a visualization and update information on the same screen by calling an API or performing a database write-back.

Area Charts

Area charts are data visualizations best used for showing cumulated totals using numbers or percentages over time. They are basically line charts that are filled in to provide a deeper view of multiple series of data.

Like traditional line charts, area charts are set against a common scale and are ideal for showing trends in data over time. You might also add a trend line or goal line to illustrate performance in a certain period against a set benchmark.

Area charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Attribution

In Marketing Analytics, Attribution refers to a set of rules that determine how credit is allocated for an action or conversion. Attribution experts often suggest that too many people rely on the last click as the sole source of a conversion. However, people should consider, measure, and give credit to multiple aspects that influence results. Those aspects could be any number of things, including multiple display ad impressions, integrated e-mail campaigns, organic search, and PPC ads.

Attribution modeling should be used to understand the behavior of website visitors. The goal of it is to determine the most effective and successful marketing channels for investment.

The Model Comparison Tool is used to compare how different attribution models impact the value of marketing channels. The calculated value will differ depending on the model used. The selection of a model depends mainly on a client’s business model and advertising objectives.

There are two types of attribution models. Custom attribution models and baseline attribution models both define how credit should be distributed to interactions in a conversion path. However, the difference between the two is that custom attribution models are created by individuals with their own set of rules. Baseline attribution models include:

  • First interaction attribution model
  • Last interaction attribution model
  • Linear attribution model
  • Time decay attribution model
  • Position-based attribution model
  • Last non-direct click model
  • Last Google AdWords click

Auditing

Auditing is a way to govern data by maintaining records of significant events in a BI platform. These records show what information is being accessed, how it is being accessed, and who is accessing or changing it. Not only does auditing ensure proper data security, it also acts as a way of monitoring usage and identifying areas for improving self-service capabilities.

Empowering users with the freedom to use data doesn’t mean giving up control. Auditing is one way of fostering an environment that encourages self-service while also governing the access to data in a centralized manner. Through auditing, we can apply the necessary security controls to ensure users are given the right data access.

Authentication

Authentication is a data reporting best practice involving the verification of a user’s identity. Along with authorization, it may be used to grant or deny access to reports, columns, and records to selected users or user groups in BI applications.

Essentially, authentication is a way of granting individuals access only to the data they need. Integration with existing authentication systems makes it easier for users to access self-service capabilities.

Authorization

Authorization is a data reporting best practice involving verification of a user’s identity and the subsequent provision of access rights. Along with authentication, it may be used to grant or deny access to reports, columns, and records to selected users or user-groups in BI applications.

Essentially, authorization is way of granting individuals access only to the data they need. Integration with existing authorization systems makes it easier for users to access self-service capabilities.

Automated Business Alerts

Critical Information via Automated Business Alerts

An alert is an automated message or notification sent via email, pager, etc., which indicates that a predefined event or error condition has occurred and that some action is needed.

Alerts allow users to receive critical business information in the quickest and most efficient possible way. For example, a store manager can be automatically informed when in-stock levels of a critical items fall below or rise above a certain level.

Benefits of Automated Business Alerts

Today’s better BI solutions you can set up custom alerts to get critical information to the right people at the right time. Developers can easily set up automatic alerts and notifications to be sent to certain users when specific data values or conditions occur in a report. For example, if inventory levels are critically low, an alert can be sent to the appropriate line manager to take appropriate action, or if a sales amount is over r1M, a notification can be sent to the sales support staff and relevant management team members.

Automated Processes

Automated Processes and Integration with Other Business Applications

In addition to pulling and presenting information in a report, developers can set up automatic database updates from within the report. For example, developers can build reports that allow business users to update inventory levels right within their inventory reports, or cancel a product shipment if a report shows that a product is out of stock.

Developers can seamlessly integrate your BI output in any of your other business applications using Web services. For example, if a customer fills out a form on your Web site and clicks “submit,” you can make sure it sends that person’s contact information straight to your CRM application.

Benefits of Automated Processes

Automated processes help BI users become more efficient in the actions they take. Instead of noticing something on a report, then having to pick up the phone or open another application to take action, automated processes allow the user to act directly from within the report.

This ensures that actions occur in a timely fashion; also, it ensures consistency between the data and the numbers occurring within the BI application and those used in the rest of the company.

Automated Report Scheduling

Automated report scheduling and delivery takes information proactively to the right users at the right time.

Report scheduling helps streamline information delivery as well as can help you optimize the running of many reports on your network, for example, during off-peak hours. Developers can add automated processes like report scheduling and delivery.

Developers can also set up automatic export (for example to Excel or PDF) and delivery of a report to specific subscribed users on a regularly scheduled bases.

For example, if sales managers need a weekly sales report, developers can automatically schedule the creation and delivery of the report to sales for every Monday at 9:00 AM. They can also specify that reports be automatically created and delivered to the file system or as email attachments.

Benefits of Automated Report Scheduling and Delivery

  • Allowing information to be distributed efficiently and consistently to the relevant end-users
  • Ensuring that heavy querying does not occur during peak hours, thereby placing a lighter load on the system or network
  • Letting end-users know, in real time, about changes in the numbers, especially if reports contain useful visualization features such as gauges, KPIs, etc.

B

Balanced Scorecard

A balanced scoreboard is an analysis technique that translates an organization’s mission statement and business strategy into specific, measurable goals, and monitors the organization’s performance in regards to achieving these goals.

First developed by Robert Kaplan and David Norton in 1992, this methodology is a thorough approach that examines an organization’s entire performance, based on the fundamental that measuring performance through financial returns only provides information about how well the organization did prior to the assessment. The future acts can then be forecasted and the right actions can be taken to produce the desired outcome. It provides feedback around both internal and external processes and outcomes in order to continuously improve performance and results.

The balanced scoreboard approach provides a clear method as to what companies should measure in order to balance out financials by recognizing some of the weaknesses of traditional management approaches. Kaplan and Norton describe the approach as “retain[ing] traditional financial measures. But financial measures tell the story of past events, an adequate story for industrial age companies for which investments in long-term capabilities and customer relationships were not critical for success. These financial measures are inadequate, however, for guiding and evaluating the journey that information age companies must make to create future value through investment in customers, suppliers, employees, processes, technology, and innovation.”

Benchmarking

Benchmarking is the process of comparing your business processes and performance metrics to industry standards. Benchmarking assists in helping set meaningful targets, gaining insight on trends across industries, determining what improvements are needed, and helping to enhance overall performance.

On its basic level, benchmarking puts business performance into perspective. In a management context, businesses use benchmarking to identify the areas that are being underperformed or over-performed. With benchmarking, organizations are better prepared to make data-based decisions about improvement initiatives, while simultaneously revealing those success stories that can form the basis of future campaigns. Benchmarking compares how cost, time, and quality are used between different organizations. The result is often a business case for making changes to lead to greater improvements.

The widely accepted benchmarking process, developed by Robert Camp, who wrote one of the earlier books on benchmarking, consists of a 12-step approach:

I. Select the subject

II. Define the process

III. Identify the potential partners

IV. Identify the data sources

V. Collect the data and select the partners

VI. Determine the gap

VII. Establish process differences

VIII. Target future performance

IX. Communicate

X. Adjust the goal

XI. Implement

XII. Review and recalibrate

Big Data

Big data refers to the ever-growing volume of data, increasing velocity in the generation of that data, and increased variety of types of data. In 2015, adoption of big data skyrocketed across all sizes of businesses, as demonstrated by Synnect Analytics State of Self-Service BI Report. And it’s not just IT teams that are looking to take advantage of all of this vast structured and unstructured data. Business users want access to these data sets as well.

There are many offerings in the marketplace attempting to address big data challenges, and it is a fast evolving marketplace. In this changing environment, customers are looking to understand how these new solutions can be combined with their traditional approaches to create new applications once thought impossible, to develop competitive advantages, and to reduce costs.

More than ever, enterprises are focused on using data to solve business problems. We are at the point where there are common design patterns and reference architectures for utilizing the various big data types of emerging data repositories. Deployment of such technologies is also much easier, especially through the cloud.

At the end of the day, big data is becoming… well, just data. The concept of big data will shift into the background from the end-user perspective. It will be par for the course. Technologists will continue to use the best tools available to them to tackle real business problems.

Big Data Sources

Big data sources are repositories of large volumes of data. Using business intelligence applications, users can quickly connect to and derive value from these sources. This brings more information to users’ applications without requiring that the data be held in a single repository or cloud vendor proprietary data store.

Examples of big data sources are Amazon Redshift, HP Vertica, and MongoDB. Emerging big data sources – such as analytic/columnar data stores, NoSQL, and Hadoop data repositories – expect to each more than double their rate of adoption and eventually exceed well over 40 percent availability in self-service tools within two years.

Bubble Charts

Bubble charts are data visualizations best used for showing three dimensions of data – comparing entities in terms of their relative values, positions, and size. Bubble charts are similar to scatter plots, but they use bubbles instead of data points.

Bubble charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Build Versus Buy

When faced with the need to embed analytics into an application, most software providers arrive at the crossroads of the “build versus buy” decision. The first step in tackling this question is to understand your requirements. Determine your desired end-user functionality, prioritize your needs, and then evaluate the feasibility of building such capabilities.

In order to properly evaluate your implementation options, it is important to qualitatively – if not quantitatively – assess the benefits and costs of each option, as well as compare the ROI for each. Your analysis should also include the opportunity cost and project risk associated with your skilled development staff spending less or more time focusing on your core product. By building a cost-benefit analysis over time, you can calculate the ROI for each buy or build option.

To some, “build” may seem like the obvious choice for embedding analytics functionality in their application. However, even if it looks like the less costly option from the investment standpoint, it may not be the most worthwhile option. Compared to coding on your own, utilizing a third-party party product may increase your cost in software licensing, but it also reduces your cost of development, both initially and ongoing.

Business Analytics

Business analytics, or business intelligence (BI), gives customers the ability to rapidly create scalable, interactive data analysis applications and self-service capabilities users can access from anywhere and on any device. BI is most optimized for supporting managerial-level decisions that require highly aggregated views of information from across a department, function, or entire organization.

Typically, technology professionals centrally manage and secure access to the data for business users. Analytics may also be embedded directly into a customer’s operational business applications, ensuring data insights are in the context of the applications users work with every day.

Business Intelligence

Turning Data into Actionable Information

Business intelligence (BI) has been defined in many ways. By the earliest definition (1958), business intelligence was seen as “the ability to apprehend the interrelationships of presented facts in such a way as to guide action towards a desired goal.”

A broader and perhaps more current definition of this discipline is this: business intelligence is the process of collecting business data and turning it into information that is meaningful and actionable towards a strategic goal. Or put even more simply, BI is the effective use of data and information to make sound business decisions.

Business intelligence encompasses the following elements:

  • Reporting: the process of accessing data, formatting it and delivering it inside and outside the organization
  • Analysis: identifying patterns and establishing relationships in a group of data
  • Data mining: the extraction of original information from data
  • Data quality and interpretation: the greater or lesser correlation between data and the real-world objects they represent
  • Predictive analysis: a branch of data mining, it attempts to predict probabilities and trends

Reporting and analysis are the central building blocks of business intelligence, and the arena in which most BI vendors compete by adding and refining features to their solutions.

The general process of business intelligence is as follows:

  • Gathering data and organizing it through reporting
  • Turning it into meaningful information through analysis
  • Making actionable decisions aimed at fulfilling a strategic goal

Data: The Raw Material

The raw material of business intelligence is the data that records the daily transactions of an organization. Data may come from such activities as interactions with customers, management of employees, running of operation or administration of finance. According to the traditional model, data from daily transaction is recorded in three main transactional databases: CRM (customer relation management), HRM (human resource management) and ERP (enterprise resource planning). For instance, a sales transaction would be recorded and stored as a piece of data in the CRM database.

A piece of data, in itself, is neutral–i.e. neither “good” nor “bad.” For instance, if you knew that rep X had received Y dollars worth or orders year to date, you wouldn’t necessarily know whether it’s a cause of panic or celebration.

Just like raw material, data needs to be processed through analysis to become meaningful. The same piece of data in the example above would become meaningful (for instance) if compared to year-to-date sales target for rep X. By doing this, the piece of data has become part of the process of analysis.

Analysis: Contextualizing the Data and Answering Questions

Analyzing data means asking it questions and getting meaningful answers. For example, the simple command “sort in descending order” on a column of data in Excel representing year-to-date orders taken by sales rep would answer the questions “Who is taking the most orders? The least orders?” The sort command has contextualized the data, making it much more meaningful in terms of the strategic goals of the business.

Of course, analysis in BI is much more complex and varied than this. The powerful and interactive analysis tools of today’s better business intelligence solutions make it easier to ask data an increasing number of questions and getting meaningful answers–including “what-if” scenarios, multidimensional slicing and dicing (XOLAP analysis), mashing up of data with geographic mapping and much more.

For example, data analysis features can answer such questions as:

  • How is my product performing by product line? What about by territory? Or by demographics?
  • What is the untapped potential of sales territory X?
  • What would be the likely impact of revenue if I eliminated territory Y and relocated Y’s rep to territory X?
  • Are my reps balancing face time with their customers with “windshield time” in an efficient way? Is there a way to improve this?

In any case, the goal of even the most sophisticated analysis features is always the same: enabling decision-makers to understand data, to spot patterns between numbers, to identify trends and the reasons behind them–simply put, to contextualize data and answer questions about it.

Making Decision and Taking Actions that Are Strategically Relevant

Interestingly, most BI projects fail not because of faulty technical implementation, but because of lack of a strategic focus. Business intelligence should be a lever that enables a company to “lift” itself more efficiently towards its strategic goals. But all too often, BI becomes an end-in-itself proposition, with project managers, CIOs or CTOs failing to look at it in light of the company’s mission.

C

Capabilities Map

A capabilities map is a tool for matching users to the analytics functionality they need. With analytics, there’s certainly a lot of functionality that can overwhelm users; this includes any of the visualizations, interactivity, and data that is displayed.

But for many users – and especially for those who are just starting out – it’s best to give them only the functionality and specific data they need to work smarter. You can release more functionality and data as adoption grows and new questions arise.

Centralized BI

Centralized BI is the system of storing and managing all business intelligence, or data, in one central location or department – typically corporate – for governance and security reasons. This model ensures data accuracy, veracity, and economy of scale.

However, companies with centralized data warehouses may also experience significant bottlenecks because all data requests go into a single queue. This creates a backlog for the team responsible for those requests – often the IT department – and ultimately results in frustration and loss of time and money. But the other option, a truly decentralized environment, can also be problematic. It may result in insufficiencies or redundancies in BI software, staff, and applications across a company.

On the one hand, organizations today want centralized data – for governance and security reasons. At the same time, organizations want a decentralized mode of operation for data discovery as well as to find new insights in a self-service fashion without increasing the workload of the reporting team.

The “best of both worlds” approach is distributed BI, a form of decentralized BI that gives individuals greater freedom to use data while still governing their access to that data in a centralized and scalable manner. In short, distributed BI makes self-service analytics possible. It allows users to discover data insights on their own without increasing the workload or creating a backlog for the reporting/IT team.

Charts

Charts are visualization tools that present analytical data in a tangible, impactful way. No BI reporting and analysis solution is complete without providing ways to visualize information. Charts and graphs help users better understand their data and provide a fast, more meaningful view in context, especially when comparing data.

It’s important to choose the most appropriate visualization tool for the type of information you want to display. Chart types range from standard flat charts – including line, pie, bar, stacked bar charts, and so on – to more advanced three-dimensional charts like heat maps, text clouds, and GIS maps.

It should be immediately apparent what users should take away from a particular chart or graph, so don’t bury the point in a complex design. End-users need to be able to glance at their dashboard and understand what they are looking at, whether targets are on track, and how to take action when they’re not.

For example, a pie chart is best used to show relative proportions between pieces of data, while a line chart is best used to track an item overtime. A heat map is ideal to show two column values of multiple rows of data, while geographic mapping is becoming essential for spatially relevant data.

Many business analytics vendors, including Synnect Analytics, will help identify which visualizations are most effective for the type of data being examined. It’s then up to the user and/or designer to customize and create the visualizations.

Cloud Analytics

Cloud analytics is a service model in which data analytics are provided through a public or private cloud. Cloud analytic applications and services are typically offered under a subscription-based or utility (pay-per-use) pricing model. It’s designed to make official statistical data readily categorized and available to users through web browsers, and is specifically designed to assist clients in extracting information from massive data.

The six key elements of analytics include data sources, data models, process applications, computing power, analytic models, and sharing/storing results. It’s often mentioned that any analytics initiative that has one or more of these elements implemented in the cloud qualifies as cloud analytics.

Columnar Data Store

Columnar data store is a type of big data repository containing structured data in columns and rows. The main benefits are that the data can be highly compressed and is easily searchable.

This emerging big data source is expected to more than double its rate of adoption and eventually exceed well over 40 percent availability in self-service tools within two years.

Commingled

Commingled data is information that is stored in a shared infrastructure, such as multiple related or connected databases. By choosing an analytics platform that supports multi-tenancy, Synnect Analytics, Software as a Service (SaaS) providers can configure security regardless of whether their customer data is stored in commingled data sources or a unique database.

