Historically, business intelligence (BI) has always been about presenting information to users by way of reports, dashboards, and other types of data visualizations. Over the years, this presentation has become increasingly sophisticated through such advances as visual analytics, embedded BI, ad hoc analysis, and other features. But despite these advances, reports and other data visualizations have remained limited in their basic nature insofar as their primary purpose – to deliver information – has not changed.
Historically, business intelligence (BI) has always been about presenting information to users by way of reports, dashboards, and other types of data visualizations. Over the years, this presentation has become increasingly sophisticated through such advances as visual analytics, embedded BI, ad hoc analysis, and other features. But despite these advances, reports and other data visualizations have remained limited in their basic nature insofar as their primary purpose – to deliver information – has not changed. That is, BI technologies move information in one direction.
But does it have to be this way?
The answer is, “no” – enter bidirectional BI. Bidirectional BI turns business intelligence into an interactive application – one that can be deeply integrated into the operational systems of an organization, where insight is transformed into immediate action. When you place input controls into BI technologies, you’re turning a traditional model of presenting information to users into one where two-way communication is built directly into the system.
Business intelligence becomes bi-directional because the users are receiving information and then acting directly upon that information through the other systems that control the operations of the organization. This type of “closed-loop” implementation helps fulfill the original promise of BI by connecting the understanding derived by end-users into actions that benefit the organization.
Without this type of integration, there’s no direct connection between the insight gleaned by the user and any action that it might require. In traditional implementations, organizations rely on the “hope” mode; you hope that everyone looks at the data presented, you hope they understand it, you hope that they glean insights from it, and you hope that they act upon it.
In a perfect world, data visualization and analysis would already exist inside applications that need it, but that’s not the way most systems evolved. Many were in place before the advent of data visualization, and only a few have integrated modern BI into their interface. Originally, business intelligence emerged as a way of providing a broad level of insight from a lot of different systems, but it has not fully realized its promise because it isn’t integrated with operational systems.
The maturation of the BI market is one component now driving interest in bidirectional BI. As companies begin to understand, realize, and adopt the functionality of BI, they are starting to ask the question: if BI is here to affect change, why can’t I directly integrate BI with my company’s systems?
There are a number of operational platform companies that have gone so far as to acquire BI systems to integrate into part of their total package. This works for companies that have adopted that specific platform, but more and more companies have diverse portfolios of operational systems and want BI incorporated into some or all of them.
Another driving force behind bidirectional BI’s momentum is the Web. Web-based BI has enabled users to interact directly with data. As a result, BI applications have become more like regular applications that accept user input. The Web also allows for flexibility in disseminating information, giving more people access to it. And, users are generally familiar with a Web interface – most people know how to work with a browser.
Bidirectional BI is powerful because it puts control in the user’s hands and, in doing so, allows companies to keep up with the ever-increasing velocity and diversity of data that affects them. It’s one thing to simply consume and analyze data; it’s another thing to be able to interact with it, add to it, change it, and affect it in some way that streamlines your company’s business processes or improves operations.
Let’s say you have supply chain data that you’re reviewing and you see something in the data that points to an impending problem. With bidirectional BI, you have the ability to act upon that information while consuming it and without ever leaving the operational application that controls it. With traditional BI that moves only one way, acting upon the information might require an offline process involving other applications, their administrators, and any number of other steps including emails, calls, or even meetings that may take days or weeks to affect that change. Conversely, this type of scenario can be completed instantaneously with a closed-loop approach, and having it take place within the context of the operational application makes things easier and more agile for users.
A common concern about bidirectional BI is that it may negatively impact the cleanliness or reliability of data. This is certainly a legitimate concern in uncontrolled systems, but the reality is that security within well-designed bidirectional BI allows you to have control and govern who has access to data, as well as who has the ability to change it. You can also place specific rules around what happens to data if a change needs to be made. If you’re not the person who governs the data, but you see something wrong and make the change, then that change can notify someone who can first verify it, thus safeguarding data cleanliness.
Business intelligence is about arming people with information so that they can make better, data-driven decisions for their organizations. Making BI bidirectional makes the process more user-friendly, causing users to access it more readily as they see it adding value to their work. Users who have a better understanding of, ownership of, and can contribute to, data meaningfully are more organized, better informed, and deliver more value to the bottom line of their organization.
When it comes to using BI today, however, many companies still add cumbersome layers to basic processes simply because they’re operating under traditional BI practices. For example, if you’re reporting on CRM data, and you need to change that data, then you might have to access the CRM database to do so, which would then take time to propagate back to the BI system that is on top of the CRM. Only bidirectional BI would allow you to work directly with the CRM system through the embedded BI technology.
Business intelligence models in some respects are still one dimensional; much of them are designed around the idea of ‘have data, present data, consume data,’ but don’t necessarily interact with or change data. But that’s beginning to change. Companies are increasingly looking to embed BI functionality directly into their operational infrastructures as a way of creating that final connection between understanding and action that drives success.
Business intelligence technology that is able to be tightly and easily integrated directly with any application is going to be extremely useful for the next phase of the BI evolution.