How to Construct and Engage a Business Intelligence Team
OK, friends, how are we defining, organizing, and deploying our Business Intelligence teams? Are you stuck in the 90s? What are your stories?
This provocatively titled article prompted the question: Fire Your BI Team. Spoiler alert: the article doesn’t suggest you let go of all your business intelligence analysts. Rather it suggests a re-thinking of the role.
Like many corporate functions, “Business Intelligence” (BI) is often set up with a transactional mindset. BI analysts essentially serve as order takers based on tickets with laughably vague requirements begrudgingly entered by the business unit. This exercise is then followed by the business unit making numerous non-substantive requests to bend the result to a vision poorly defined in the first place. That is, if the business unit even cares about the ticket anymore.
The case made in Fire Your BI Team is to re-define the BI analyst’s role. Rather than an order-taker, the BI analyst should be considered a creative problem solver (of course). Rather than the business unit providing prescriptive requirements, the business problem should be explicated. Tell the BI analyst your problem, the author says, and let them figure out how to leverage the data to provide a solution.
This is largely correct, of course, but not novel or even specific to BI. Any collaboration within an organization should start with the participants identifying the business problem. Rather than tell your HR partner to give Jane a raise, note that you’re concerned about retaining Jane because she’s such a high performer.
Other considerations around constructing and engaging analytics teams are examined in McKinsey’s Building an Effective Analytics Organization. McKinsey considers centralized vs de-centralized organizational structures for analytics teams, as well in-house vs outsourcing, and reporting lines. As is usually the case with McKinsey, valuable insights from experience across multiple organizations are shared and may provide some guidance in your case.
In my experience, though, the key insight is one neither of these articles touches on. An effective analytics project is usually accomplished by a small team with diverse skill sets that are given ownership of a well-defined business problem. Put a BI analyst, business analyst, and data engineer together and let ‘em rip.
Yes, for sure, there are certain projects that require a larger, more structured approach (e.g., building a load balancing optimization tool and deploying to production), but for straight-forward insight generation, the small-team approach really is the 80/20 solution.