Comparing Today’s BI Models

If we look at the main Business Intelligence tools and software available to today’s company, we see the following models: free BI, open-source BI, on-demand BI, legacy BI and dynamic Web-based commercial BI solutions. Each model has benefits and drawbacks, as we will see directly.

Free BI

Are there free BI solutions out there, ready for download, that can get a company squared away with business intelligence? Yes, and some of them offer a high value, in the sense that they are truly free, are easy to connect, don’t require much maintenance and are robust enough to give the company the tangible benefits of BI. For the sake of this discussion, we will only consider solutions that have these advantages–it’s up to the project manager shopping for BI to ensure this so that the company doesn’t get hooked on a “free” product that turns into a money-pit and is hard to work with to boot.

A good free BI solution has many advantages. First of all, it requires no upfront costs, which is a boon, especially in these hard economic times. Then, a good free BI solution will not be resource-intensive to build, use and maintain–which would conversely negate the savings enjoyed initially. As far as connection, it will offer the possibility to link to various types of traditional and nontraditional data sources, such as databases, Web services and flat files.

And it will do so out of the box.

Some of the better free BI solutions are Web-based, are embeddable into other software solutions or applications and are relatively robust and feature-rich. This means that if they are used to their fullest or near-fullest potential, they have the capability to empower many users within the organization to report on and analyze their data in a reliable, efficient and creative way, giving the company a chance to become much more competitive than prior to having BI. Features like tabular, cross-tab and free-style reports, drill-down and drill-through capability, rich visualization features and the ability to export reports to common formats such as Excel, CSV, Word, HTML, etc. will be in many cases all that the users need to become measurably more effective in their tasks.

Lastly, the better free BI solutions will not require upgrading to a commercial solution, but will leave that as an open option for the company. And at upgrade time, with a good free BI solution, reports won’t need to be rebuilt, making the transition smooth and seamless.

A good free BI solution has very few drawbacks. If the alternative is no BI or investing time, resources and cash into a risky project, free Web-based BI available for download is the proverbial no-brainer. Sure, you may not have at your disposal the latest in dashboarding, complex data analysis tools, robust user-driven ad-hoc reporting, OLAP or advanced visualization tools like heat maps, but as far as the basics, you will be covered. And you will have these basics out of the box.

Open-Source BI

Commercial open-source BI works in this manner. The vendor takes a product that was created in the open source community and makes it their own so that they can market it. Since by definition an open source product cannot be sold, commercial open-source vendors make money through services, support and any add-ons that they have built themselves. So, although they are not selling the core product, they are still selling something.

One of the things that makes a pure open-source model attractive is the flexibility it offers for customization, although this comes with a substantial flip-side. The buyer has access to the source code, so his team can add, modify or delete anything they want. But here’s the rub: as soon as they do this, they’re deviating from the source. So at that point they either need to become active participants in the community (submitting their changes for everyone else to use), or they have to move further away from the core and hope not to run into any major landmines within the source that they need to ultimately fix by themselves.

Another initial lure of working with a commercial open-source vendor is the low cost of entry–since the product is ostensibly free. However, the services and support that commercial open-source vendors provide is essential to helping the client get started. Once the client goes down this services and support road, however, they face the challenges described above (they’ve deviated from the source) and now they’re even more dependent on the vendor for services, support and add-ons.

Another negative is the fact that there is no real accountability if something goes wrong. Who do you turn to if there is a major problem with the product? Can you go back to the community to get the bug fixed? Possibly, but you’re likely not going to get a resolution very quickly. Can you go back to the vendor? Perhaps–unless they’re also waiting for the same fix from the open source community. This sort of bottleneck actually happens quite frequently in the commercial open-source market: the same bug exists within the commercial open source as does the main open source project. The customer can’t get their situation resolved by the vendor, because the vendor is waiting for the community to fix the problem–with the sense of urgency of a more or less voluntary community.

Therefore, the fact that services and add-ons still have to be paid, plus the uncertainty of how the project will be supported in case something goes wrong often makes open-source BI risky.

On-Demand BI

On-demand BI (as well as software as a service or SaaS) is another model that has become popular in recent years. The way it works is by offering some of the benefits of BI without the hassles of hosting an application in-house. In other words, the vendor keeps, hosts and manages the application, while the client uses and pays for the application on demand through the Web. There is a whole host of companies that have specialized in on-demand business intelligence solutions or components, and, like open-source BI, on-demand has become a viable alternative to more traditional models.

The drawbacks of on-demand BI stem primarily from three factors. Firstly, the greatest majority of them are not targeted either for the smaller company nor for the larger one. The smaller company can be just as well served by free BI–let’s remember that on-demand BI is not free–while the larger one is better served acquiring, refining and maintaining their own in-house BI applications.

Then, to quote analyst Boris Evelson, “BI is still an art much more than a science. It still takes an army of consultants to pull it together, and whether I’m hosting [BI] somewhere in the cloud or doing it in-house, I’m still going to go through exactly the same difficulties. And as long as I’m doing that, why would I want to release–or lose control over–my BI installation to a third party vendor?”

And lastly, there is the issue of safety. Putting critical data and sensitive information in someone else’s hands and outside of the company’s own firewalls, as is the case for a solution hosted in the cloud, is not something that all firms are willing to do, which is not hard to understand.

Traditional Legacy BI

By definition, traditional legacy BI is ill-suited to all but the largest firms. This is because, as we have mentioned, it grew out of the needs, the budgets and the timelines of blue-chip companies.

Aside from the huge expense associated with buying, implementing, maintaining and upgrading traditional legacy BI, there are a host of other issues that make these solutions less than ideal.

The many mergers and acquisitions that these companies have undergone in recent years mean that much of their product offering is technologically heterogeneous (at best) and uncertain (at worst). Of the tens of disparate BI products offered by some of the legacy firms, which ones will run on the same technological platform? Which will force the client’s IT department to implement, learn and maintain products borne of radically different philosophies? And most importantly, which will still be supported by the vendor next year–and how can the client know before he buys?

Another drawback is that the prevalent licensing model for legacy BI is user-based. This means that if the firm wants BI to be truly pervasive–as it should be–there will be a substantial cost.

But there is another BI model that is much more advantageous to today’s firm than all of the ones we have just mentioned.

Dynamic Web-Based BI

Perhaps the best fit for today’s average firm is with dynamic, Web-based BI solutions offered by vendors that have been specializing since their inception on this kind of software. The main differences between this model and those we have just discussed are:

  • The solutions are ready out of the box, unlike open-source BI. And although they offer several degrees of customization, they are much more predictable (in the good sense of the word) as to what they can do and what they will look like once implemented
  • The solutions are modular, meaning that a company can buy only as much BI as they need, unlike some of the legacy-BI solutions
  • The solutions allow companies to host and own their own BI application, unlike with on-demand BI products.

The advantages of these newer, Web-based BI solutions are many, namely:

Easy to get started – From pricing to connection to set up, the better Web-based BI solutions save companies time and money. They do not require multiple consulting trips, and they can easily connect to one or more of the most common data sources that a midsize firm is likely to use–from databases to Web services to flat files. In some cases, these vendors have their solutions available for free-trial download, so the decision-maker in the midsize firm can test the solution with his company’s own data and evaluate it against the backdrop of his own technological architecture and real-life issues.

Easy to use – Solutions that were born to run on the Web–and are not adaptations–have the easy feel and navigability of the Internet. This is advantageous to both the report developer, who can prepare feature-rich, dynamic reports with little coding and using a wizard-driven development approach, and to the end-user, who will find reporting and analysis intuitive. In turn, this will benefit adoption: if the solution is adopted enthusiastically by as many users as possible, decision-making will become more efficient and (ultimately) the firm will become more competitive.

Powerful and interactive – BI companies that were sufficiently forward-thinking ten years ago to bet all their chips on Internet technology are often the ones that introduce or are quickest to adopt the features that are truly useful. For this reason–and without departing from their easy-to-use philosophy–these BI vendors offer features like interactive dashboards, powerful visualization tools like heat maps and GIS maps with drill-down and drill-through capability, animated charts and graphs, intuitive OLAP analysis and user-driven ad-hoc reporting.

Complete and modular – Today’s firms–especially midsize firms–should not be forced to buy more BI than they need or to settle for less for fear of buying too much. The key to their ability to buy just as much as they need is in the modular nature of the solutions offered by vendors. For instance, a company that has only a moderate amount of data and whose users share similar requirements may just need a managed reporting solution and should not be forced to buy a product that incorporates ETL, data marts or a data warehouse. Conversely, one that has complex data and many users with different needs can explore the possibility of acquiring a whole platform, as long as it’s complete, it’s technologically unified and it features components optimized for different tasks–such as, for instance, managed reporting, ad-hoc reporting and ETL/data integration. In this sense, the components of a good Web-based platform lend themselves to being points along a firm’s growth in size and data complexity–meaning that the company can get only what it needs now, and may plan on getting more BI in the future as data volumes and operations grow.

A good value – This point is easy to dismiss as intuitive–but it’s all but. Value is the ratio between benefits gained compared to effort required. Failing to measure either one right from the onset is why so many BI projects fail–either because, down the road, they yield no strategic value, or because they require too much effort for the benefits they bring. So, let’s start with value. A BI solution is valuable when it allows a firm to reach its strategic goals more efficiently. It does so by making data easily available, easily processed, easily understood and easily acted upon. It does so when it sifts through the white noise of less-than-critical information and pinpoints a vital action that a decision-maker must take. It does so through tools like KPIs, dashboards, automated alerts, meaningful visualization and analysis tools. And the main points of effort for a BI solution are upfront costs, IT costs, maintenance costs, licensing costs and upgrade costs. Good Web-based BI solutions bring value by offering the company all these benefits; while minimizing the effort required–through being much less expensive and resource-intensive, and through being licensed to empower as many users as a firm needs without per-user fees.

Consumers

Consumers are business users who prefer a defined analytics experience. They usually have a standard set of metrics and KPIs they need to track in order to do their jobs more effectively or to measure performance against target.

Consumers can be at any level of the organization, from the C suite to the factory floor. In many ways they’re the toughest persona on Synnect Analytics of Self-Service to satisfy because they don’t necessarily know what they want, but they’ll know it when they see it. While a consumer’s analytics experience is focused on consumption of pre-determined dashboards and reports, they still want to interact with the content and “play” with the numbers. They also want to personalize the display to their liking.

Examples of information consumers include:

  • A CEO who wants to review high-level performance metrics at the organization
  • Line worker who needs to see production versus target volumes being produced on the plant floor
  • Nurse who needs to monitor the readmission floor of a hospital on an iPad

Creators

Creators are business users who value a managed experience where they can query different data sets, create their own data visualizations, and share them with others.

Creators make up the majority of users in businesses. They are often responsible for delivering the weekly reports to measure their team’s performance to other departments and upper management. They want to supplement standard metrics with new dashboards and reports. It’s not uncommon for consumers to become creators over time as they become more comfortable with the data.

Examples of creators include:

  • An HR manager who may not want to look at not just employee hires across the company, but also by department
  • Production manager who wants to review just-in-time inventory metrics to ensure what bottlenecks are holding up the assembly line
  • Head of Surgery who wants to examine and compare the number of surgeries performed per week

CRM

CRM, short for customer relationship management, is a database that records a company’s daily customer-related transactions. CRMs can help customer representatives to provide better service, close more deals, and increase revenue. According to the traditional business model, day-to-day data is also recorded in two other transactional databases: HRM (human resource management) and ERP (enterprise resource planning).

However, transactional databases are not considered optimal for business intelligence. This is for a variety of reasons, including the fact that data is not optimized for reporting and analysis. Additionally, querying directly against these databases may slow down the system and prevent the databases from recording transactions in real time.

CSS

CSS is a web standard for customizing styling, branding, and visualizations. Along with HTML and JavaScript, it is one tool for building a consistent theme and layout in BI applications.

D

Dashboards & Dashboarding

Dashboards–sometimes called IT dashboards or corporate dashboards–are single screens in which various critical pieces of information are placed in the form of panels. Like dashboards in a car, they allow the end-user to have a unified view of the data and information that matters to “drive” the business forward.

If a dashboard is useful, a Web-based dashboard is even more so. Blending the power of a desktop application with the flexibility and the navigability of the Web, its panels can be as diverse as:

  • Business metrics such as charts and graphs
  • Key performance indicators or KPIs
  • GIS maps
  • Web sites, news, RSS feeds, real-time stock or currency quotes
  • Personal reminders
  • In short, most anything that can be displayed on the Web

Advantages of Dashboards

Dashboards are valuable because they transform business data into critical information that jumps out to the user, who can then make sense and act on it immediately.

  • Fast and effective decision-making – Gives executives, managers and analysts convenient immediate access to key performance metrics, which help them monitor performance and processes for a greater understanding of the business.
  • On demand, accurate and relevant information in line with business priorities – Dashboards clearly communicate business objectives throughout the organization and allow users to see progress towards those goals. This keeps everyone focused and informed. With a personalized layout, users only see the information that is most important to them, and they can filter out information that is not relevant.
  • Focused identification of problems, inefficiencies or negative trends for immediate action and improved performance Users can immediately see any problems and drill down on charts and links to explore detailed information and analyze data in real time, to determine root causes and to correct negative trends.

Best Practices Tips

As the “new face of BI,” a dashboard is an attractive feature for prospective buyers of business intelligence. Some go as far as almost thinking that a corporate dashboard has magical properties. It’s like a business talisman: just get it and in no time your decision-making will become more effective and your company more competitive. This is, of course, not the case. To be effective, dashboards need to be implemented smartly and with a view towards the company’s strategy.
Let’s look at some best-practice tips to ensure you get the most out of your dashboard investment.

Do #1: Let the Dashboard Be Business-driven and Focused

Ask yourself: what competitive goals are you trying to achieve through this tool? What specific processes are you trying to make more efficient? What critical information are you trying to make more readily available and why? Be ruthlessly specific. The more surgically you zero in on precise tactics, the better your chance to achieve your strategy.

Example: you want the inventory of the top-10 SKUs to always remain optimal, so that you’re not out of goods while never getting overstocked. You set up a dashboard that shows this information in intuitive eyeful–in graphic form and of course in real time.

Don’t #1:

Don’t make the dashboard into a slightly less unprofessional version of solitaire. Too much freedom and too little focus, and your users will spend time on it for entertainment with your BI investment going to waste.

Do #2: Let the KPI Be Your Friend

What’s a KPI? It’s a key performance indicator–a color-coded dot or gauge that “indicates” if your “key” items are “performing” well or if they need corrective action. Set a threshold (e.g. minimum month-to-date sales) for the critical items; when you’re on the good side of the threshold, the KPI shows you a green dot–all OK. When you’re on the wrong side of the threshold, the KPI turns red–time to take action.Example: you want to have an optimal in-stock level of your top 10 SKUs. Have 10 KPIs that alert you without even having to read numbers. Green: all is going well. Red: either too much or too little inventory.

Don’t #2:

Don’t use too many KPIs. The “K” stands for “Key.” Prioritize and use KPIs only for your key items, otherwise your dashboard will become too cluttered and important information will fail to jump out to your users.

Do #3: Make Your Dashboard Actionable

The thermostat in your car reads 38 degrees. Does knowing that make you any warmer? Not unless you can act on the temperature-control lever. Without being able to act on what you see, a dashboard is as useful as than the morning paper–it informs you but it does not give you a chance to do something about what you read. Give yourself the power to see the information, understand what it means to your goals and act on it without leaving the application.Example: one of your inventory-level KPIs is red. Time to reorder. Instead of leaving the application, looking up the vendor, entering another program and placing the order, you just click on the “reorder” button right from your dashboard.

Don’t #3:

As you implement BI, don’t foster a culture of “knowers.” Foster one of “doers.” Remember that it’s actions that impact the bottom line, and that knowledge is only the prerequisite–albeit a critical prerequisite.

Do #4: It’s a Web, Web World, Although…

With the Web taking over the world of BI, it’s become chic to malign desktop applications. Yes, having dashboards on the Web is almost essential today, making it easier to access them, share them and work on them from virtually anywhere. However, the best Web-based dashboard software still retains the features of a desktop application–flexible, easy to use, powerful, interactive, with that “dedicated” feel to them.

Example: you should be able to move your panels around without refreshing the screen (thanks to technologies like AJAX), plus drill down, drill through and have persuasive and impactful features like Flash-powered charts and graphs.

Don’t #4:

Don’t set up a Web-based dashboard that looks and feels like an Internet site from 10 years ago–a static, read-only tool whose usefulness is greatly watered down.

Do #5: Make Dashboard Software Available to Everyone

Us BI industry insiders may not realize it, but it’s still out there. That culture where reporting and analysis is the domain of a few techies or upper management. For it to be useful, dashboard software should be available to every decision-maker in your company. And if you are smart about the way you manage your people, most your employees should be treated as decision makers.

Example: there’s no reason why your warehouse managers, your HR personnel, all your sales-force and your finance department (to name but a few), should not have access to dashboards making their jobs more efficient.

