Questions your BI initiative need to answer

Not long ago I had the privilege of supporting a Business Intelligence initiative of a client. Working closely with the CIO of the business we started the journey with a health check of their data, examining data sets, feeds from number of applications, integrity checks, and assessment of their operational and management reporting. There were several obvious transformational opportunities to be taken advantage of. However, the sales cycle went on for a prolonged period, and although we continued to work progressively toward a common goal, I sensed that my client wasn’t getting what he was after.

 One day during a routine follow up meeting, he posed an interesting question – “tell me something about my data that I don’t know”. Reacting swiftly we commissioned a small prototyping exercise to impress him and his colleagues from the commercial side of business. We thought through and articulated questions pertinent to their operations which eventually led to some interesting insights and patterns from the prototype data set (which happened to be real data).  Coupled with some intuitive aesthetically appealing graphics and a narrative, a number of pointers emerged, upon which key business decisions could be made going forward.

The question however continued to linger on in my mind, and having given it a little more thought, I realised how important it was, and perhaps the only important question we should have focused on from the start! The fundamental purpose of any BI initiative is to figure out the story data tells and more importantly what actions can be taken to leverage the value of information it provides.

Often when you take on such a project, teams get far too deeply engrossed on the technical aspects of making it happen. These include selection of tools , data definitions, the feeds from a number of operational applications, the ETL into a data warehouse, dimensions, cubes, dashboards, and the list goes on. And even after a “successful” conclusion of the project the business benefits are only a fraction of what could have been if right questions were asked and given the central focus. Many BI initiatives thus end up serving only the operational, must-have reactive reporting.  Don’t get me wrong, all those technical aspects are important and critical. Without a strong foundation of clean data, continued systemic integrity, a good understanding the business logic that manipulates the data, and the tools to access in multiple dimensions, we won’t even have a start. But the cost of ignoring the critical questions can be expensive nonetheless.

So how do we ensure we ask the right questions? The more successful businesses have this ability to identify and articulate these unique questions coming, usually, from top down. But more often, people involved in the business don’t even know what questions to ask. Therefore the BI consultants have an added responsibility of helping their clients identify and articulate those upfront. There are several ways in which one could articulate these questions. For example one can use dimensions such as those proposed in a book titled “Integrated Business Intelligence”  [1]. These are broadly categorised as (my interpretation):

• Interaction based analytics – based on all forms of interaction the business has with its customers to profile the customers or end-users. Typically this is captured via CRM, but may be outside.
• Behavioural – based on online behaviour, where applicable. Analysis of click streams, patterns etc. This slightly different from first point as in this is focused on behaviour displayed during interaction, whereas the first point about subject of interaction
• Financials – based on £’s – sales, margins, share of business and other such KPI’s
• Supply chain – based on analysis of partners’ contribution to fulfilment of product or service within a supply chain.

Or dimensions based on suggestions made by Tom Davenport in one of his HBR papers [2]

• Time – Past (reporting) , present (surveys) or future (predictive)
• Focus – What happened, Why it happened and How to make it happen differently, if desirable
• Correlation or Causation

The dimensions mentioned above are neither exhaustive nor mutually exclusive but can act as a structured guideline.  Most industries / sectors have a common set of critical parameters that need managing for the businesses to stay competitive in that marketplace. This has led to many niche providers to come up with industry specific templates – you can Google key words like ‘travel specific BI’ or ‘insurance specific BI’ to check some out, for example.  Even so , I believe every business has some unique metrics which might not be covered in ready templates but give them their competitive advantage. The above mentioned structured approach to formulating questions will help identify gaps if any.

To conclude, any BI initiative should consider identification of business critical questions, preferably at the start, which help build a storyline that is simple, insightful, credible, and above all presses emotional buttons to instigate desirable action from customers, internal users and management. A structured approach gives you the best chance to cover all the bases.

[1]  A Butler Group book – Integrated Business Intelligence

[2] HBR May 2014 – Tom Davenport, a distinguished academic and author specializing in analytics