Business Intelligence Underground

Most of the activity in Business Intelligence Projects is underground – that is, down the tunnels of IT. To much of the business, BI Is about the analytics tool or dashboard they see on their desktop – they have little understanding if just how much effort goes on behind the scenes to connect meaningful data to those tools and ensure that it is available to them in a timely manner.

In reality, BI is a combination of: Data + Business Requirements + Tools + Design + Transformation .

Data – Getting the business to fully appreciate the intricacies of a robust data management system can be a challenge – and quite frankly they do not need to know. However, they do need to take responsibility for the quality of data going forward. Piecing together historical data is not as easy as combining a few tables – the real issues are in the years of poorly controlled data entry, meaning erroneous data has been entered into the wrong fields and once combined – it creates a high error rate for the data set.

Once data has been cleared of historical errors it must be maintained in a clean state going forward – this means locking down software fields with tighter validations and training users not to just plug random text in anywhere that's available. It also means assigning data stewards to manage new data and changes to existing data definitions in each area of ​​the business. These changes are then aggregated as a central repository, managed by the BI Competency Center [BICC].

Business Requirements – The next challenge is in transforming data so that it is usable for trending and analytics. A lot of data in its current form can not be disaggregated across months, weeks, days without semantic input from the business – for example, how to deal with erratic sales of broadly varying magnitude. Selecting the right KPI and the attributes of those KPI is not as well understood as one might think. Too often, KPI used for reporting purposes are adopted by default for dashboards. Yet, many of these KPI are not qualified to performance management.

BI Tools – The layering of the BI tool on top of data only accounts for 10-20% of any BI project. It's all the users see, so it is often most of what they appreciate in terms of work effort in a BI project. I realize this sounds simplistic, however this perspective is less from ignorance and more from a lack of time to stop and think about what a BI project involves. A little bit of upfront education goes a long way in setting more reasonable expectations from business users.

Design – often overlooked is the design element of dashboards. Dashboard design is a highly skilled art from – but one that is reliably easy to learn once one is educated on key principles of visual perception and design.

Transformation – lastly, there are the behavioral changes to ensure BI becomes pervasive throughout the business. Weaning users of MS Excel spreadsheets onto over complex BI tools is not the easiest. A simpler route is to get them to appreciate the value of dashboards as a new way of working – seeing dashboards as a strategic performance system and not just a graphical report.

So what's Missing?

The main elements of BI that I see missing that I feel could contribute significantly to overcoming many of the issues inherent in the above Phases:

  1. BI Education and Co-regulation – BI is an organizational capability, even if it starts out as a local initiative. Creating a BICC does not require dedicated roles – rather it is a meeting of contributing minds across both IT and business units to ensure a common understanding of data and BI projects.
  2. BI Strategy – few businesses develop a comprehensive BI strategy. Without a shared view of how BI will be implemented and used in an organization, a fragmented set of BI capabilities quickly pervades the organization and IT get hammered with often duplicate and contradictory demands.
  3. Strategic Performance Management – without a robust connection between the strategy and the action in a business there is no clear line of performance measurement upon which to track success and apply transformational developments.
  4. Dashboard Design Skills – to transform dashboards from graphical reports to highly contributing performance management systems

In summary – Business intelligence is 10% presentation and 80% integration and 10% education.



Source by Gail La Grouw


Comments are closed.