Don’t #5:

Don’t end up paying for tens of user’s licenses, or worse yet, tens of user’s licenses that end up unused because of failed adoption. Shop for a vendor that allows you to deploy dashboards to unlimited users–e.g. through a server-based licensing model.

In Summary

In the end, remember that the dashboard is just a tool. The easier it is to use, and the more directly it makes your employers’ life easier, the more it will be adopted. And the more it is adopted, the more positively it will impact your business.

Data Discovery

Data discovery is the capability to uncover insights from information. It is revolutionizing how people work with data. Instead of relying on data scientists as users, today’s BI applications empower anyone to be an analyst.

Advanced data profiling techniques help business users make sense out of the data, providing recommendations and prioritizations to help filter out the noise and make it as easy as possible to discover insights. The next part of data discovery is sharing insights with others in a collaborative way so everyone else can benefit.

BI is moving away from a top-down model, where executives or IT decide what people should see in company-wide dashboards or reports. Those reports have their place, but companies are recognizing they need a bottom-up approach to analytics as well – where all employees are empowered with the right information and tools to make informed everyday decisions in their domains.

Data Enrichment

Data enrichment is a method of preparing data so it is ready for analysis and exploitation. For most business users, data enrichment is best done with a data prep tool designed for them, not data scientists or engineers.

This tool should allow users to access data in various systems (e.g., on-premise, cloud, spreadsheets), identify the data sets needed, blend the data, and get the data into the right format for analysis. From there, users can enrich the data further, combining like pieces of data into common fields (data blending), or adding new calculations as a new field.

Data prep needs to be appropriate for most any business person who is required to do analysis today – which is pretty much everyone at this point. This allows users to get their answers much more quickly and independently without IT assistance.

Data Governance

Data governance is a control that ensures that data entry by a business user or an automated process meets business standards. It manages a variety of things including availability, usability, accuracy, integrity, consistency, completeness, and security of data usage. Through data governance, organizations are able to exercise positive control over the processes and methods to handle data.

Data governance also describes the evolutionary process for a company, changing the company’s way of thinking and setting up the processes to handle information so that it can be utilized by the entire organization. It also includes using technology when necessary in many forms to help drive the process. When companies want to gain control of their data, they empower their people, set up processes, and seek help from technology to do it.

Some goals of data governance include:

  • Increasing consistency and confidence in decision-making
  • Decreasing the risk of regulatory fines
  • Improving data security
  • Maximizing the income generation potential of data
  • Designating accountability for information quality
  • Enabling better planning by supervisory staff
  • Minimizing or eliminating re-work
  • Optimizing staff effectiveness
  • Establishing performance baselines to allow improvement efforts
  • Acknowledging and holding all gain

Data Integration

Your data sources are where your transactional and corporate data reside. To report, analyze and act on this data, you need first to connect to your data sources and bring them together.

There are many different ways to bring data together. From various kinds of connectors to ETL tools (extract, transform and load), from mashups to Web services, from datasource-neutral BI solutions to ones requiring massive meta-infrastructures, you have many models to choose. In general, however, an ETL provides a means of collecting, optimizing, and storing that data to better serve your company’s reporting and analysis needs.

A small company working with few pieces of data from homogeneous sources can have the flexibility to manage this data in different ways. However, the higher your data volume and the more diverse your data sources, the harder to organize, manage and ultimately rely upon this data. This is when it is useful to switch to an ETL.

How ETL manages and creates a process around your data:

The extract step in an ETL job reads the data from one or more data sources. A good-quality Web-based ETL is “data source neutral” and is capable of reading data from almost any data source, including databases, flat files, spreadsheets, RSS/ATOM feeds and Web services.

The transform step in an ETL job manipulates the data gathered in the previous step. Here, data is combined, cleaned up, processed and optimized for reporting and analysis.

The load step in an ETL job takes the data collected and optimized and writes it back out to one or more destinations. In a good ETL, these can be almost any data source, including databases, flat files, spreadsheets, and Web services, RSS/ATOM feeds–just as is true of the extract step.

When Does Data Integration or ETL Become Necessary?

It is of course possible to report directly against your databases or data source(s). However, there is a point past which data volume, diversity of data sources and other important considerations make it desirable to have a data integration or ETL. If you are a data architect, developer or database administrator, here are some of the questions you need to ask yourself in this regard:

  • Is the volume of your data growing noticeably?
  • Is your company using an increasing number of data sources?
  • Do you need a convenient way to integrate your data across different applications?
  • Do you want to find a way to make your data more accurate and easier to understand?
  • Are you searching for an efficient way to manage or create a process around your data?

If you have answered any of these questions in the affirmative, you may need to look into acquiring a data integration or ETL tool

Data Journalism

Data Journalism is a type of journalism where reporters make use of large databases to produce stories. This reflects the increased role that numerical data is used in the production and the distribution of information, and the increased interaction between journalists and fields such as design, computer science, and statistics. The use of data journalism helps tell a complex story through the use of infographics and data visualizations.

Data journalism has been widely used to connect several concepts and relate them to journalism. It is often seen as having a number of stages leading from simple to complex uses of new technologies in the journalist process.

The areas covered by data journalism include:

  • Computer-assisted reporting and data-driven journalism (journalists make use of large databases to produce stories)
  • Infographics
  • Data visualization
  • Interactive visualization
  • Taking interaction a step further
  • Database journalism (pieces of information are organized in a database)

Data Mart

A data mart is the access layer of a data warehouse that is used to provide users with data. Data marts are often seen as small slices of the data warehouse. Data warehouses typically house enterprise-wide data, and information stored in a data mart usually belongs to a specific department or team.

The key objective for data marts is to provide the business user with the data that is most relevant, in the shortest possible amount of time. This allows users to develop and follow a train of thought, without needing to wait long periods for queries to complete. Data marts are designed to meet the demands of a specific group and have a comparatively narrow subject area. However, narrow in focus doesn’t necessarily mean small in size. Data marts may contain millions of records and require gigabytes of storage.

Advantages of using a data mart:

  • Improves end-user response time by allowing users to have access to the specific type of data they need
  • A condensed and more focused version of a data warehouse
  • Each is dedicated to a specific unit or function
  • Lower cost than implementing a full data warehouse
  • Holds detailed information
  • Contains only essential business information and data and is less cluttered
  • Works to integrate all data sources

The creation and use of a data mart leads to a great summarization of data. A much broader range of data is available with data warehouses; however, this data is generally not summarized, can make it difficult to sort through masses of data, and increases query times.

Data Mining

Data mining is the process of analyzing data from different sources and summarizing it into relevant information that can be used to help increase revenue and decrease costs. Its primary purpose is to find correlations or patterns among dozens of fields in large databases.

Data mining software is one of many analytical tools for reading data, allowing users to view data from many different angles, categorize it, and sum up the relationships identified. The ultimate goal of data mining is prediction and discovery. The process searches for consistent patterns and systematic relationships between variables, then validates the findings by applying the patterns to new subsets of data.

Data mining consists of five major elements:

I. Extract, transform, and load transaction data onto the data warehouse

II. Store and manage the data in a multidimensional database system

III. Provide data access to business analysts and IT professionals

IV. Analyze the data by application software

V. Present the data in a useful format (graph, table, etc.)

The process of data mining is simple and consists of three stages. The initial exploration stage usually starts with data preparation which involves cleaning out data, transforming data, and selecting subsets of records and data sets with large number of variables. Then, identifying relevant variables and determining the complexity of models must be done to elaborate exploratory analyses using a wide variety of graphical and statistical methods

Data Modeling

Data modeling is a method of preparing data so it is ready for analysis and exploitation. Today’s BI applications include powerful tools that automate the traditionally time-consuming work of preparing, modeling, and profiling data for analysis.

Most of these apps use the OLTP (online transaction processing) data modeling approach. OLTP systems focus on fast query processing, while maintaining data integrity in multiple access environments and effectiveness, measured by number of transactions per second. An important attribute of an OLTP system is its ability to maintain concurrency. To avoid single points of failure, OLTP systems are often decentralized.

Synnect’s data modeling tool automatically groups data into intuitive bins, such as income bands from salaries; age ranges from ages; or weeks, months, or quarters from dates. Users can quickly and intuitively assess data distributions and patterns using dynamic or custom bins – all without the complex data modeling effort of years past.

Data Profiling

Data profiling is a method of preparing data for visual analysis. We’ve all spent countless hours grappling with spreadsheet software to manipulate text, set up pivot tables, or create charts. Advanced BI applications, like Synnect Analytics, simplify data complexity and analysis by automating the tasks of data profiling, modeling, and blending.

In Synnect Analytics, a recommendations engine suggests best-fit visualizations, while dynamic binning creates logical data groupings. Altogether, these profiling tools shape the data in a way that makes sense for analysis, empowering all users to make data-driven decisions regardless of their technical expertise.

Data Scientist

A data scientist analyzes data to help a business gain a competitive edge. A data scientist represents someone who has evolved from a business or data analyst role. The training and education is similar, with a foundation in computer science, mathematics, programming, modeling, or analytics. Data scientist typically have a strong business acumen, and are able to communicate findings to both business and IT users in a way that can influence how an organization approaches challenges.

Data scientists have been described as “part analyst, part artist” by Anjul Bhambhri, vice president of big data products at IBM. He says “a data scientist is somebody who is inquisitive, who can stare at data and spot trends. It’s almost like a Renaissance individual who really wants to learn and bring change to an organization.”

What differentiates a data scientist from a traditional data analyst is the number of sources used to look at the data. A data analyst typically looks at data from a single source, and a data scientist will examine and explore data from multiple sources. He or she does not simply just collect and report data, but also looks at it from many angles and perspectives, determines what it means, and recommends ways in which the data can be applied.

As an interdisciplinary subject, data science draws out information from a broad range of areas including:

  • Data mining
  • Cloud computing
  • Databases and information integration
  • Signal processing
  • Learning, natural language processing, and information extraction
  • Computer vision
  • Information retrieval and web information access
  • Knowledge discovery in social and information networks
  • Information visualization

Data Source Types

The data companies analyze through business intelligence comes from a diverse type of data sources. The most common of these are:

  • Databases
  • Flat files
  • Web services
  • Other sources such as RSS feeds

Databases

Databases are the most traditional kind of data source in BI. There are many different kinds of databases, and many vendors providing databases with different architectures and different features. Common databases used today include MS Access, Oracle, DB2, Informix, SQL, MySQL, Amazon SimpleDB and a host of others.

Traditionally, transactional databases—namely the ones that record the company’s daily transactions, such as CRM, HRM and ERP—are not considered optimal for business intelligence. This is for a variety of reasons, including the fact that a) data is not optimized for reporting and analysis and b) querying directly against these databases may slow down the system and prevent the databases from recording transactions in real time.

In some cases, companies use an ETL tool to collect data from their transactional databases, transform them to be optimized for BI and load them into a data warehouse or other data mart. The main downside of this approach is that a data warehouse is a complex and expensive architecture, which is why many other companies opt to report directly against their transactional databases.

Flat Files

Few companies today have not used Microsoft Excel spreadsheets. Their ease of use and widespread employ makes them as popular as ever.

Data Storytelling

Data storytelling is a method of visually presenting data to make it more understandable and easy to digest. Visualizations such as charts and graphs guide users toward a conclusion about their data and empower them to make a decision based on that conclusion.

How does storytelling work through visualizations? Let’s start with our brains. How the brain best learns and retains information is reliant on understanding how it processes the information coming in. As we see information, it forms a visual pattern so that we quickly draw attention to key observations.

So ideally, it makes sense that users can grasp the meaning of data when it is displayed in visual form, rather than spreadsheets or numbers scattered on a document. As you think about it, you’ll soon realize that most numbers should always be presented in context. For example, sitting alone, 6.2% can literally mean almost anything. But taken in context, explained by story, not only does it mean something specific, it can move your audience to actually do something about it.

So the first rule of storytelling with data is find and present the context. This will almost always involve multiple series of numbers related to your data point. The second rule is to always remember you are speaking to an audience – and that audience is seeking to hear something that informs them, that moves them to action when necessary, and that reassures them when all is well.

If you can get a tight focus for that audience, then you can narrow your story and tailor it just for their needs. For a broader audience, you’ll want to accommodate a wider spectrum of understanding. This can be done with multiple charts, drill downs, dense information displays, etc. There are a number of viable ways to show multiple-variable data.

But the most important thing is to simply ask your audience, “What information do you need from me and what form do you need it in? What do you need or want to understand about this data?” Get them to draw it on a whiteboard, or lead them through a short example or two.

It can help to think of your BI dashboard the way a screenwriter thinks of a storyboard: A progression of data and images that leads the viewer from a beginning to an end. And like any good storyteller, you develop a story over time, with multiple drafts, and much editing.

Data Tables

The data table is perhaps the most basic building block of business intelligence. In its simplest form, it consists of a series of columns and rows that intersect in cells, plus a header row in which the names of the columns are stated, to make the content of the table understandable to the end user. This type of table is known in BI as a tabular report. A tabular report is used primarily–but not exclusively–to record information.

Data tables are the most basic component of BI, and still one of the most useful.

For example, if you are the sales manager of a company, you may have a tabular report in which five columns represent order dollar amount, order quantity, salesman and territory.

Another common type of data table in BI is the cross-tab report. With a cross-tab report, data starts to be grouped and organized in a more summarized way, making it more intelligible and therefore more useful for BI.

A cross-tab report is a data table in which there is not only a header row, but also a column (typically the left-most) that groups data in an intelligent way. Using our example of the sales manager’s report, a cross-tab version would group data by salesman (column on the left), and display the others as total dollar amount, total order quantity and territory.

The Benefits and Drawbacks of Data Tables

A data table is in many senses the lowest common denominator of reporting in BI. Although its role is more skewed towards reporting than analysis, it still holds a vital role in business intelligence, provided its limitations are correctly understood.

  • Ability to store large numbers of records in an easy, intuitive format – This is one of the reasons why Excel spreadsheets are still popular–and will likely remain popular for a long time to come. Records–including historical records–can exist on a spreadsheet for as long as the spreadsheet is saved, and are therefore always accessible to the user whenever required.
  • Ability to construct data summaries to make analysis easier – A spreadsheet or data table can form the basis for a summary cross-tab report (as in the example we gave above), moving the table closer to being an analysis tool rather than a mere means of reporting.
  • Ability to add a number of analysis capabilities, such as sort, drill-down and drill-through – When these features are added to a data table, its role to not just present data but to make data understandable in the form of information is greatly enhanced.
  • Limited capability to make key information jump out – The main drawback of a data table–especially a tabular report–is that it tends to present a vast amount of data in a neutral way, and it leaves it up to the end-user to sift through it or analyze it to make business sense of it.

Data Visualization

No reporting and analysis solution is complete without providing ways to visualize information. Charts and graphs as well as more advanced data visualization tools help users better understand their data and provide a fast, more meaningful view in context, especially when comparing data.

From a wide range of standard flat charts–including line, pie, bar, stacked bar, and so on–to more advanced, three-dimensional and dynamic Flash charts or to features like heat maps, text clouds and GIS maps, visualization features offer a way to present data and information in a tangible, impactful way.

The Benefits of Data Visualization

Humans are visual animals. Even in our common language, to “see” means to “understand.” And in today’s fast-paced business environments, scanning through rows of data can be time-consuming and impractical. Many businesspeople want easy visual tools to see accurate, real-time business-critical information. They may need to start with the big picture and further explore the details as needed. Or, they may need to spot exceptions and identify emerging trends to take immediate, appropriate action.

Cutting-edge visualization tools show high-level summaries of important data. They present information in clearly defined spaces using shape, size and color to provide context and meaning to the user, who can identify trends and get insight at a single glance.

Lastly, data visualization has great persuasive power. To show two figures on a data-table and to display these same figures side by side on a chart or graph is quite different in terms of impact.

Data Visualization Best Practices Tips

  • Be selective about which types of data you want to visualize. Data visualization can be viewed as a visual form of analysis with the purpose of making critical items jump out at the user. If too much data is placed in visualization form, it may lose its impact and its usefulness. Stick with the critical items.
  • Choose the most appropriate visualization tool for the type of data and/or information you want to display. For example, a pie chart is best used to show relative proportions between pieces of data; a line chart is best used to track an item overtime; a heat map is ideal to show two column values of multiple rows of data; geographic mapping is becoming essential for spatially-relevant data, etc.
  • Leverage the power of the Web to place visualization tools outside of the strict confines of your reporting interface, so that your users are further encouraged to consume key information. For instance, place relevant information displayed in charts and graphs in your company’s wiki or intranet, on blogs, CRM, etc.

Data Warehouse

A data warehouse is a central storage for all data that an enterprise’s various business systems collect. Developing a data warehouse includes production of systems that can extract data from operating systems and integrate data from one or more disparate sources. Additionally, the installation of a warehouse database system provides users flexible access to the data.

Data warehouses possess five key characteristics:

I. Data from multiple operational databases is combined.

II. Data is certified to be of higher quality. Low-quality data is cleansed before entering the warehouse.

III. Data is read-only. It cannot be changed by end users.

VI. Data is historical and represents a series of snapshots depicting the state of businesses at different points in time.

