Q&A with Jill Dyché: "Our COE Is About to Kick the Bucket!"
Can you prevent business users from circumventing the work of your analytics center of excellence?
- By Jill Dyché
- July 8, 2016
In the past decade, the analytics center of excellence has taken off. Companies have embraced the COE model for analytics, BI, and data warehouse deployment, citing the benefits of centralized skill sets, economies of scale, and competitive panache.
However, like many ideas that were originally fresh, in some companies the COE is has become a bit stale. I think Dave K.'s question represents many COE managers' frustrations.
Nine years ago I was promoted by my company, an oil and gas conglomerate, to start an analytics center of excellence. Like many organizations, we had to overcome siloes with business units using various BI and analytics tools in different ways. It was expensive and nobody kept track of what was going on.
I did a lot of missionary work back in those days, eventually convincing business unit managers to use our COE to define requirements, handle design and development, and track progress of various analytics projects across departments. Most of the business units got on board, and we've done a pretty good job managing the pipeline (so to speak).
That is until about a year and a half ago when skunkworks projects started popping up -- first in IT and then increasingly in the business units. At first these project teams assured me that they were building "prototypes" and nothing would go into production. Now there are several projects using open source, loading new data formats into big data platforms, and leveraging the cloud. I'm sure there are some I don't even know about.
My fear is that BI and analytics will become the Wild West again, a very scary thought -- and I'm from Texas! What should I do?
--Dave K., Houston
Wow, Dave, I get it. You've been a pioneer -- indeed, a maverick -- at your company, and you don't want to hang up your spurs just yet! The various teams that seem to be launching the skunkworks projects probably didn't intend to cut you out of the loop. They're probably just humming that Wynn Stewart country classic, "I Keep Forgettin' That I Forgot about You."
In fact, the longer you wait, the more likely it is that these upstart teams will "forget" to engage the COE. They'll simply move to the next phase of their "prototypes" until someone requests access. It's a slippery slope.
Dave, it's on you to catch these cowboys and round 'em up! We need them to recognize your team's skills and experience and realize that you can help them. We need them to stop singing that Dan Hicks ditty, "How Can I Miss You When You Won't Go Away?"
One thing I'm noticing more often is that analytics capabilities no longer require a single purveyor of software and skills, a function the COE has served for the past decade or so.
Indeed, many companies are turning to a "marketplace" model for analytics, where separate, small teams offer fine-grained analytics services based on specialized skills or tools. These teams needn't fall into a centralized organization. In fact, what's more important than the reporting structure is the collaboration and rules of engagement between them.
Another quality of the COE-turned-marketplace is its embrace of a consulting mindset. I vividly remember watching the head of a struggling COE painstakingly walk through his team's development process for the third time, insisting that as soon as other teams understood "how we've been writing the code," everyone would get on board.
Be willing to let go of your centralized development methodology (people care more about the "what" and the "why" than the "how") and embrace the service model, making sure members of the marketplace are rewarded for collaborating on work efforts and sharing best practices.
Moving from the COE to the analytics marketplace model requires considerable thought, a deliberate plan, and a bit of political maneuvering. It also requires willingness to embrace people and teams who might be doing things differently. Ivory tower development must cede to a culture of cocreation. Then just follow the great Johnny Cash's advice and deliver innovative analytics capabilities "One Piece at a Time."
Whipsawed by a work conundrum? Email me a question about analytics programs, data management, organizational issues, or culture at [email protected].
Jill Dyché has advised clients and executive teams on their analytics and data programs for as long as she can remember. Longer, in fact. She’s the author of four books on the business value of technology and regularly talks to teams about what keeps them up at night. Ambivalent about analytics? Maddened by management? Constricted by your culture? Check out Jill’s Q&A column, Q&A with Jill Dyché, here.