The Business Intelligence Chronicles Part 7: The Need for Speed - Call for Special Forces or the Infantry ?

In all organizations generally two ideological camps emerge around the best way to implement BI : The first one driven essentially by the ERP implementation methodology which almost always precedes a BI implementation is process focused. It wants all specs to be nailed, users requirements documented, the last 'i' dotted and 't' crossed before development can commence. The second one is driven by the kind of close to the users analytical team methodology which had existed within user departments (like Sales or Finance) before ERP implementations made IT centralized. Here the focus is on speed. Specs are at best considered a necessary evil to be completed after the development is complete.

The first camp is considered as "Process Fiends", "Slow moving Bureaucrats" by the second camp which in turn is considered "Bunch of cowboys","Bulls in a China shop" by the first.

My personal approach is to effectively leverage the strengths of these two what appear to be conflicting ideologies. The larger BI group needs to be split into 2 components -

• a small "Special Forces" contingent which can quickly scout new options , build prototypes, execute limited precision deployments for a selected target audience. Focus is on 4 S's : Scout, Speed, Scalable and Selective.
• The larger "Infantry" component focuses on scaling up some of the early wins of the Special Forces component , establishes a process framework around it and launches them for mass deployment. Focus is on 3 P's: Parameters, Process, Performance.

For example , when deploying BI capabilities on a handheld device like a Blackberry, instead of getting caught up in the specification definition process, a small "Special Forces" contingent was deployed to "scout" the technology, "speedily" build a "scalable" prototype for a "selective" deployment to key stakeholders.

Subsequently the "Infantry" stepped in to scale up the "wins" of the "Special Forces" for deployment to a wider user base.They worked on the detailed specifications ("parameters"), Job Scheduling, Database Tuning ("performance"), Training documentation ("processs") etc.

Some projects executed by the "Special Forces" may never reach the "Infantry" stage and there may be others which start straight with the "Infantry"

In my opinion this two pronged approach is better, a "win-win" rather than the single pronged approach of either making your team just a Special Forces contingent (support issues when their "wins" are ramped up) or just an Infantry contingent (slow development leading to waning user interest).

What do you think ?