V. Data warehouses are frequently large and usually in the multi-gigabyte range.

There are two different approaches to data warehousing: top down and bottom up. The top down approach spins off data marts for specific groups of users after the complete data warehouse has been developed. The bottom up approach builds data marts first, then combines them into one data warehouse.

There are many benefits of data warehouses. They provide the opportunity to:

  • Collect data from multiple sources into one database so a single query engine can be used to present data
  • Improve the quality of data
  • Provide a single common data model for all data regardless of the data’s source
  • Maintain data history
  • Make decision-support queries easier to write
  • Restructure the data to be easily understood by business users

Decentralized BI

Decentralized BI is the system of storing and managing business intelligence, or data, independently by respective department, rather than centrally. Essentially, every division handles its own data requirements and any analytics applications they need with little corporate involvement.

A truly decentralized environment can result in insufficiencies and/or redundancies in BI software, staff, and applications across a company. On the one hand, organizations today want centralized data – for governance and security reasons. At the same time, organizations want a decentralized mode of operation for data discovery as well as to find new insights in a self-service fashion without increasing the workload of the reporting team.

The “best of both worlds” approach is distributed BI, a form of decentralized BI that gives individuals greater freedom to use data while still governing their access to that data in a centralized and scalable manner. In short, distributed BI makes self-service analytics possible. It allows users to discover data insights on their own without increasing the workload or creating a backlog for the reporting/IT team.

Designing for Color-Blindness

Color-blindness occurs when a user is unable to see visual differences in charts and other graphics. Typically, a visual difference like a title background color displays a distinction between the title and the background. However, for a color-blind user, that visual difference is no longer noticeable.

A rule of thumb is to use a maximum number of six colors in data visualizations. Any more and it becomes very difficult for users to see the differences. It is also recommended to test screenshots of visualizations for color-blindness using online tools such as Color Brewer. These tools contain pre-selected and/or pre-tested color sets that you can leverage.

Display Picture

Display picture, often abbreviated as DP, is what users see on the screen of their device, whether it’s a desktop monitor, smartphone, or tablet.

We’ve all been on our mobile device and tried to touch one thing but ended up selecting something completely different. This can be frustrating and time-consuming. That’s why, once a user reaches your BI dashboard, it’s important to remove any barriers to information consumption – including any click paths to dive deeper and explore other visualizations.

When designing display pictures for BI, make sure to incorporate spacing between filters, links, and visuals. The small screens on most mobile devices make it challenging for users to select objects that are right next to each other. Touch targets should be at least 48×48 DP, and they shouldn’t overlap. This ensures that users will be able to reliably and comfortably target them with their fingers.

Drill Down and Drill Through

Drill down and drill through make reporting powerful and useful.

Drill down and drill through are two extremely powerful features in business intelligence. They both give the user the ability to see data and information in more detail–although they do so in different fashions.

Drill down is a capability that takes the user from a more general view of the data to a more specific one at the click of a mouse. For example, a report that shows sales revenue by state can allow the user to select a state, click on it and see sales revenue by county or city within that state. It is called “drill down” because it is a feature that allows the user to go deeper into more specific layers of the data or information being analyzed.

Further levels of drill down can be set up within the report–practically as many as supported by the data. In our example, the drill-down can go from country to state to city to zip code to specific location of stores or individual sales reps. Typically, the look and feel of each level of the report is similar–what changes is the granularity of the data.

Instead of taking the user to a more granular level of the data, drill through takes him to a report that is relevant to the data being analyzed, also at the click of a mouse. For example, a tabular report that shows sales revenue by state can allow the user to click on it and reveal an analysis grid of the same data, or a heat map representing the data in visual form. It is called “drill through” because it is a feature that allows the user to pass from one report to another while still analyzing the same set of data.

Benefits of Drill Down and Drill Through

  • Gain instant knowledge of different depths of the data – Drill down gives the user a deeper insight of the data by letting him see what makes up the figures he’s analyzing. For example, in mere seconds, drill-down answers questions such as: of my National sales figure, which states are performing better? Which states are underperforming? And within each state, which territories are driving revenue?
  • See data from different points of view – Drill through allows users to analyze the same data through different reports, analyze it with different features and even display it through different visualization methods. This greatly enhances the users’ understanding of the data and of the reasons behind the figures.
  • Keep reporting load light and enhance reporting performance – By only presenting one layer of data at a time, features like drill down lighten the load on the server at query time and greatly enhance reporting performance–while offering great value to the end-user.

Dynamic Dashboards

Dynamic dashboards provide owners or authors with the ability to seamlessly update and add new BI content. This can include importing/exporting abilities, customized dashboard views via drag and drop, and integration with platforms or existing web technologies like Salesforce or Google Analytics.

A dashboard serves as a visual display of the most important information needed to achieve an objective. Today’s BI dashboards are leveraging richer content experiences through interactive elements like video, light boxes, overlays, and slicers to promote user engagement and help derive further insights. Capabilities like zooming, drill-down, and filtering are becoming more commonplace as well. Some dashboards even utilize social feeds to heighten brand loyalty and deliver real-time actions.

E

Elemental Development

Creating BI Applications through Pre-built Elements

Elemental development is an approach to report-building introduced and used by Synnect Analytics. It consists in employing pre-built elements to create BI applications such as reports, analytics, dashboards, visualization features, etc. The elements are pre-defined within the development environment. Instead of hard-coding, the developer creating the BI application selects the predefined elements from a menu, organizes them in a parent-child hierarchy, then specifies attributes pertaining to them.

Elemental development: pre-built elements are used and arranged in a parent-child hierarchy by developers to efficiently create feature-rich BI applications.

For example, in the creation of a data-table, the developer would select a data-table element, a report-body element, then column elements rather than building these from scratch.

In addition to being pre-built, elements are reusable. There are a host of benefits to elemental development, first of which is perhaps productivity and speed in building BI applications.

Embeddability

Embeddability/customization/integration: One of the major ways embedded analytics initiatives differ from standalone analytics projects is the need to integrate with the application environment. Embeddability refers to how tightly you integrate analytics into the overall user experience, the existing application security, and the application workflow.

Application providers typically want to offer a seamless user experience within the context of their existing application and brand. Achieving this involves a high level of customization – in other words, white-label and control the look and feel of the application to make it your own, and tailor the functionality so every user has access to the capabilities they need.

Embedded Analytics Defined

Embedded analytics is the integration of analytic content and capabilities within applications, such as business process applications (e.g., CRM, ERP, EHR/EMR) or portals (e.g., intranets or extranets). The goal is to help users work smarter by incorporating relevant data and analytics to solve high-value business problems and work more efficiently.

Embedded analytics differs from traditional business intelligence (BI), which focuses on extracting insight from data within the silo of analysis. While traditional BI has its place, the fact that BI applications and business process applications have entirely separate interfaces forces users to switch between multiple applications to derive insights and take action.

Instead, embedded analytics puts intelligence inside the applications people use every day. This improves the analytics experience and, in turn, makes users more productive by combining insight and action in the same application.

Said another way, business intelligence is a map that you utilize to plan your route before a long road trip. Embedded analytics is the GPS navigation inside your car that guides your path in real time.

Analytics may be embedded within business applications and workflows in several ways, each with varying levels of integration. The Embedded Analytics Maturity Model depicts these methods in four stages. The model begins with a standalone analytics application, where no embedding takes place, and ends with infused analytics, the deepest and most advanced form of embedding.

Deeper integration of analytics within applications is correlated to improving the user experience, increasing end user adoption, and differentiating the product.

End-User in BI

No matter how cutting-edge a BI application is, and no matter how well it is built and implemented, it is ultimately the end-user who has to make the most out of it. The business intelligence end-user can be defined as a decision-maker (of any level within the company), who does not necessarily possess IT skills and who uses business data and information from the BI solution to guide his actions.

The true test of the usability of a BI solution is with the nontechnical end-user.

The success that a BI solution will have in propelling the organization forward depends in large part on how it is received by end-users. Adoption makes or breaks a BI project. And adoption is, in turn, dependent on three factors: ease of use, usefulness and cost.

Ease of Use: The First Requirement of BI End-user Adoption

“Being adopted” is not the goal of business intelligence. The goal is to help end-users solve problems, eliminate inefficiency and achieve the company’s strategic goals. Adoption is merely a condition to this end, albeit a necessary condition. A well-implemented BI solution that is squarely and intelligently aligned with the company’s strategy has indeed the potential to make a tremendous impact–if adopted.

The first condition to adoption is ease of use. New technologies tend to make some new users anxious, especially if they are perceived as coming with a steep learning curve. If a newly-implemented BI solution is (or even comes across as) complex to learn and use, you can rest assured that many end-users will be reluctant to adopt it, and will instead fall back on what’s familiar.

Today, there is no reason why a business intelligence product should be hard to use. Especially with Web-based solutions, user interfaces should mirror the easy and intuitive navigability of the Internet. Important information and recommended actions should pop out to the end-user without requiring him to sift through pages of data or reconciling multiple tools. Likewise, analysis should be intuitive, letting the user filter, sort, drill down and drill through data at the click of a mouse without any technical knowledge required.

Usefulness: BI Must Solve Real Problems

Even if initially adopted, a BI solution will quickly lose its following within the organization if it does not provide real solutions for the end-users. “Does this make my job easier and does it make me successful at what I do?” is the question that needs to be answered in the affirmative through the business intelligence solution.

And how does a BI solution make personnel more productive and more successful? By making it easier for them to spot, understand and act upon critical situation while making as many routine tasks as possible automatic.

Cost: Avoid User Fees

The more expensive a good, the less of this good will be bought, goes a fundamental law of economics. Likewise, BI user fees–a throwback of the time when BI solutions were all desktop–discourage widespread end-user adoption. Yet, most BI vendors still charge by the user, forcing companies to either be conservative in estimating who gets access to BI or to waste license fees on employees who end up not using the solution.

The best licensing model is server-based, which allows companies to empower as many end-users as they need at no additional cost.

ERP

ERP, short for enterprise resource planning, is a transactional database that records a company’s information on daily business processes. According to the traditional business model, day-to-day data is also recorded in two other transactional databases: HRM (human resource management) and CRM (customer relationship management).

However, transactional databases are not considered optimal for business intelligence. This is for a variety of reasons, including the fact that data is not optimized for reporting and analysis. Additionally, querying directly against these databases may slow down the system and prevent the databases from recording transactions in real time.

ETL: Extract, Transform & Load

A Web-based ETL works like a Web service to help you integrate your data.

In business intelligence, an ETL tool extracts data from one or more data-sources, transforms it and cleanses it to be optimized for reporting and analysis, and loads it into a data store or data warehouse. ETL stands for extract, transform, and load.

There are many different models of ETL tools in today’s BI market, from complex, specialized products to light, Web-based solutions that work easily with multiple data sources.

Benefits of a Web-based ETL

A Web-based ETL gives you these unique benefits:

  • Fully Web-based Data Integration – With a Web-based ETL, you can not only seamlessly integrate your data, but also integrate the ETL with your other BI applications-regardless of vendor or brand. Use the ETL as a Web Service, launch ETL jobs from any standard-type processes and Web processes. Integrate the ETL into your business processes and workflows tied to triggers and alerts.
  • Unique Web Data Sources – Do more with a diverse set of data: a Web-based ETL gives you easy connections out of the box with Web Services and other Web-oriented data sources (e.g. SalesForce.com, Google Docs, RSS and ATOM feeds). Today’s most cutting-edge, Web-based ETL tools connect with relational databases and flat-file data sources.
  • Elemental Development Methodology – The same concepts you use to define logic in reports, templates, process files, etc., can be applied and even reused/shared in a Web-based ETL.
  • Optimization for BI and Reporting – Look for a Web-based ETL that is designed to work with data geared towards reporting, analysis and visualization. In particular, there are Web-based ETL tools that are created and marketed by companies specializing in BI; apart from truly optimizing data for reporting and analysis, this type of ETL will integrate seamlessly with your other BI applications.

How ETL manages and create a process around your data:

The extract step in an ETL job reads the data from one or more data sources. A good-quality Web-based ETL is “data source neutral” and is capable of reading data from almost any data source, including databases, flat files, spreadsheets, RSS/ATOM feeds and Web services.

The transform step in an ETL job manipulates the data gathered in the previous step. Here, data is combined, cleaned up, processed and optimized for reporting and analysis.

The load step in an ETL job takes the data collected and optimized and writes it back out to one or more destinations. In a good ETL, these can be almost any data source, including databases, flat files, spreadsheets, and Web services, RSS/ATOM feeds–just as is true of the extract step.

When Does Data Integration or ETL Become Necessary?

It is of course possible to report directly against your databases or data source(s). However, there is a point past which data volume, diversity of data sources and other important considerations make it desirable to have a data integration or ETL. If you are a data architect, developer or database administrator, here are some of the questions you need to ask yourself in this regard:

  • Is the volume of your data growing noticeably?
  • Is your company using an increasing number of data sources?
  • Do you need a convenient way to integrate your data across different applications?
  • Do you want to find a way to make your data more accurate and easier to understand?
  • Are you searching for an efficient way to manage or create a process around your data?

If you have answered any of these questions in the affirmative, you may need to look into acquiring a data integration or ETL tool.

Extended BI

Extended BI is the complex commercial network of customers, suppliers, and distribution partners who interact with BI solutions.

The days of the “self-contained” business are over. Since the value of a BI solution is tied to its ability to access relevant data wherever it resides, companies today are increasingly expanding their BI capabilities across the “extended enterprise” to collaborate with external stakeholders through innovative web portals and custom applications.

Internal data now represents only a small subset of information required to make strategic and operational decisions. Extended BI is empowering companies in any industry to collaborate with customers, suppliers, and distribution partners to grow revenue, enhance services, and achieve a competitive advantage.

Extensibility

Extensibility is the capability to extend and enhance a BI application or data model. Open-architecture solutions provide developers with unmatched flexibility to customize the look and feel of their applications and extend the functionality of the platform to meet custom requirements.

Extensible applications like Synnect Analytics employ custom themes, CSS, and JavaScript so customers can create unique user interfaces aligned with their brand. Third-party data visualizations and controls (including JQuery) may also be incorporated to meet specific requirements of a particular use case or industry.

Additionally, platforms with plug-in models can be extended to execute business logic, implement proprietary algorithms, or comply with desired data handling procedures.

Extranet

Extranet is an external-facing website containing information that is typically only accessible on an organization’s intranet (or internal network). Access to an extranet is controlled and limited only to authorized users such as partners, vendors, suppliers, or customers.

Extranets are a way for important stakeholders residing outside the organization to access internal information. Some BI applications allow for the placement of data visualization tools on a company’s extranet so that users can more easily consume key information.

F

Forecasting

Forecasting is the act of analyzing and mining data in order to predict what will happen in the future. Forecasting is typically accomplished using a BI application such as Synnect Analytics.

Forecasting can provide essential data to any business, no matter the industry. For example, in the case of a facility manager, such an application might provide real-time KPIs to give users the insight they need to make data-driven decisions – to track performance of every asset, accurately predict points of asset failure, and save maintenance costs to enhance equipment life.

Funnel Charts

Funnel charts are data visualizations best used for showing stages in a particular process (i.e., sales process) or identifying potential problem areas within an organization’s process.

Funnel charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

G

Gartner Magic Quadrant

Gartner Magic Quadrant is a series of market research reports published by Gartner, Inc. These reports have become a go-to BI industry resource for an objective perspective on the technology and service markets. The Magic Quadrant provides an annual qualitative analysis on the direction, maturity, and participants of technology industries.

In 2015, Gartner split off Advanced Analytics from the Business Intelligence and Analytics Platforms category, making the former its own magic quadrant – an indication that these capabilities are increasingly in demand.

Gateway to Analytics

Gateway to analytics is Stage 1 in the Embedded Analytics Maturity Model. This method of embedding provides access to the analytics in the process application via an embedded single sign-on. With single sign-on integration, the main application serves as the user’s “gateway” to the analytics application.

In the gateway model, the analytics application has integrated security with the core application. Users only need one set of login credentials, which are passed from the core application to the analytics application via single sign-on. Note that there are still two applications, but the access to analytics is embedded in the core application. It’s still a separate experience, however, because users have to switch back to the core application if they actually want to put insights to work.

There are a variety of instances when this model of embedding is appropriate. One example is when you have multiple applications and are creating a single analytic application that accesses data from one or more of these. With single sign-on implemented, users coming from any of the applications can access analytics. This can work well if the analytics application is in the cloud and serving up data from on-premise and cloud applications.

A second example is when you are offering the analytic application as a distinct commercial offering that customers need to purchase separately from your core application – and the “not-fully-integrated” application aligns with your commercial approach.

A third reason is that this model is simply an intermediate step in your development process and you intend to integrate analytics deeper in the future.

Gauges

Gauges are data visualizations best used to show a range. They are ideal for when you have an absolute floor and absolute ceiling value and you want to show where the value lies within that.

Gauges are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

While some users still prefer gauges, these attention-grabbing visualizations are notorious for taking up valuable space and providing limited information, since they present data on a single dimension. All they really tell you is whether something is on target, above target, or below target.

