Centers of Excellence vs. Communities of Practice in Transformation

Most digital transformations fail (87.5% fail to “meet their original objectives”). There can be many reasons for this, but in portfolio companies, one major factor is often culture. That is, the culture causes the transformation to fail. Organizations fall back to what they know, even if it’s not working.

I generally get involved with portcos when their growth has outpaced their maturity. The most likely candidate for change? The culture of software delivery (some call this DevOps).

There is no one right answer for what needs to be changed. I like to fall back on some of the great capitalist philosophers like Deming and Goldratt (author of The Goal), both systems thinkers, for guidance. Goldratt describes in a lecture how to take a holistic approach to change an organization. One message was that “local implementations of TOC (theory of constraints) are fragile”, meaning it is necessary to look at the problem holistically (systemically). If we only make a local optimization, we will run into problems where our local implementation ends.

Why holistic?

In the DevOps movement we’ve learned that Communities of Practice are more effective for transformation than Centers of Excellence. Goldratt explains that if we don’t take a holistic approach, the organization will “will revert back and deal with each issue in isolation”.

Where can we see examples of this? Centers of Excellence (CoE). With CoE we change the culture within a specific silo, the rest of the organization goes on as usual. When the ideas try to spread to the rest of the organization, the rest of the organization will not have the structure to support that culture. They may have a “I know someone that can fix this” or a “just make a ticket” culture. Those are the tools they have used to get to this point. When there is a problem, those are the tools they will instinctively reach for.

I had a client transforming because they were unable to keep up with their current workload. Instead of just adding more people, or slowing down the work, a decision was made to change the way they worked. To do more with the same.

The workload increased during the transformation because people were learning a new way of working and were not able to deliver projects as fast yet, the client panicked and wanted to cut back on the transformation. They wanted to revert back to the old way of dealing with each issue in isolation (project vs. product approach). This was despite the fact that being unable to keep up in this manner was the exact reason they were transforming in the first place!

The problem was their culture. The solution was to embrace Communities of Practice. This means spreading the new culture throughout the organization in a holistic way so that changes in one area of the system support changes in another. The system needed to be evolved and adapted to support both the old and new ways of working during the transition. It required the efforts of the community to support this effort. This requires a lot of strategic thinking.


Luckily, in crisis there is opportunity. People are more willing to put up with short term pain if they know it will lead to better outcomes. They need clear guidance on the outcomes and how they will get there. Goldratt emphasized that the solution must be simple. This sounds odd at first when dealing with complex problems, but in reality, transitioning from 47 lone wolves each handling issues individually to teams of people responsible for an area of expertise, brings simplicity. No longer is there a question of responsibility or workload. They are easier to understand in terms of teams.

Or as Goldratt says: “the more complicated the problem, the simpler must be the solution or it will never ever work”. Systems thinking for the win.