Back to Excel - A Framework for Building Responsive BI
For decades, business users were able to analyze small sets of data independently using off-the-shelf tools like Microsoft Excel. Business users and technical non-IT personnel were able to perform a variety of tasks to extract value: from adding charts/pivots based on what best meets their needs, to using excel formulae and VBA code to incorporate business logic, and to embedding excel charts into presentation tools such as powerpoint to tell compelling business stories.
The Need for Responsive BI
Over time, with the exponential growth in the amount of data being produced and collected by businesses, every company now has huge quantities of valuable data that exceed Excel’s maximum row limit and analysis capabilities. As data grows in complexity and comes from disparate sources, business teams have been forced to give up their independence or turn to complex business intelligence tools that are rigid and cumbersome. IT departments have become overwhelmed with the growing number of change requests, and are often unable to meet the needs of businesses who want rapid changes and customized outputs. In many organizations, this dynamic has turned BI from a business driver into a bottleneck.
We are now firmly in the consumer era of B2B analytics. Business users and executives are demanding ‘self-service BI’ solutions that allow for rapid customization without the need for technical know-how and costly “system enhancements”. The challenge then is to enable B2C-like flexibility while retaining B2B-grade capabilities.
3 Key Strategies to Building Responsive BI
Building a responsive BI solution starts with setting goals based on established benchmarks that can function as the guidepost for product development. Although generic BI and visualization tools have made good strides in enabling customization, research reports have repeatedly confirmed that business users prefer the convenience and flexibility of excel over other solutions. Given that level of entrenchment, it's probably fair to say that most business users will never want to give up their ability to use Excel. This poses a challenge but also presents a clear roadmap for BI design; a roadmap backed by the collective experience of billions of Excel users. So, what can the BI solutions of tomorrow learn from Excel? We believe there are at least 3 things every solution should do:
Allow business users to choose the type of visualization
Every user perceives and interprets data in their own unique way. A bar chart may work for someone but another may prefer a simple table. This chart type bias is not trivial, since the user who prefers a specific chart type will now have to transform the visualization prior to using it in other avenues such as powerpoint. Most BI tools do have the capability to change chart types quickly but the capability is still targeted at “analyst” type users who are well-trained to use that specific BI tool. Whether building new BI capabilities or evaluating potential solutions, a combination of the “best visualization by default” and a “user level override” of that default should be the minimum requirement.
Empower business users to build and customize their own dashboards
Most dashboards are built on a lowest common denominator approach where everyone gets to see the same set of KPIs laid out in a specific navigation. Most BI solutions now offer some level of customization based on user role or profile. Still, we know that even peers in the same role prefer to look at their business differently from their colleagues. This is especially true in customer-facing roles such as sales where each market can be unique. BI solutions should allow each individual user to decide the collection of KPIs that make up their personal dashboard. Additionally, solutions should allow the user to modify the layout of the dashboard by moving, resizing, and annotating the KPIs to meet their unique business needs.
Enable technical non-IT users to edit the business logic quickly
Businesses are dynamic in nature. Its needs can change month to month, and sometimes from hour to hour. Many times, these changes require a rethink of the algorithm or business logic powering the internal BI solution. In Excel, such changes were done quickly by an anyone who was able to edit the related formulae efficiently and independently. Pure visualization tools fail in this regard since they rely on pre-aggregated data fed by another system upstream which handled the bulk of the business logic. As a result, business needs get de-prioritized in the name of change management and users have to find an alternate option aka Excel. BI solutions should focus on enabling excel formula-editing type capability into their business logic layer. This can take many forms. Within Infinity™, we allow technical non-IT users to write SQL directly from the UI. This allows our customers to publish new insights quickly and also to make business logic changes efficiently.
Business complexity can increase, data volumes can grow, and big data technologies can emerge. But the basic objectives of a responsive BI solution remain the same. They need to maximize user autonomy and minimize time to change. And the enduring power of Excel offers some important lessons towards achieving these goals.