So what are the other alternatives to gauges? Line charts show trends over time; bar charts are great for comparisons; bullet charts are good for targets; and if preferred, you can opt for a mishmash of several types of combinations.

Geographic or GIS Maps

Geographic or GIS maps lets users visualize data in a spatial way.

Geographic maps let users make better decisions through geographic visualization and analysis of data. It allows report developers to feed in data and plot points on a map such as an actual street address or larger geographic area like zip codes, counties, states, countries, etc. Areas can also be color-coded not unlike a heat map, base on the underlying data to show how specific areas are performing (e.g. for a sales organization) or how they are affected by an event (e.g. appreciation of real estate).

Most data has a component that can be tied to a place: an address, postal code, global positioning system (GPS) location, census block, city, region, country, etc. Geographic mapping lets you visualize, analyze, create and manage data with a geographic component. And you can build compelling maps that help you visualize patterns, trends and exceptions in your data.

Benefits of Geographic Visualization

Using geographic mapping and visualization, users can visualize, explore and analyze data, revealing patterns, trends and relationships that are not readily apparent in other analysis features. Mapping can help you better answers questions such as:

  • Where are my customers?
  • What is the environmental impact of a new development?
  • Where should I put new stores or facilities?
  • How can I maximize my sales route or that of my reps?
  • Who is impacted in an emergency?
  • What are the highest traffic areas of a city?

The better geographic visualization tools offer the possibility to download boundary information from commonly available sources (e.g. the US Census) without having to build everything by hand. Also, they enable drill-down and drill-through to other reports, as well as the placing of markers on the map that are tied to relevant metrics.

Google BigQuery

Google BigQuery is a web service that enables interactive analysis of large datasets and works in conjunction with Google Storage. It’s an Infrastructure as a Service (IaaS) that may be used complementarily with MapReduce.

Google BigQuery solves the issue of querying massive datasets in a time-consuming and expensive way. It enables quick, SQL-like queries against append-only tablets, using the processing power of Google’s infrastructure. Users can control access to both the project and the data based on the business’s needs.

The program assists in helping protect data with its replicated storage strategy. All data is encrypted both in-flight and at rest and can be protected with strong access control lists (ACLs) that can be configured and controlled. Google BigQuery also enables users to access, save, and share complex datasets, and allows users to set limits and specify what permissions others have on the dataset.

Governed Data Discovery

Governed data is information that is managed and secured by a governing department, typically IT, before users may access it. Centralizing and managing the use and deployment of data (and its derivatives) ensures data integrity and security; in effect, users will only be working with trusted, credible data.

Today’s BI tools must balance the need for governance with the need for speed of data delivery. Data governance tools in BI applications usually include auditing, authorization, and authentication.

Governed data discovery is a way to address business users’ requirements for easy data delivery while at the same time satisfying IT-driven requirements for managing and securing that data.

Many organizations are unable to implement the data discovery tools their end-users demand because they don’t have the enterprise features necessary for achieving both data governance and speed of delivery. Governed data discovery delivers all of these capabilities within the same platform or application. In this way, s can access and analyze information that has been secured and determined credible.

Gradient

Gradient, or variant, is a non-uniform colored background that incorporates several gradually changing hues, usually from light to dark. Gradients are not recommended for use in data visualization backgrounds because they can distract from the data being displayed.

So that data is easy to see, it’s best to utilize a “flat user interface” that features two-dimensional illustrations and lessens the use of shadows, gradient textures, and three-dimensional treatments. This helps users to process the content quickly without distraction.

H

HeatMaps

Heat maps use cell size and color to display complex information in an intelligible way.

The heat map is one of the most useful and powerful data-analysis tools available in business intelligence. It is a visualization feature that presents multiple rows of data in a way that makes immediate sense by assigning different size and color to cells each representing a row. A color slider at the bottom or on the side of the heat map allows the end-user to easily spot the high and low outliers in the column represented by color.

For example, suppose you manage a sales force of 100 reps. On a data table, you have 100 rows, each representing one of your reps, plus a number of columns displaying (for instance) year to date dollars sold, year to date orders taken, etc.

If you were to represent this data on a conventional bar chart, the resulting visual would be so cluttered it would be practically useless. Instead, on a heat map, you can assign a cell to each rep, year to date dollars sold to cell size and year to date orders taken to cell color. The heat map will immediately sort your data by cell size, thereby allowing you (in our example) to see which rep has sold the most dollars year to date.

Color will immediately show you which reps have taken the most or the least orders; this is further facilitated by the color slider, which will intensify the colors on the high and low end of the spectrum and let the user see outliers.

What to Look for in Web-based Heat Maps

To be even more useful, a Web-based heat map feature within a BI solution should have the following benefits:

  • Easy integration with other applications such as dashboards
  • Easy setup even by end-users within an ad-hoc reporting environment
  • Flexibility to let the end-user choose which data columns to display through cell size and color
  • Drill-down and drill-through capabilities

HOLAP

HOLAP (Hybrid Online Analytical Processing) is a combination of ROLAP (Relational OLAP) and MOLAP (Multidimensional OLAP). HOLAP allows storing part of the data in a MOLAP store and another part of the data in a ROLAP store.

HOLAP can use varying combinations of ROLAP and OLAP technology. It typically stores data in both a relational database and a multidimensional database, depending on the preferred type of processing. The databases are used to store data in the most functional way possible. For heavy data processing, the data is more efficiently stored in a relational database, whereas multidimensional bases are used for speculative processing.

HP Vertica

HP Vertica is an analytic database management software company. Vertica is a columnar storage platform designed to handle large volumes of data, which enables very fast query performance in traditionally intensive scenarios. The product improves query performance over traditional database relational database systems, provides high-availability, and petabyte scalability on commodity enterprise servers.

HP Vertica is also part of the HP Haven platform, which is focused on analytics big data sources, including petabytes of structured and unstructured information. The integration of IDOL and HP Vertica within HP Haven allows users to connect to a variety of business, machine, and human data sources and perform both standard and predictive analysis.

HP Vertica’s design feature include:

  • Column-oriented storage organization
  • Standard SQL interface with many analytics capabilities built in
  • Compression to reduce storage costs
  • Support for standard programming interfaces
  • High performance and parallel data transfer
  • Ability to store machine learning models and use them for database scoring

HTML

HTML is an industry-standard programming language employed and supported by the majority of modern websites and web browsers. Along with JavaScript and CSS, HTML is used to create customized styling, branding, and visualizations in BI applications. Applications like Synnect Analytics, which produce industry-standard HTML5 output, are instantly accessible on Apple iOS, Google Android, and Windows mobile devices.

I

Iconography

Iconography is the use of images and graphics (icons) to denote common user actions such as “close” or “delete.” These simple icons can help users quickly discern which specific action to take.

Iconography is frequently used in BI applications to support a good user experience. For example, dashboards might be integrated with labeled icons in the navigation pane. They usually feature a small graphic or image followed by a pithy description so that users know exactly what they are looking at.

Information Delivery

Information delivery is a core functionality in any embedded analytics application. Improving how data is presented to business users is often the main reason software providers want to take on an embedded analytics project. In addition to satisfying users’ informational needs, the look and feel of these capabilities should align with the style of the embedding application.

Information delivery typically includes the following capabilities:

  • Dashboards and data visualizations: A range of visualizations, such as charts, gauges, heat maps, and geographic maps, enables users to quickly draw conclusions and monitor key performance indicators. These can be presented in the context of a single chart or in a collection of visualizations in a dashboard.
  • Reports: This is a tabular display of data, often with numerical figures and grouped within categories. Interactivity can include dropdowns and filters for users to view specific slices of data.
  • Mobile: Capabilities are made available to users on mobile devices, ensuring accurate visual display of information as well as compatibility with mobile device features such as touch input.
  • Scheduling and exports: Dashboards and reports can be scheduled for delivery, used in conjunction with thresholds/alerts, or exported to other formats for printing or offline access.

Infused Analytics

Infused analytics is the third and final stage in the Embedded Analytics Maturity Model and represents the deepest form of embedding. Here, analytics is infused as a natural part of the application. It is embedded within user workflows and becomes a core part of the overall user experience.

The infused analytics model is for application providers who want to position analytics as a core capability, by bringing together insight and action into the same context. It is correlated with greater realization of strategic benefits to the organization.

One way to infuse analytics is to provide analytic content “in the moment” or in existing application screens where users are making decisions and taking action. An example is to provide customer churn risk scores or purchase history in a customer service application so support representatives can offer personalized support.

Another infused analytics approach is for users to interact with analytic content that leads them to immediately transact or take action based on the analysis they are performing. For a sales manager who conducts geographic territory analysis, this approach enables him to immediately and efficiently re-assign territories from within the analytic interface, instead of having to jump out into the main application.

Of all the forms of embedding, the infused analytics model is experiencing the most growth. It makes analytics a core component of all software applications, so that it is impossible to tell the difference between an application and the embedded analytics content.

Just as B2C applications embed analytics as a natural part of the user experience and workflow, B2B applications will continue to move away from bolt-on approaches to more infused implementations. Users will simply be using their application – not two applications – to both inform and perform their work every day.

Inline Analytics

nline analytics, the most popular form of embedding, is Stage 2 in the Embedded Analytics Maturity Model. In this model, the analytics functionality appears inside the overall user interface of the application, creating a better user experience.

Inline analytics is often implemented as a reports tab or module. Another example would be a dashboard on the homepage of the application that users see directly upon logging in. With analytics integrated at the presentation tier of the application, it is ideal for the look and feel of the analytics functionality to match the user interface of the main application.

Application providers choose the inline analytics model when users demand easy and frequent access to analytics. Most third-party analytics applications can be embedded using this approach, so there are a lot of options available to “bolt on” to an application. Users are also very comfortable with the reports module approach (as used by Salesforce.com, for example), so it’s not surprising that inline analytics is the most common model for embedded analytics.

Interactive Data Viewer

The IDV gives end-users the opportunity to perform “quick and dirty” analysis of business data.

The Interactive Data Viewer, IDV for short, is a robust data-analysis feature that enables business end-users to do two main things:

  • Perform instant calculations and sorting of data in a table
  • Immediately visualize the results graphically

Although Web-based, the IDV offers the interactivity of a dedicated desktop application. There are three powerful controls that the user has at his fingertips.

First, there is the ability to aggregate or isolate the rows and columns to be analyzed by simply control-clicking on the entries.

Then, the user can choose which calculation to perform by selecting from a drop-down menu. This menu contains the most useful types of business and statistical calculations, already preset and without any need for constructing formulas.

Data may be viewed in table, graph or table-and-graph format, to tailor them to the needs of the user and his audience.

Next, the user has the flexibility to choose the type of graph in which he’d like to see the results, so that the information presented can have the most visual impact. Chart types include line, bar and pie-all with a three-dimensional option.

The Benefits of the Interactive Data Viewer

Simple to use, flexible and interactive, the IDV is the perfect tool to answer business questions accurately and quickly.

  • Quickly select rows and columns you want to see from the row- and column-selector, eliminating clutter
  • See the results of your analysis displayed in graphic format in real time
  • Choose from a high number of pre-set calculations to compute average, median, standard deviation and other common computations on the fly
  • Easily make the IDV part of a dashboard or other report
  • Point-and click for advanced calculations–users can choose data-sets pertinent to the business at hand and perform one-click, pre-defined calculations to such data-sets.

For more complex, personalized analysis where, for example, custom columns and filters need to be created, there are other features like the Analysis Grid that yield the necessary power and flexibility.

Intranet

Intranet is a private internal network that may only be accessed by staff within an organization. Some BI applications allow for the placement of data visualization tools on a company’s intranet so that users can more easily consume key information.

J

Javascript

JavaScript is a programming language employed and supported by the majority of modern websites and web browsers. Along with HTML and CSS, JavaScript is used to create customized styling, branding, and visualizations in BI applications.

K

Key Perfomance Inidactors

Key performance indicators or KPIs are exactly what the name suggests. They are visual indicators in the form of color-coded shapes that are tied to a pre-defined, critical threshold. When the threshold is crossed, the KPI’s function is to alert key personnel so that they can take the necessary action.

Here’s an example. A bestselling product needs to be always near optimal inventory level. The inventory manager sets 1,000 units in stock as the critical threshold: if the item drops below 1,000 units in stock, the manager needs to be alerted and the item immediately reordered. By setting up a KPI as a way to ensure this, the inventory manager will see a green dot next to the item if inventory levels are OK; as soon as they drop below 1,000 (the pre-defined threshold), the dot will turn red, and the inventory manager will immediately know he needs to place a reorder.

The same mechanism is used by today’s companies for monitoring lead levels, sales target, revenue, or anything critical that can be quantified and to which a threshold between good and poor, noncritical and critical, etc., can be set.

Advantages of KPIs

The KPI is perhaps the single most intuitive visualization feature available in BI. For example: green=good; red=action needed! The more a critical piece of information jumps out to the relevant personnel, the better the chance to take corrective action before the situation develops into a real problem.

With critical information, users should not have to sift through pages of data, essentially analyzing it every time a decision about something important needs to be made. The BI solution should have a mechanism in place through which the piece of information is served up to the user in a meaningful way–as it is in the color-coded KPI. Green means something, as well as red, that the user will immediately know and recognize without having to ask himself.

Here are two caveats about KPIs:

  1. Do not turn too many items into KPIs. KPIs should be reserved for the truly critical items–for instance, the performance of your main product lines if you are a rep, not the performance of every individual SKU you sell. The risk is to water down the effectiveness of the KPI through visual and mental inflation–if you have too many items “crying wolf” at you, you are less likely to pay attention to the ones that are of high importance.
  2. If you can, make KPIs actionable. For instance, in our example of the inventory manager monitoring in-stock level of the bestselling SKU, when the KPI turns red the interface should offer the option to reorder without leaving the application, thereby saving time and making the process more efficient.

L

LDAP

LDAP, or lightweight directory access protocol, is an Internet protocol BI applications use to look up information from a server. Some of these applications support single sign-on integration for any security framework or application, including LDAP.

M

Managed Reporting

Managed reporting lets developers create powerful, feature-rich reports.

Managed reporting is a model of business intelligence (BI) in which reports are built and distributed by report developers. In other words, it is technical users with an IT background, knowledgeable about subjects like SQL queries and CSV language, who develop the reports.

Managed reporting is therefore “managed” in the sense that end-users receive reports that are built and distributed from the top down by technical users, who “manage” the process, ensure that users have what they need and correct any bugs or flaws that the report may have. It stands in contrast to ad-hoc reporting, the model by which nontechnical end-users prepare the reports.

The Goal of Managed Reporting

Managed reporting’s goal should be to leverage BI technology as efficiently as possible to ensure that end-users achieve the company’s strategic and tactical goals. Here too, like elsewhere, the emphasis is on the company’s strategic and tactical goals and the facilitation of their attainment. Like all things in BI, technology should be the means to a clearly-defined end, not the end itself.

For example, let’s say a manufacturing company is implementing BI and they choose managed reporting as the principal model of their solution. Their strategic goal is to become a leader in their vertical–which entails managing the supply pipeline, implementing total quality control in manufacturing and on-time delivery, minimizing inefficiencies and maximizing sales.

What to look for in a good managed reporting solution

A managed reporting solution, in this case, would be focused on the achievement of these goals by providing end-users with an easy and efficient way to perform tasks such as:

  • Ensuring that manufacturing components are ordered on time, for example, through an automated lead-time and in-stock level report with actionable reordering alerts for critical items
  • Identifying inefficiencies in manufacturing and correcting them, through features like dashboards, visualization tools, gauges and actionable key performance indicators (KPIs)
  • Spotting untapped potential in product lines, customer segments or territories, with tools like interactive heat and geographic maps
  • Maximizing sales reps’ face time with customers, by making reporting automated, as well as by mashing up customer data with geographic maps to plot most profitable sales routes

Although these goals can in theory be also achieved through ad-hoc reporting, the power, the features and the developer know-how make them an ideal fit for managed reporting.

MapR

MapR is a big data platform that integrates Hadoop, an open-source software framework, and Spark, an open-source data processing engine. The platform includes global event streaming, real-time database capabilities, and enterprise storage.

Mashups

“Mashup” started as a popular word in the world of club music. For several years, people have been combining rock music clips with rap clips into mashups for dance clubs and parties. Really fancy mashups merge several different types of music-rock, rap, techno-all into one dance tune. A mashup is also a kind of video. Just like with a music mashup, a video mashup uses images from different sources, merging them together and superimposing them to one-another to form a new, dynamic and often surrealistic effect.

Mashups have moved beyond music and video and into the world of the Web and business intelligence. In the Web world, mashups have the same basic idea as with dance music and video-combining otherwise discrete components into a single aggregate. BI mashups fall into different categories, such as overlay mashups, widgets and dashboards.

Overlay mashups merge data from more than one source into a single user-controlled feature; a typical example is combining business metrics (e.g. sales by rep) with Google Maps. Widgets are small HTML chunks of a third party site that are embedded and executed as an element of another-for example, a chart from a BI site or a You Tube video may be embedded and displayed in a news site or blog. A dashboard, instead, is a single Web page containing various panels each displaying a different object that may be a graph or chart, a Web site, RSS feeds and so on.

