In my previous two posts I have tabled the concept of the Cobra Effect in BI and taken a closer look at Excel spread marts—the Cobras of BI, and the incentive EDW, aka the single version of the truth.
This week, let’s walk through the journey of creating and rolling out this single version of the truth. Will the EDW incentive prevail over Cobras of BI?
The first 3-6 months, creation to first release of the EDW incentive: After considerable back and forth and negotiation on requirements, data issues, metric definitions, data governance and data architecture build, the first release EDW or first of many integrated data marts is ready to go live. In addition, an initial set of dashboards and handful of reports are ready for release. Perhaps a little late and more expensive than originally anticipated, the new BI solution (EDW and Dashboards) is faster, accurate, scalable and has reduced manual labor through robust automation.
The new BI tool is capable of fantastic visualizations and the self-service dashboard creation capabilities seem to address most needs of analysts. IT has created robust processes to support usage, enhancements, additions and changes. There is even a roadmap describing how and when additional data marts will be built over a period of time expanding the scope of the EDW to cover core business functions. With the release of each new data mart and dashboards, corresponding spread marts (cobras of BI) will be sunset i.e. thereby eliminating cobras of BI completely over time.
All and all the mood is extremely positive as the approach and incentive of a single version of truth seems to be working.
Fast forward a few more months, has the EDW incentive worked? Business processes have evolved, data and requirements have changed challenging initial design of the data mart(s). Changes and evolution of needs; data, calculations, reports and dashboards have over whelmed IT’s limited capacity. At the same time, pressure to expand the EDW and create new data marts with additional data has increased. Limited IT capacity, evolution of needs and data challenges has pushed out roadmap timelines, increased costs and eroded confidence. After all, it’s going to take longer and business users will need to wait to get their highly functional reports and dashboards. Also, the Cobras continue to thrive with no replacement solution in sight.
New breed of cobras of BI? Business leaders, managers and analysts have a business to run and limited patience. Analysts revert to what they do best; they start building new spread marts using Excel (with or without IT support) by extracting and transforming data from EDW and new sources to produce reports. Using data from EDW, Excel spread marts can be built faster giving rise to a new breed of Cobras. The new spread marts have better data quality, to a certain degree, as part of the data comes from the trusted EDW. Nonetheless these spread marts pose a risk of departing from the Single Version of Truth over a period of time.
Despite all efforts and the incentive of “Single Version of Truth,” spread marts continue to thrive. Clearly there are tremendous benefits that analysts and business users continue to derive with the use of Excel. Agility and freedom to look at data in different ways to better understand business issues and opportunities. Reasoning with data lies at the heart of driving continuous improvements. The real disadvantage of Excel spread marts lies in the viral & untraceable consumption of data rather than the freedom of analysis.
So the bigger issue at hand is not to restrict use of Excel but to find a way to address the viralitity and traceability of Excel-based BI. In the next post in this series we’ll a take a closer look at modern Excel – Power Pivot, Power View and associated services of Power BI. This modern Excel platform improves on the agility and freedom of analysis but at the same time provides a way to govern the use of information produced from Modern Excel.
Catch up on past posts in this series:
And see you next week for Part Four!