When we delegate responsibility, we’re frequently told, we must also delegate commensurate authority. Otherwise, we can’t hold delegates accountable.

As with so many other what-we’re-frequently-told statements, it’s more plausible than accurate.

Start with the organizational chart. What it does is document operational delegation — how the CEO has delegated responsibility (and, in theory, authority) for everything that has to happen for the organization to succeed.

Imagine you’re Director of Sales for a company that sells a variety of widgets, gadgets, and doohickeys. The company’s strategic plan calls for a 20% year-over-year increase in Gadget sales. You’re responsible, so you ask your top Gadget sales reps what they need to hit the target.

“Fix the product,” they tell you. “Our gadgets are the worst in the industry.”

So you convene a Gadget Redesign Committee, which dutifully creates new-and-improved Gadget specifications. You bestow them to the Director of Gadget Engineering.

Who bestows them right back, because he, not you, is responsible for and has authority over gadget specifications.

His strategic target is a 10% increase in gadget manufacturing efficiency. Your designs would reduce it. And you’re invading his delegated-authority turf.

Oops.

No matter how you draw the boxes and connecting lines, anything important that has to happen crosses organizational boundaries. Successful products, for example, require research-and-development, design, engineering, supply chain management, manufacturing, distribution, sales, marketing, and advertising, and I’m sure I missed a few others.

Which illustrates why, as pointed out last week, there’s no such thing as a perfect org chart.

One solution is to create cross-functional teams to address boundary-crossing responsibilities. These work well when pointed at problems that require collaboration among those competent at the different disciplines required for success at whatever it is that needs to be successful.

But cross-functional teams are less well suited to when the company needs to be competent at something, with “something” encompassing organizational capabilities like: Project management, architecture, process design and management, organizational change, and quality assurance, to name a few.

That’s where Centers of Excellence (CoE) and Communities of Practice (CoP) come into play. Both are boundary-spanning entities whose goal is to improve organizational competence at a specified discipline.

Where they differ: CoEs are official entities with assigned leadership, budgets, and goals. CoPs, in contrast, are, as the name implies, true communities — self-organized, with goals defined informally and by consensus, and with the time and effort required for their success donated by community members on top of their official responsibilities.

Compare a CoE to, say, a Project Management Office (PMO).

With a PMO, the CEO has delegated both responsibility and authority for improved project success. Having authority, the head of the PMO is in a position to define and enforce the use of the company-standard project management methodology.

But as the head of every other organization has the equally delegated authority to run their operation as they see fit, including how they manage projects, conflict is built into this arrangement, right down to its core.

The head of the Project Management CoE, in contrast, is responsible for improving project management practices with no enforcement authority. Lacking authority, the CoE has to make use of a more subtle toolkit, one that includes internal marketing, persuasion and influence, a curriculum with certifications along the way, and one-on-one selling to managers throughout the enterprise.

When you can’t dictate you have to communicate and deliver benefit.

How about CoPs? As self-organizing entities they have a lot in common with volunteer organizations outside a business setting: They depend on the enthusiasm, commitment, and evangelism of people deeply committed to the practice in question. Usually, they’re even more dependent on a highly skilled, even-more-committed, charismatic leader to hold everything together, especially as many volunteers, in the absence of financial compensation, expect payment in the form of ego gratification.

CoPs can and do provide a lot of value. But they’re an unreliable organizational solution, because expecting a CoP to happen on its own is just trusting to luck. If a useful one does form, encourage it. Just don’t rely on it.

Can’t execs drive the process? No, for a simple reason: The act of encouraging a handful of experts to form a CoP instantly transforms it into a CoE, by definition.

By all means support CoPs whenever they form. But as a reliable way to increase organizational competence, CoEs are the way to go.

One last thought: Imagine the company has chartered several CoEs. Some work better than others. Should you charter a Centers of Excellence Center of Excellence?

Repress the thought. That way lies madness. Or at a minimum, uncontrolled recursion.