In the world of Web 2.0-and consequently BI 2.0-mashups are a useful, interactive tool in the hands of the business end user. First of all, they provide a way to integrate different data and visualize the results in a coherent and persuasive manner. Then, when enhanced by technologies like Flash and AJAX, and when featuring drill-down and drill-through capability, they become some of the most versatile and visually-powerful BI instruments in the decision-maker’s arsenal.

Let us now explore overlay mashups, widgets and dashboards separately, taking a look at what they are, what technologies they employ and what benefit they give the business end user.

Overlay Mashups

Overlay mashups are the combination of two or more data sources for use in a single Web-based feature or application.

A common example of the overlay mashup is the placement of charts, graphs and other business-relevant data onto a map. Google, for example, publishes an open API that allows software vendors and developers to mash up geographically-relevant information with Google Maps. Your BI application may show pins on a Web-based map to represent customer locations. It may color those pins to reflect customer segments. When you mouse over a pin, interactive charts or graphs may pop-up dynamically showing recent sales activity and customer service information for that location.

Map mashups are the first, and perhaps most intuitive example of overlay mashups. Maps are so familiar that overlaying a drillable chart or graph onto a map is a simple extension of the paradigm. The next steps beyond maps will probably also require an intuitive context. For example, an architectural extension of the map concept might include a mashup of live security video feeds and entry/exit statistics overlaid onto premises maps and floor plans. The floor plan context may also be used to support shop-floor materiel movement, productivity and quality charts and graphs.

All of these examples involve spatial contexts such as maps and floor plans. As business intelligence users become more accustomed to mashups, the model may eventually be extended to show inventory and lead-times overlaid on a non-geographic supply chain model. Productivity, quality metrics and other worker performance management metrics could be overlaid on work-flow and process maps.

There are two essential ingredients for overlay mashups to work. The first is an abundance of open APIs, such as the one published for Google Maps. In addition to the open APIs, the BI tools themselves need to be modular in design and Web-based. XML and HTML are the natural vehicles to make this level of integration happen.

Widgets

Widgets are small HTML chunks from a third-party site that are embedded and executed on a Web page. An everyday example of this would be a You Tube video that the user can watch by clicking on it without exiting the page in which the video is embedded.

Widgets fall into the category of mashups because information from two different sources or servers-the one hosting the site, plus the third party source from which the widget comes-is combined.

The role of widgets is primarily to add variety and dynamism to a site. For example, a financial-market blogger may embed a NASDAQ real-time quote chart on his Web page to add a useful and dynamic element to it. Or, a sports news Web site may have a streaming video embedded as a widget to show the week’s best touch-down.

In BI, widgets have the same function than in any other Internet environment, which is to give users access to useful content from a third party site without opening another Web browser. In the past, access to third-party Web content of this type was primarily effected through links, which involved leaving the page being currently viewed or having to open a new window.

From a technical standpoint, widgets are fairly easy to execute. All that is required is access to the server of the host page and that of the site producing the widget. With their easy deployment and Web 2.0-like dynamic feel, it is not illogical to foresee that this dynamic and easily-executable form of mashups will keep growing in the BI world.

Dashboards as Mashups

IT or corporate dashboards can also be viewed as mashups because they represent a means of bringing together different visual representations into a unified picture for an end user. Dashboards are becoming increasingly common in business intelligence applications, but we are still in the early stages of this trend. The dashboards we see today include interactive and drillable charts grouped together into a single window. The most flexible dashboards are built with AJAX (Asynchronous JavaScript and XML) which allows individual objects (charts, graphs, etc) to be refreshed and repositioned without refreshing the entire browser window.

Currently, many dashboards represent data from a single database or data warehouse. While the dashboard itself may contain several individual charts and graphs, the underlying data most often comes from a single data source. For example, a country sales manager might use his dashboard to show the current sales pipeline, quarterly projections, regional sales performance and sales by product. All of these individual report objects are populated by the company’s data warehouse.

With the more sophisticated BI products, however, data can be also brought together on a dashboard from different data sources – for example different databases, plus Web services, flat files, RSS feeds, etc.

Benefits of Mashups

In the world of Web 2.0 and BI 2.0, mashups that are interactive and dynamic are appreciated by business end users. Interactivity can come in the form of drill-down or drill-through capability, as well as in the ability to use the Web-based feature or solution as a dedicated desktop application. And, a mashup can be made dynamic through the use of Flash technology, for example, by enabling key performance indicators, graphs or charts to update dynamically when the page is opened or refreshed.

Business intelligence features that have an interactive feel and intuitive, browser-based navigation help spread BI to a wider and deeper range of business end users. This model, called pervasive BI, is defined by accessibility of reporting and analysis tools to non-technical end users across organizations. Although mashups and dashboards are not the only features conducive to pervasive BI, they are a great example of how by integrating familiar Web 2.0 capabilities into a BI solution help more end users have access to the data and information they need

Metadata

Metadata is data that summarizes information about other data. In traditional BI, metadata is very structured; anytime the fields change, the metadata has to be updated as well.

However, today’s BI applications, such as Synnect Analytics, are able to work with user-defined fields, eliminating the need for a metadata layer. Data definitions may be changed on the fly with the use of what’s called a “repeat element.” This element design allows for a more agile, flexible approach to BI report authoring.

Metadata Management

Metadata management is the harvesting and governance of metadata across an enterprise. Metadata management can help businesses understand how report figures are calculated or how changes impact the data.

Mobile BI

Mobile BI refers to the recent trend of business users accessing their data and dashboards on mobile and tablet devices. This change is forcing designers to re-evaluate the user experience (UX) and replicate the same relevant content usually seen solely on desktop devices on mobile.

It’s important to know the differences in desktop and mobile application development. Since the mobile market is constantly changing in terms of devices, operating systems, and support for different types of content, the QA process can be tedious and time-consuming. Developing apps that comply with web standards like XML and HTML5 and work across multiple user platforms can make or break a product.

You’ll also want to utilize user interface elements like responsive design so visualizations look great regardless of mobile device, as well as functionality native to those devices – like clicking on an address to search for the location in Google Maps, or clicking on a phone number to start dialing. Even the inputs need to be mobile-friendly so users can easily access, type, and make changes to visualizations on a smaller screen.

Finally, it’s important to differentiate the goals of your desktop applications versus your mobile applications. If the desktop application is meant to show summary reports with tons of drill-downs for manager types, while the mobile application is meant to enable workers in the field to create orders and fire off automated emails and alerts, then the functionality as well as look and feel will be very different.

Look for a BI tool that supports both types of applications so you can reuse as much content as possible without sacrificing functionality or ease of use. With the right features, mobile UX can be fast, agile, and simple. Features like responsive design, scrolling, and “swipe” will surely propel this trend forward for the next several years.

MOLAP

MOLAP (Multidimensional Online Analytical Processing) is an online analytical processing (OLAP) that indexes directly into a multidimensional database. It processes data that is already stored in a multidimensional collection in which all possible combinations of data are reflected and each of which can be accessed directly.

Advantages of MOLAP include:

  • Effective data extraction
  • Fast query performance due to optimized storage, multidimensional indexing, and caching
  • Compact for low dimension data sets
  • Smaller sized data collection due to its compression techniques
  • Natural indexing due to array models
  • Automated computation of higher level aggregates of data

MongoDB

MongoDB is a leading NoSQL database that empowers businesses to be more agile and scalable with their data. Application developers particularly like the flexible document-oriented store, which allows for the direct storage and retrieval of structured, semi-structured, and polymorphic data without the need for complex object relational mapping.

NoSQL solutions are one of the newer and most popular approaches for handling large-scale and flexible data requirements.

When used in conjunction with a business analytics platform like Synnect Analytics, MongoDB enables report authors and BI application developers to:

  • Extract value from big data and all their data sources through data access and mashup capabilities
  • Report and create visuals on data stored in MongoDB and other hierarchical data sources through data flattening
  • Continue to access MongoDB capabilities, including Find and Run commands, Aggregation Framework, Map Reduce, and writebacks

Multi-Tenancy

Multi-tenancy is a software architecture in which a single application serves multiple customers (or tenants). Multi-tenancy gives companies the ability to create an application just once and deploy it to many customers.

BI applications that support multi-tenancy, like Synnect Analytics, provide fine-grained user access control across every layer of the application − reports, charts, application functions, and data columns and rows. Single sign-on integration is supported for any security framework or application, including LDAP, Windows Active Directory, and custom databases that store user profiles.

Applications like these do not necessarily store user information, thus simplifying security integration and removing the overhead of continuously synchronizing user profiles. Software as a Service providers can easily configure security regardless of whether their customer data is stored in unique databases or in multi-tenant/commingled data sources.

N

Native

Native refers to a software application or data format that is designed to run on a specific system (i.e., software is native to a BI platform if it was designed to run on that platform). This attribute is essential to successful mobile BI. Embedding analytics applications within native mobile apps enables them to offer the same functionality and interactivity as the browser-based content.

NoSQL

NoSQL is a type of big data repository for handling large-scale and flexible data requirements. NoSQL-based solutions are one of the newer and most popular approaches for storing and retrieving structured, semi-structured, and polymorphic data without the need for complex object relational mapping. The leading NoSQL database in the marketplace today is MongoDB.

O

ODBC

ODBC, short for open database connectivity, is a high-performance application programming interface (API) designed for accessing relational data stores. JDBC (Java database connectivity) is an API designed specifically for the programming language Java.

OBDC/JDBC is used to acquire data from big data sources, independent of database and operating systems. Along with other data connectors like Hubspot and Eloqua, OBDC/JDBC empowers users to analyze and visualize their data from a variety of disparate sources.

OEMs

OEMs are manufacturers who resell other companies’ products under their own names and branding. In the BI arena, OEMs are typically Software as a Service (SaaS) providers who embed business analytics applications into their software. These embedded applications typically come with custom licensing and a suite of technical advisory services. OEMs will license BI products in order to sell more software, increase revenue, and reduce costs.

OLAP

OLAP is an acronym that stands for On Line Analytical Processing–a somewhat fancy term for another fancy term, multidimensional analysis. OLAP is the process of analyzing data from different dimensions, which is why the objects to be analyzed are called OLAP cubes.

What this means, in simple terms, is prioritizing the way data is shown by a given column. For example, if you have a table of data about sales, you can analyze it by product type (a dimension), by demographic (another dimension), by geographic region (another dimension), etc. The data you see can even be always the same, but it is prioritized by whatever column you place first–which we call a dimension. In practice, this requires the data dimensions to be pre-calculated.

Multidimensional databases are usually quite complex architectures. In these databases, intersections of relevant data become more apparent so that the data is easier to group, summarize and analyze. For example, OLAP allows an analyst to answer questions like “how many computers have been sold in Canada this year?” and “of those sold in Canada, how many were sold to people over 50?”

Now, there are “cube viewers” that can be accessed via Web services–such as Microsoft Analysis Services–that make the process a lot lighter and easier.

The Benefits of Web-based OLAP Analysis

  • Powerful, complete analysis that answers “why” and “how” questions – Multidimensional analysis allows for a great degree of understanding of the reasons behind the data. This is what is called “slicing and dicing” of data–namely looking at it at different levels of detail as well as from different perspectives to understand it better.
  • Interactive access to OLAP data – Users can slice and dice cube data all on the fly by various dimensions, measures and filters all within their Web browsers.
  • Zero-footprint access to OLAP analysis – With Web-based access, IT personnel can set up connection to OLAP cubes in minutes. When users access the OLAP reporting interface, no client software has to be installed. They easily perform on-demand analysis in the Web-browser.

OLTP

OLTP (Online Transaction Processing) is a data modeling approach that is used to facilitate and control everyday business applications, as well as support high transaction applications.

OLTP systems focus on fast query processing, while maintaining data integrity in multiple access environments and effectiveness, measured by number of transactions per second.

Most of the applications seen today are OLTP based. The primary system features consist of instant client feedback and high individual transaction volume. An important attribute of an OLTP system is its ability to maintain concurrency. To avoid single points of failure, OLTP systems are often decentralized.

Open Architecture

Open architecture is an environment that provides software developers with unmatched flexibility to customize the look and feel of their applications and extend the functionality of the platform to meet custom requirements. Custom themes, CSS, and JavaScript can be employed to create unique user interfaces aligned with any brand. Open architecture platforms can also incorporate third-party data visualizations and controls (including JQuery) as well as plug-in models.

Open architecture allows for the easy deployment and scaling of BI implementations. Unlike traditional BI systems that require proprietary servers or a heavy metadata layer with ETL processing, these implementations reduce both the cost and time to deliver value to end-users. In addition, the lightweight web architecture enables organizations to utilize widely known techniques to deploy and scale implementations, minimizing costs and investment in proprietary technologies.

Applications can be deployed to any major infrastructure, whether on-premise, hosted at a data center, or in a cloud-based infrastructure such as Amazon Web Services or Microsoft Azure.

Open Source

Open source is one of the major models for business intelligence software. A software vendor takes a product that was created in the open-source community and makes it their own so that they can market it.

Since by definition an open-source product cannot be sold, commercial open-source vendors make money through services, support, and any add-ons they have built themselves. So, although they are not selling the core product, they are still selling something.

Open-source BI has several appealing attributes, but it can also be quite risky. One thing that makes a pure open-source model attractive is the flexibility it offers for customization – although this comes with a substantial flip side. The buyer has access to the source code, so his team can add, modify, or delete anything they want. However, as soon as they do this, they’re deviating from the source.

At that point, they either need to become active participants in the community (submitting their changes for everyone else to use), or they have to move further away from the core and hope not to run into any major landmines within the source that they need to ultimately fix by themselves.

Another initial lure of working with a commercial open-source vendor is the low cost of entry – since the product is ostensibly free. However, the services and support that commercial open-source vendors provide is essential to helping the client get started. Once the client goes down this services and support road, they face the challenges described above (they’ve deviated from the source), and now they’re even more dependent on the vendor for services, support, and add-ons.

Another negative is the fact that there is no real accountability if something goes wrong. Who do you turn to if there is a major problem with the product? Can you go back to the community to get the bug fixed? Can you go back to the vendor?

This sort of bottleneck actually happens quite frequently in the commercial open-source market: the same bug exists within the commercial open-source as does the main open-source project. The customer can’t get their situation resolved by the vendor, because the vendor is waiting for the community to fix the problem – with the sense of urgency of a more or less voluntary community.

Therefore, the fact that services and add-ons still have to be paid, plus the uncertainty of how the project will be supported in case something goes wrong, often makes open-source BI risky.

P

Parameterization

Parameterization is a best practice for designing BI dashboards, including mobile BI. Filters and queries allow users to select from options – or parameters – to get different views of data very quickly.

While there may be many different filter options for the desktop, filters on mobile should be kept to a minimum. Add a search option on a dropdown list, for example, to allow users to find the values they need. Providing the right set of filters will help your users make better use of their mobile reports and dashboards.

Parameters

Parameters are the simplest method for posing queries in BI applications. Most systems will present a list of parameters for users to choose from when asking questions of their data. Through guided menus, business users can easily request specific information.

While selecting from predetermined parameters is probably the easiest way to query data, it is also the least flexible. More advanced users will instead want to input their own custom queries.

Performance Dashboard

Performance dashboards are common management tools used to gauge performance and enable business people to measure, monitor, and manage the key activities and processes needed to achieve business goals.

They can be designed and developed to direct a wide range of objectives, from monitoring the usability of a global organization’s business strategy to keeping a check on a department’s ability to achieve targets.

According to Wayne Eckerson, a performance dashboard provides three main sets of functionality:

  • Monitor critical business processes and activities using business performance metrics that alert users to potential problems.
  • Analyze the root cause of problems by digging into relevant and timely information.
  • Manage people and processes in order to improve decisions and optimize business performance.

Pie Charts

Pie charts are data visualizations best used for when you want to show a one-to-many comparison between the different data sets being represented. They can illustrate relative magnitudes, frequencies, or percentages.

Pie charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Alternative visual styles to traditional pie charts include an exploded pie wedge chart, for when you want to emphasize important data, and a donut pie chart (a pie chart with a hole in the middle), if you want to insert a design element in the center to support the information being displayed.

Plug-Ins

Plug-ins, such as .NET and Java, are used to extend a BI application’s capabilities to execute business logic, implement proprietary algorithms, or comply with desired data handling procedures. Insights from multiple databases and application sources can be displayed together by blending data into a single visualization, dashboard, or report.

Polar Charts

Polar charts are data visualizations best used for displaying multivariate observations with an arbitrary number of variables in the form of a two-dimensional chart. Polar charts are also known as radar charts, web charts, spider charts, and star charts, among many others.

Polar charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Portals

Portals are a way to display data from disparate systems in a single place where anyone can access the data and applications they need to do their work. These portals connect to data sources and applications in order to present and distribute information to any audience.

Portals in BI applications, such as Synnect Analytics, may also be integrated with any security model to adopt the user rights and roles already established. Some platforms offer this secure portal access on mobile devices, as well.

