Don’t fight stupid. Make more awesome.

Fighting bad or stupid practices or behaviors head-on is the hardest way to make an organization change, and it fails most of the time. I learned that to build something demonstrably better, in a way other people can adopt and own, is what actually moves a culture. Most of the time when people are saying no, what they really mean is they don’t know how to say yes. The work is to give them a way.

I refined the framework below over many years and shared it publicly at DevOpsDays Boston in 2011 and O’Reilly Velocity in 2012.

Start small

Pick the smallest project, with the most receptive people, where you can prove the idea works. Call it an experiment. A small experiment is easy to ignore, which is exactly what you want. You are trying to build something real before the organization’s antibodies notice.

Create champions

Get your boss on board first. They have to be willing to defend the work while it is still small. Then spread credit as far away from yourself as you can. The goal is for other people to feel ownership of the change, talk about it as theirs, and pull more people in.

At Amazon, I was fortunate to have Werner Vogels, Rick Dalzell, Charlie Bell, and Kim Rachmeler as executive sponsors for GameDay and the broader availability program. I am fortunate to have been mentored by all of them on how to make a huge and critical cultural change.

Use metrics to build confidence

Find a number that supports the change and use it ruthlessly. Time from commit to deploy works. So does cost of an outage. Numbers give people something safe to hold onto, and they let other people make the case on your behalf without needing to repeat your argument.

Celebrate successes

Tell the story with data. Be positive about people. Attack the problem and the constraints around it. Leave room for resistors to come around without losing face. The best outcome of a long disagreement is that the other side flips to your position so quietly they do not even register it as a change. That is winning.

Exploit compelling events

An outage, a compliance mandate, or a reorganization is an opportunity to push for the change you have been building toward. The bigger the upheaval, the more receptive people are to ideas they would have rejected the week before.

GameDay was the first place I applied this

I built GameDay at Amazon as a way to shift a punitive operational culture toward one based on learning. I started with the smallest groups of developers who were receptive, ran controlled exercises that injected real failure into critical infrastructure, and built up from there.

The Waveland lesson

During my deployment as a task force leader after Hurricane Katrina, I spent time at a volunteer kitchen in Waveland, Mississippi that was feeding thousands of people a day. FEMA kept trying to shut it down because no one was “in charge.” Eventually someone from one of the emergency management agencies said, “Make them all site directors.” The next time FEMA asked who was in charge, somebody answered “I’m a site director,” and the supplies started flowing.

The constraint was real, and the fix was a single word.

How this shows up now

The same rule shapes how I think about open-source community building. Chef won adoption because we built something developers wanted to use, and the community built on top of it. We rarely argued that the old way was broken. We didn’t have to.

It is also how I evaluate founders as an investor. I look for the ones who build tools so good their users cannot imagine going back. They spend their energy on making the better thing.

Don't fight stupid. Make more awesome.