I just read a column by automotive journalist Karl Ludvigsen in which he described a meeting between a senior Volvo executive and an influential industry analyst. As the exec later told the journalist, "'It's always interesting to meet people like that in person.'" As Ludvigsen explained in his column, "It wasn't a compliment."

Analysts take their share of abuse for preaching an idea broadly while not knowing much about who is listening. In their defense, analysts often face triple duty in creating research, consulting closely with clients and then trying to communicate a lesson learned to the masses.

Sometimes it's just not going to fit. I have met hundreds of analysts and count dozens as friends. I regularly seek out their advice, but I too get leery when I'm told of a new business undertaking we really need to organize and optimize company resources. I'm seeing fewer three letter, one-size-fits-all acronyms than I did five years ago, but one that gives me the jitters is COE, otherwise known as a center of excellence and sometimes as a competency center.

There might be a real gnawing need for a COE to address fragmented projects and manage resources and training from pooled expertise. But you're not going to get there by coming back from a conference waving a research report or white paper in the air.

Plus, there are analyst advocates for all kinds of COEs, for BI, for integration, for process management and more. We were talking about this on DM Radio last week with host Eric Kavanagh and our guests Jill Dyché of Baseline Consulting (who has a COE article coming up in our January print issue) and David Lyle of Informatica.

Jill recounted the case of the guy who proposed a COE because, "we wanted to get back on everyone's radar." Oops. That sounds a little like getting a bunch of warlords together to decide who's the toughest, the kind of visibility that can set a movement back a few years.

In Jill's experience, a COE needs to time its appearance, earn the right to exist before it hangs out a shingle, and the best way to do that is to bring some kind of success story to the table as something that can be discussed and built around. "How do you ask for more funding when you haven't delivered anything in six months?"

Touché, Ms. Dyché. (I'm sure she's heard that one before.) David Lyle (who's written a well-reviewed book on this topic and has another coming out soon), agreed with Jill that we need to draw boundaries around what the COE deals with, justify whatever discipline we're changing and have an agent of change that keeps the mission of the COE in mind and in perspective.

But we all agreed there are no silver bullets. Politics, personalities and resistance are the real enemies, and that's something you're just going to have to navigate for your unique organization. Eric Kavanagh related one of those funny horror stories about a database administrator who'd wound herself so deeply in the logic of BI at her workplace that she figured she'd made herself irreplaceable. I'm not sure how you overcome an entrenched personality like that, but you want to be running a COE as a center of knowledge, not a person -- or worse, a personality.

I mentioned just last week that it can be a relief when we finally confront a nagging problem of fragmented efforts with a program like data governance. I still think that's true but I also think the COE is more speculative than governance; they're certainly not the same thing. So before you raise a flag, it's wise to do a little political snooping to figure out allegiances because in the end you're going to hitch your wagon to someone. This isn't an area where there's ready technology to bail us out.

And while you're at it, take anything Jill or David or Eric or I say with a grain of salt. We don't want to come off sounding like a bunch of know-it-all experts.

(If you have a political lesson to share, sound off below and help us all out.)