Predictive Analytics

Predictive Analytics is a business intelligence technology that produces a predictive score that informs actions that customers should take. It’s an area of data mining that is related to the overall prediction of future probabilities and trends. For example, an insurance company is likely to take into account potential driving safety factors, such as age, gender, and driving record when distributing automobile insurance policies. Predictive analytics helps organizations identify the most efficient marketing campaigns and website behavior to increase customer responses, conversions and clicks, and to decrease interruptions.

The three types of models in predictive analytics include:

  1. Predictive model: models of the relation between the specific performance of a unit in a sample and one or more known features of the unit
  2. Descriptive model: quantifies relationships in data in a manner that is used to classify customers or prospective customers into groups
  3. Decision model: describes relationship between all the elements of a decision, including the known data, the decision, and the results, to predict the results of decisions involving many variables

Premise-Based

Premise-based software is located at a company’s physical location (e.g., on premise) and controlled by that company. The premise-based model, however, has largely fallen by the wayside in favor of virtual models, where software resides in the cloud.

For on-premise software, revenue happens when you ship the box. To illustrate the point, let’s pretend your software is actually a box of cereal at the grocery store. With cereal, the important thing is how many boxes you shipped to the store; you don’t care too much whether people are actually eating it or how they are eating it. You only really care about the metrics up until the boxes are sitting on the shelf at the store.

Consumers are able to pick out different foods and ingredients that they assemble at home to make their own meals. Similarly, with on-premise software, either the end consumer or the internal IT organization is responsible for picking the right things, putting them together, maintaining them with updates and patches, and deciding when they’re available for the business to use.

It’s the responsibility of the consumer to create a successful solution here, not the software manufacturer – just as it’s up to the individual to whip up a tasty meal at home, not General Mills.

Pyramid Charts

Pyramid charts are data visualizations best used for showing comparisons of data, using the thickness of their layers to denote relative values.

Pyramid charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Q

Queries

Queries are requests for information from a database. Today’s business intelligence systems enable business users to easily connect to and query data sources provisioned for them by their technology teams.

Menus typically guide users through queries by presenting a list of pre-determined parameters. More advanced BI systems allow the user to define (customize) the query or require the query to be written in special query language.

R

Relational databases

Relational databases are data sources that display information in tables (also known as relations). BI applications are able to retrieve information from these databases in order to prepare and provide that data for analysis. Examples of relational databases include SQL Server, Oracle, MySQL, PostgreSQL, and DB2.

Relational databases are reliable, easy to create and use, and support data integrity. These benefits have made them a mainstay in the BI industry. However, as big data increases in volume and becomes more unstructured, relational databases are giving way to other big data solutions like NoSQL.

NoSQL-based solutions are one of the newer and most popular approaches for storing and retrieving structured, semi-structured, and polymorphic data without the need for complex object relational mapping. The leading NoSQL database in the marketplace today is MongoDB.

Report Building

Elemental development means high productivity for report developers.

To enable end-users to see, understand and act upon their data, reports have to be built first. Traditionally, report-building was slow and cumbersome: upper management in the sales, finance, marketing, HR or operation departments tapped IT with one or more report requests; IT interpreted the requests and weeks, sometimes months later the reports were delivered, more or less in line with the main points of the original requests.

Clearly, this old model of report building was a bottleneck. With the advent of modern BI solutions, this model is rapidly becoming obsolete. Through managed and (even more so) ad-hoc reporting, dynamic Web-based BI solutions are used to place more and more reporting power and flexibility in the hands of the end-user.

If we momentarily leave aside ad-hoc reporting, which is the end of the spectrum in which end users have full control of their reports, let’s say a few words on the most innovative report-building model available in today’s BI.

The Advantages of Elemental Development in BI

Synnect Analytics has developed a unique and innovative paradigm for report application development–a concept that we have termed Elemental Development (ED).

Implementing an ED-centered environment is based on an extremely high level, re-usable XML-based language that fits specific business intelligence needs. And, this XML-based language can be thought of as a dictionary for application development. Traditional high-level languages such as Visual Basic or C focus on providing a flexible and robust framework for creating applications of almost any type. An ED language, however, is formed for a specific type, or class, of application. Once built, it can be reused to rapidly develop similar applications with a huge savings in development and maintenance time and costs.

This is because ED standardizes and simplifies the development process.

Benefits for Report Developers

The ED approach offers the following valuable benefits for developers:

I. Streamlined Development –This approach takes advantage of the self-documenting, intuitive and descriptive nature of prebuilt elements. For example, if you see an Email element in a report, you quickly know that the report form sends email. If you see a User Role element, you know that role-based security is implemented, and you can determine how that security is set up by looking at the attributes that describe the User Role element. Also, less actual coding is required. Using wizards and drag-and-drop, developers can easily build complex reports (for example, with drill-down, data grouping and filtering) without having to build complex SQL queries, subroutine calls or advanced command constructs.

II. Increased Productivity and Faster Deployment –Accomplish report development in a matter of hours instead of the weeks and months that may be required of other development tools due to:

  • Ease of use and reusability of elements
  • Logical and hierarchical layout of elements, which makes it easy to understand and manage layout and functionality of larger reports
  • Ability to change report layouts or functionality ‘on the fly’–just by modifying elements and attributes in the report definition
  • Based on well-known, non-proprietary open technologies and standards like XML,.NET, SOAP, Web Services, and so on

III. Scalability –This approach leverages multi-tier development and multi-tier deployment inherent in Web-based applications, which is by nature more scalable than license-based or traditional development.

Report Developer in BI

Elemental development makes building reports a high-productivity process.

Although in most companies the report developer is part of the IT department, his strategic thinking impacts the whole company. We can look at the report developer’s role as that of a giver of knowledge. Sure, in many companies, end-users create their own reports, especially in today’s time when technical skills are more commonly widespread even among business users. But a professional developer’s skills–when used effectively–can be the primer to competitive success.

The main task of the report developer is, as the name suggests, to prepare reports. But to do so effectively, he needs to take the following into close consideration:

  • The company’s strategic goals
  • The roles of the various departments within the company’s strategy
  • The pieces of information that are vital to each department and to the company as a whole
  • The actions that should be made easier, quicker and more efficient to positively impact the company

Developing a Report the Smart Way

The better report developers have a knack for zeroing in on the problem that reports are meant to solve. In other words, they start with the most important goals that the report’s user have to meet–again, tied to strategy–and they back out a solution that meets those goals as efficiently as possible. In the continuum of reporting, analyzing and action-taking, good report developers focus their applications around the last point.

For instance, a good inventory report for a buyer may feature prominent gauges showing in-stock level of the most profitable products, with automatic reordering built in or at least with an option for a one-click reorder action. Such report could also feature a list of the top-25 products that are overstocked and/or the top-25 that are low in stock.

Part of the report can also consist in automatic business alerts to be sent to the inventory manager in case in-stock levels of critical products fall below a predefined threshold, so that the manager can place an immediate reorder without leaving the application.

With this approach, important information is channeled proactively to the problem-solver; with less efficient reporting, the problem-solver has to sift through data of various levels of importance before (hopefully) finding the critical items needing immediate attention.

So, although we call this “report developing,” we can easily see that the more useful the report, the more it focuses on analysis and action-taking for critical items–rather than on “reporting” literally defined.

The Best Tools of the Report Developer

Technology is the means to an end, and the smart report developer understands this. For a developer to become a company-wide hero, he has to make the life of people using his reports as easy as possible, and their jobs as efficient as possible. So, it doesn’t matter how visually appealing or technically-powerful the report is; what matters is what problems it solves and how efficiently it does so.

Still, robust web-based reporting and analysis features are a great weapon in the developer’s arsenal, since they expand the gamut of what he can offer end-users. This is especially true if the features are easy to build and do not require much manual coding. This way, the developer can focus on the true goal of the report–what problems it is meant to solve–rather than on the intricacies of developing the report manually.

A good reporting solution should enable a developer to do the following:

  • Offer end-users an easy and intuitive way to interact with their data, to spot the pieces of information most critical to their jobs and to act on them immediately. Features like dashboards, heat and geographic maps with drill-down and drill-through capability, actionable key performance indicators (KPIs), automatic report scheduling and business alerts go a long way to achieving this.
  • Set up a variety of reports for different functions within the organization easily and quickly, and with as little manual coding as possible. The best BI solutions have vast libraries of pre-built elements that cover most of today’s cutting-edge reporting and analysis features. In addition, these solutions offer developers the possibility to further extend the report’s features and functionality through plug-ins.
  • Maximize reporting performance so users can see and use their reports in a flash instead of wasting costly time waiting for the report to load. There are reporting-performance factors that are outside of the developer’s control–such as hardware and network bandwidth. But especially within a Web-based BI application model, factors such as shared elements, dynamic filtering and type of data-processing engine allow the developer’s decisions to have an impact on report performance.

Reportnig in BI

Reporting means collecting and presenting data so that it can be analyzed.

When we talk about reporting in business intelligence (BI), we are talking about two things. One is reporting strictly defined. The other is “reporting” taken in a more general meaning.

In the first case, reporting is the art of collecting data from various data sources and presenting it to end-users in a way that is understandable and ready to be analyzed. In the second sense, reporting means presenting data and information, so it also includes analysis–in other words, allowing end-users to both see and understand the data, as well as act on it.

Reporting can be classified in many different ways. One is to differentiate reporting by the role of the person(s) preparing the report: managed reporting is reporting prepared by technical personnel such as developers; ad-hoc reporting is instead the realm of the nontechnical end-user. Another way in which reporting can be classified is by identifying the most important features of a report, such as data tables, cross-tab reports, visualization features, etc.

The Goal of Reporting (Strictly Defined)

If the flowchart of business intelligence is to see, understand and act upon data, reporting’s goal is the first–to enable end-users to see data so that they can analyze it and make it understandable through analysis. Reporting deals with data, while analysis is what turns the data into information.

For example, a sales report may include rows representing sales reps and columns showing orders taken, units sold of each major product line, revenue-dollars generated, percentage of target achieved, etc

Benefits of Reporting

Reporting is the necessary prerequisite of analysis; as such, it should be viewed in light of the goal of making data understandable and ready for easy, efficient and accurate analysis.

  • Collecting and presenting data ready to be analyzed, including historical data that can be tracked over time
  • Empowering end-users with the knowledge to become experts in their area of business
  • Having the underlying figures to back up actions and explain decisions

Reporting Best Practices 

  • Prepare your reports with the next step in mind, which is analysis. Format your data in a way that is conducive to quick and intuitive analysis, and name your columns in a way that is both consistent and understandable to the end-users.
  • If you are the report developer or system administrator, don’t confuse the end-user with too many objects. Work with them to understand what their needs are, and give them a report that is as clean as possible, eliminating clutter and unnecessary pieces of data by carefully selecting the objects end-users can see and report on.
  • Be mindful about sensitive data and use role-based security, authentication and authorization to grant or deny access to reports, columns and records to selected users or user groups.
  • Leverage the power of the Web to create reports that are powerful, interactive and easy to navigate, while placing as little a burden as possible on the system.
  • Place your reporting layer on top of the data sources that are most used in your company. In this sense, a reporting solution that is data-source neutral allows you to bring together data from traditional and non-traditional data sources, such as databases, Web services, RSS feeds, Excel, etc.

Reporting Performance

Reporting performance is generally understood to be the speed and efficiency with which reports are generated by the system when end-users perform a query. Performance depends on a variety of factors, including system bandwidth, number of concurrent users, volume of data to be presented, etc.

Naturally, some of these performance factors are “environmental,” that is, pretty much outside of the report developer’s control. Others, however, are not, and the smart report developer should be aware of these factors and use them to his advantage to create reports that perform efficiently and that don’t place undue burden on the system.

Let’s take a closer look at the two categories of factors affecting reporting performance–one environmental, one under your control as a developer.

Environmental reporting performance factors

These are the factors that are generally outside of the report developer’s control, and have more to do with the technical structure and architecture of the system on which BI runs.

  • Hardware (CPU/RAM). The more robust a hardware solution, the better the overall report performance will be.
  • Network bandwidth. Connectivity to the reporting server will play a large role in how responsive each report request will be.
  • Operating system and database platform. The main factors are the overall file management performance and data load performance, and the connectivity to the database.
  • Number of concurrent users. As with most web-based applications, user load shows a linear growth when it comes to response times.
  • Data schema. Complexity of the database or data source schema will affect performance.
  • Data Volume. Columns, rows and size/type of data value can burden your processor.
  • Report definition complexity. The number of elements and attributes in a report will affect loading time, just like the more information a web page requests, the longer it will take to load.

Dashboards can make reporting lighter on the system while offering end-users great benefits.

Performance factors in the developer’s control

Now, here is the good news. Although the factors just mentioned are in general outside of the developer’s control, there are elements within a Web application model that depend (at least in part) on the developer, who can therefore affect reporting performance.

  • Interactive paging. This allows the server to only load the proper amount of data and not become burdened, while letting the users view the information they request.
  • Drill down, drill-through (on-demand data). Besides being beneficial to the end-user, these features make the report faster and more efficient, lightening the reporting load by given only the current state of requested data.
  • Reusing elements. Reusing or sharing elements within a report allows for more complex reporting environments without adding burden to the server.
  • Dashboard visualization. With a dashboard, the user gets only the information he needs, while the load on the system will be lighter. Also, dashboards present grouped data, aggregations, KPIs, etc., further compounding the technical benefits.

What these factors in the developer’s control have in common is that they create reports that limit the amount of data they present.

ROLAP

ROLAP (Relational Online Analytical Processing) uses multidimensional data models to analyze data, and does not require the pre-computation and storage of information. ROLAP tools access the data in a relational database and generate SQL queries to calculate information.

Advantages of ROLAP include:

  • Considered to be more scalable in handling large data volumes
  • Faster loading than the automated MOLAP loads
  • Data is stored in a standard relational database and can be accessed by any SQL reporting tool
  • Tools are better at handling non-aggregatable facts
  • Possibility to successfully model data that would not fit into a strict dimensional model
  • Leverages database authorization controls

S

Salesforce Analytics

Salesforce Analytics refers to the analytical capabilities within the Salesforce platform. This includes reports and dashboards that measure against key performance indicators, like sales opportunities, leads and wins.

Scalability

Scalability is the capability to make a software application available to those who need it and want it. Scalable applications can help increase user adoption across an organization, and may even increase the time users spend engaged in self-service activities.

Scalability makes insights accessible to everyone. Dashboards and reports are available to any number of users, not just restricted to the data analysts or executives in an organization. In fact, many companies are embracing the concept of the extended enterprise − sharing insights with field operations, customers, and external partners in order to promote a common understanding of how to drive business forward.

Scalable applications may be easily deployed on the web via .NET and Java application platforms. Applications can be deployed to any major infrastructure, whether on-premise with end-users, hosted at a data center, or in a cloud-based infrastructure such as Amazon Web Services or Microsoft Azure.

Scatter Charts

Scatter charts are data visualizations best used for showing the overall relationship in a large amount of data. The data is displayed as a collection of points, each having the value of one variable determining the position on the horizontal axis, and the value of the other variable determining the position on the vertical axis. Scatter charts work best when you have an integer value on both the Y- and X-axis – because if you don’t, a scatter chart simply looks like a line chart without the line.

Scatter charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Scorecard

Historically, software scorecards were a direct result and visual representation of the theoretical balanced scorecard approach to business strategy created at Harvard Business School. Scorecards are a valuable component of any company’s business analytics toolbox, and play an essential role in helping management teams make better decisions for the future.

Software scorecards are also well-known by the disciplined top-down organizational planning process. This process intends to pinpoint the key performance metrics that best signify an organization’s progress towards their strategic goals, and then trickles down through the organization to all supporting metrics, workgroups, and individuals. Software scorecards highlight individual accountability for contributing to and achieving an organization’s strategic goals.

A Balanced Scorecard is one type of business scorecard method that is used to measure a company’s business performance against strategic goals and determine whether its business operations are linked to its objectives, strategies, and vision. When this method is used with business analytics, it’s possible to drive strategic decision-making across the company.

Self-Service BI

Self-service BI (business intelligence) is a software tool or application that empowers business users to analyze data, visualize insights, and obtain and share information in the form of reports and dashboards – without the help of IT.

The goal of self-service BI is to make business users become self-reliant and less dependent on their IT organizations to answer business questions as they arise. In many cases, IT provides these analysis tools and applications to the business. When a business user needs that data and information, they can get it themselves.

There is a massive unmet need for self-service analytics, which has prompted 25 percent of businesses to go around IT and purchase their own BI tools in the hopes of solving their own problems. However, that introduces new governance challenges around security, data quality, and reliability.

Self-Service Data Prep

elf-service data prep is a way for business users to rapidly access, blend, and prepare data for analysis without the help of IT. It is designed to give both data analysts and non-technical users a logical view of data enrichment.

Synnect Analytics is an example of a BI solution that simplifies data preparation. It enables customers to connect, acquire, and blend data from many sources, whether on-premise or in the cloud; cache that data in a high-performance self-tuning repository; and prepare the data using smart profiling, joining, and intuitive data enrichment.

Programs like DataHub include connections to cloud-app sources like Salesforce and Marketo; cloud-platform sources from Amazon and Google; and on-premise sources including HP Vertica, PostgreSQL, and ODBC-standard databases.

Social BI

Self-service data prep is a way for business users to rapidly access, blend, and prepare data for analysis without the help of IT. It is designed to give both data analysts and non-technical users a logical view of data enrichment.

Synnect Analytics is an example of a BI solution that simplifies data preparation. It enables customers to connect, acquire, and blend data from many sources, whether on-premise or in the cloud; cache that data in a high-performance self-tuning repository; and prepare the data using smart profiling, joining, and intuitive data enrichment.

Programs like DataHub include connections to cloud-app sources like Salesforce and Marketo; cloud-platform sources from Amazon and Google; and on-premise sources including HP Vertica, PostgreSQL, and ODBC-standard databases.

Social Intelligence

Social intelligence is the analysis of data from social sources, such as Twitter, Facebook, and LinkedIn. These are often used with traditional reporting and business intelligence methods to help organizations make better data-driven decisions.

Companies use social intelligence for insight on how consumers think and behave. As social technology matures, social intelligence will take on a broader role, including informing competitive strategy. Social media can help companies overcome some of the limits of older intelligence gathering approaches. This typically involves collected information from a range of public and copyrighted sources, drawing out insights using time-tested analytic methods.

Sparkline Charts

Sparkline charts are data visualizations best used for showing many trends at once, as assets of small timelines. They are great for showing variation in some measurement in a simple and condensed way. A prime example of a sparkline chart is the market summary of the U.S. DOW Jones and S&P 500 stocks.

Sparkline charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Stacked Bar Charts

Stacked bar charts, or stacked columns, are data visualizations best used for showing the relationship of individual items to the whole – comparing the contribution of each value to a total across categories. They are especially useful when real estate is limited.

Stacked bar charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Stacked Columns

Stacked columns, or stacked bar charts, are data visualizations best used for showing the relationship of individual items to the whole – comparing the contribution of each value to a total across categories. They are especially useful when real estate is limited.

Stacked columns are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

Standalone Analytics Application

The standalone analytics application represents Stage 0 in the Embedded Analytics Maturity Model, where the analytics are not embedded into the core application at all.

Similar to the traditional business intelligence model, analytics is delivered to users in a completely separate application. The only integration is data integration between the main data-generating application and the analytic application. Data access is typically provisioned through a data extract, an API, or a data export.

From the user standpoint, a standalone analytics application results in a disjointed experience. Users have to work with two separate applications, which likely look and operate differently and have no security integration. A familiar example is exporting data from the application for analysis in Excel, and creating a new copy of data along the way. Once the data changes, the Excel data becomes outdated.

Common instances of standalone analytics applications are when the product generating data has no business user interface or when the data comes from applications that you cannot embed into. An example of the former is Google Analytics. The service tracks visitor activity on a website. This visitor activity gets fed into the Google data store, but business users can only access that data by logging into the Google Analytics website.

Star Schema

In data warehousing and data marting, star schema is the simplest form of a dimensional model, with data organized into facts and dimensions. A fact is an event that is counted or measured (such as a sale or login), and a dimension contains reference information about the fact (such as date, product, or customer). A star schema is represented by surrounding each fact with its associated dimensions, and the resulting diagrams resembles a star.

The star schema consists of one or more fact tables referencing any number of dimension tables and is effective for handling simpler queries. They are optimized for querying large data sets and are used in data warehouses and data marts to support OLAP cubes, BI and analytic applications, and ad hoc queries.

Structured Data

Structured data is information that is highly organized (e.g., an Excel spreadsheet). In the world of business intelligence, structured data is preferable because it is easy to connect to and search, especially when stored in a relational database. Compiling this data and preparing it for analysis is also easy to accomplish.

On the other hand, unstructured data (e.g., an email) does not work well in a database format. It lacks organization and consistency, making it difficult to store, compile, and prepare for analysis.

System Administrator in BI

The system administrator’s roles include connection, security, maintenance and more.

The system administrator has an essential role in business intelligence. Intimately knowledgeable of the IT infrastructure and architecture on which the BI layer is to reside, he is often the one who evaluates and selects the BI solution for the company. His main tasks also include connecting the solution to the company’s data sources, establishing security and maintaining the solution so that it runs smoothly.

Let’s take a closer look at the main tasks of a system administrator, along with the different ways in which various BI models impact them.

Connecting and Integrating BI

When it comes to business intelligence, the first and most important task of a system administrator is to connect the BI solution to the company’s data sources and ensure that it works with the company’s IT architecture. Depending on the BI vendor and the technology behind the solution, this can be more or less demanding.

Some BI solutions–especially those adapted from an older legacy model–are quite demanding in this regard, requiring setup and maintenance of complex meta environment against which reporting and analysis occur. Also, this same type of BI solution may only work with particular database or data source while not working with others–or doing so only when complemented by additional data-integration software.

Conversely, there are BI solutions that are data-source neutral and that can easily be integrated with the company’s current IT environment. This makes the system administrator’s job easier in many regards: integrating the application is quicker and easier, and no resources need to be devoted to meta environments, data warehouses, connectors or data integration tools. Furthermore, this type of solution frees the system administrator to acquire the latest in data-source technology, since data-source neutral BI solutions are as much likely to work with them out of the box.

Maintenance

Maintaining the BI solution and ensuring smooth functioning is another important task of the system administrator. The more “moving parts” IT has–different applications and solutions, databases, data sources or data marts/warehouses, meta environments, etc.–the more complex this task.

A big factor that makes the system administrator’s job easier or harder in this regard is the internal consistency of the BI solution between its components. Naturally, the more complete the solution (e.g. a BI platform), the greater the chance that there may be some inconsistencies between the individual components’ technologies.

The solutions available today range from the seamlessly consistent to the mish-mash of different technologies. For instance, BI solutions that were conceived as Web-based tend to have a single, unified technologies; conversely, BI platforms that are the result of mergers and acquisitions between different companies with different philosophies are often less than consistent, and call for more knowledge and effort on the part of the system administrator and his team.

Data Security

It generally falls to the system administrator to set up data security. The purpose of data security is to ensure that only the designated personnel, roles or departments are allowed to view certain records. The obvious example: general employees are barred from seeing certain financial records such as salaries and bonuses.

Data security is handled differently by different BI solutions. The better solutions enable the system administrator to manage security at the so-called “granular level,” meaning that security can be established not only at the role or report level, but down to the record level. This gives the system administrator a great amount of flexibility in terms of empowering as many BI users as possible while keeping sensitive reports, columns, rows or even records from being accessed by all but the appropriate personnel.

T

Text Analytics

Text mining is the process of deriving insights from text. This information is typically obtained through determining patterns and trends within text through methods such as statistical pattern learning. It typically involves the process of structuring the input text, deriving a pattern within the structured data, and finally evaluating and interpreting the output.

The goal of text mining is to essentially turn text into data for analysis with applying natural language processing (NLP) and analytical methods. To accomplish this, text mining involves information and data retrieval, lexical analyses to study word frequency distributions, pattern recognition, tagging and annotation, information extraction, data mining techniques, visualization, and predictive analytics.

Some subtasks of text mining include:

  • Information retrieval or identification
  • Recognition of pattern identified entities: features such as telephone numbers, e-mail addresses, quantities, etc.
  • Relationship, fact, and event extraction: identifying associations among entities and other information in text
  • Sentiment analysis involving discerning subjective material
  • Quantitative text analysis

Transactional Systems

Transactional systems are databases that record a company’s daily transactions. The three major transactional databases include CRM (customer relationship management), HRM (human resources management), and ERP (enterprise resource planning). For instance, a sales transaction would be recorded and stored as a piece of data in the CRM database.

Transactional systems are not considered optimal for business intelligence. This is for a variety of reasons, including the fact that a) data is not optimized for reporting and analysis and b) querying directly against these databases may slow down the system and prevent the databases from recording transactions in real time.

In some cases, companies use an ETL tool to collect data from their transactional databases, transform them to be optimized for BI, and load them into a data warehouse or other data mart. The main downside of this approach is that a data warehouse is a complex and expensive architecture, which is why many other companies opt to report directly against their transactional databases.

Turning Data into Actionable Information

Analyzing data, in general, assumes that the data has already been presented, or “reported” on–in the strict definition of the word. Analyzing literally means “taking apart,” i.e. sifting through something, breaking it down in its components to better understand it. Analysis in business intelligence is therefore the art of understanding data by “taking it apart” and asking it relevant questions.

Or, put an even better way, reporting presents data; analysis turns data into information. Information that, to be useful, can be then acted upon in the interest of the company’s strategy.

We can look at analysis as the simple act of asking your data questions. Take a table of data, for example, showing you a column of sales reps’ names and another column displaying total orders taken. The data is neutral. You can’t immediately make business sense of this simple table, unless you ask it questions. Now, ask the table “who has taken the highest amount of orders?” by sorting the second column, descending. Now the data has turned into information. A simple sort has been your way to ask your data a question, and you are therefore armed with the piece of information that (say), Jones is your top-performing sales rep.

Naturally, analysis can be much more complex than this. It can involve looking at your data from multiple dimensions (OLAP), spotting trends and exceptions, and even predicting future patterns. Regardless, what all these techniques have in common is that they turn neutral data into meaningful information.

The Goal of Analysis

As we have said, analysis turns data into information. In business intelligence, this means asking relevant questions of your data so that you draw the necessary knowledge to make business decisions and take actions that further the company’s strategy.

The Benefits of Analysis

  • Analysis is the step that lets users understand their data, turning it into information; without analysis, data loses its context and much of its meaning.
  • Analysis empowers users to ask questions of their data; this is the main way in which users are said to “interact” with the data. In this sense, the more the analysis interface allows users to obtain meaningful questions of their data, the more it is interactive.
  • Analysis lends the necessary answers that guide business end-users to making the correct decisions and taking appropriate action.
  • Analysis highlights the critical factors and points the end-user towards them. By doing so, it facilitates prioritization and makes the business process more efficient.

Analysis Best Practices

  • Leverage the power of the Web to make data analysis features easy and intuitive to navigate thanks to the familiarity of the Internet
  • Empower as many end-users as possible to analyze the relevant portions of your company’s data. Do so by choosing a Web-based solution that is licensed to be distributed to unlimited end-users without additional cost, as is the case with server-based licensing.
  • Use technology smartly. Technology and the features deriving from it are tools–ask yourself what goal the tools are meant to achieve, and design your analysis interface to attain those goals.
  • Set up your reporting and analysis solutions to point the end-users to the most critical items, using features like dashboards, key performance indicators (KPIs), automated business alerts, etc.
  • Make your analysis actionable, so that the cycle “see, understand and act” is rendered as efficient as possible.

Typography

Typography is how text looks after the font and colors are set. It accounts for both the color of the font as well as the color of the background.

When designing typography for a BI dashboard, it’s important to choose text and background colors that complement one another. Color adds another dimension to your text, whether for good or bad (i.e., you wouldn’t want to have dark text on a dark background).

U

Unstructured Data

Unstructured data is information that lacks organization (structure) and consistency (e.g., an email). These attributes make the data difficult to store, compile, and prepare for analysis. With unstructured data, all of these tasks can be very time-consuming – and therefore costly. Unlike structured data, unstructured data does not work well in a database format.

User Experience

User experience, commonly abbreviated as UX, describes the overall interaction a user has with a company, product, service, and even with sales and support lines. Every little detail matters for the customer user experience.

Of course, UX is a common practice for software applications, but it hasn’t yet been fully embraced in the world of business intelligence and analytics. Beyond colors, fonts, and interactivity, UX is about having a deep understanding of who your users are and how they prefer to work with data.

To improve the UX of analytics, identify your personas and determine the experience that will help each of them be more productive. Logi Analytics developed the Continuum of Self-Service model to facilitate this process. The three personas on the continuum – consumers, creators, and analysts – each use data to make decisions in different ways, and therefore they each have different UX needs.

Regardless of where your users fall along the continuum today, or even where they move over time, you can refer to this model to match each persona to the appropriate functionality.

Along with the user interface, UX is important for driving broad user adoption. For example, not all users are “pivot-table” compliant. Focus in on the target user and their needs and capabilities, and deliver an analytics experience that is engaging and fits that end-user.

User Interface

User interface, commonly abbreviated as UI, describes the way a user accesses and interacts with a system. It’s the physical characteristics of the application, whether digital or tangible; it’s literally the part that you interact with, the face of the system to the user, and how they think of it. A UI should meet the needs of the user with simplicity and elegance.

Flat UIs have become fundamental in creating successful user experiences because they offer a straightforward, minimalistic approach to design. Utilizing two-dimensional illustrations and lessening the use of shadows, gradient textures, and three-dimensional treatments is aesthetically pleasing and helps users to process the content quickly without distraction.

Along with the user experience, UI is important for driving broad user adoption. For example, not all users are “pivot-table” compliant. Focus in on the target user and their needs and capabilities, and deliver analytics in a UI that is engaging and fits that end-user.

V

Visual Analytics

Visual analytics presents data in a way that is understandable and easy to digest. Visual analytics applications like Synnect Analytics feature intuitive visualizations and charts that enable users to quickly uncover key patterns and trends and make informed decisions.

These applications bring data to life, enabling anyone to easily combine data from many sources and discover and share insights in real time. Users can learn more about data distributions and outliers with powerful features like in-cell graphics and column profiles. They can also consult a recommendations engine to find the visualizations that best fit their underlying data.

These visualizations may be static or animated. Static visualizations provide users with basically everything you want them to see without requiring them to take any action. Usually, static visualizations display information that isn’t going to change. Animated visualizations provide users with an option to drill-down deeper into the information being presented, often requiring them to take some sort of action to make it happen.

Visual Exploration

Visual exploration drives faster insights through visual data discovery. It is a way of working with data – in the form of visualizations such as charts and graphs – to quickly uncover key patterns and trends.

Users can learn about data distributions and outliers with powerful features like in-cell graphics and column profiles. Visual analytics solutions, such as Synnect Analytics, may also include recommendations engines that use best-practice-based algorithms to automatically suggest best-fit visualizations for the underlying data.

Visual Workflow

Visual workflow, sometimes referred to as write-back, is a common analytics feature within software applications. It incorporates transactional capabilities directly within the analytic user interface. Analytics is sometimes very tightly integrated with application functionality.

Examples of visual workflows include charts embedded on an existing application page to guide user behavior; a report with editable data cells where users can update the displayed data; and a visualization with selectable regions (on a map or area of a scatter plot) allowing the user to perform an action on the selected records.

Visualizations

Visualizations are charts and graphs displaying performance metrics in a way that is understandable and easy to digest. They have the ability to tell a story through images by guiding users toward a conclusion about their data and empowering them to make informed decisions.

How does storytelling work through visualizations? Let’s start with our brains. How the brain best learns and retains information is reliant on understanding how it processes the information coming in. As we see information, it forms a visual pattern so that we quickly draw attention to key observations. So ideally, it makes sense that users can grasp the meaning of data when it is displayed in visual form, rather than spreadsheets or numbers scattered on a document.

Visualizations come in two forms: static and animated. Static visualizations provide users with basically everything you want them to see without requiring them to take any action. Usually, static visualizations display information that isn’t going to change.

Animated visualizations provide users with an option to drill-down deeper into the information being presented, often requiring them to take some sort of action to make it happen. Typically animated visuals leverage flash technology or HTML5 to render the data, while static visuals render the data as an image so they don’t require any special browser add-ons to be viewed.

W

Whisker Charts

Whisker charts are data visualizations best used for statistical analysis and to show the distribution of a dataset. The “whiskers,” shown as lines that extend vertically from the boxes, denote variability outside the upper and lower quartiles. This type of chart is also referred to as a box and whisker plot.

Whisker charts are one of many types of data visualizations used to organize and present data in a way the audience can understand and take action on. To ensure visualizations have real business value, it’s important to know which types of visualizations are best suited for a given data set.

White-Labeling

White-labeling is the ability to make embedded analytics look like your own application, not someone else’s. When it comes to BI user interfaces, white-labeling is usually a top requirement. Application providers are sensitive about ensuring the look and feel of the analytics matches their application, their company, and their brand. Good white-labeling can increase user adoption of BI applications.

Wireframes

Wireframes are an essential step in the dashboard design of a business intelligence application. A wireframe represents a draft, not an exact match, of the final layout. It shows what the user interface will look like through rough images or screenshots. Knowing where everything is going beforehand is a huge development timesaver and provides users with a much better experience.

Write-Back

Write-back is the ability to update source systems, such as databases, to maintain systems of record while staying within the context of a BI application. This is essential to the concept of embedded BI.

With support for database write-backs and the ability to initiate backend processes from within the analytic content, end-users can update source systems directly in the same context as their analysis. A good BI application will enable these data updates in a secure and managed way.

Synnect Analytics

A complete business intelligence and analytics solutions.

Synnect Analytics is a division of Synnect that focuses on Business Intelligence consulting and Analytics solutions. We focus on analytics tools that help with: Data Visualization, Data Management, Customer Intelligence, Enterprise Resource and much more. Our platforms are developed in house through alliances with industry leading companies.

We have adapted a cloud strategy with most of our tools for efficiency and minimal infrastructure implementation.

We Simplify the Solutions

By Capabilities

By Industries