In 2002, Google decided to eliminate managers. It went terribly (as you might guess) and managers were reinstated. Still, Google remained so skeptical of management that, in 2008, they assigned a team of researchers to go figure out whether management was useful or not.

The effort was code-named Project Oxygen. The result of that research is some of the highest fidelity insight into what makes great managers great that we have ever had.

During uncertain, difficult, extraordinary times, your job as an engineering manager fundamentally changes.

When the world is predictable and the team is fully engaged, you can focus on output, set and model a high bar, give frequent feedback, and push the team to be the best that it can be. But when the world is upside down and everyone is living a reality they didn’t sign up for, the best thing you can do is be totally supportive.

“Ugh, what are we even going to talk about?”

How many times have you looked at your calendar, realized that you have a one-on-one meeting with someone on your team, and fallen into a spiral of anxiety about how you will fill the time? Maybe you end up talking about their current project’s status, or you grab a couple of questions from that “101 questions to ask in a one-on-one” list and ask those, or, almost the worst of all, you just chat about life and industry news for 45 minutes.

We had suffered a massive outage. It was so big that it made headlines across our industry, caused tangible customer churn, and sent ripples of distrust through the businesses we served. While nobody would wish for this to happen, it is something that will happen to practically any technology business eventually, and what is important is how you respond.

I thought that my company responded perfectly. They openly apologized, took responsibility, and enacted changes through the product engineering organization to shore up the stability and reliability of our systems. It was the right thing to do.

It was as if a light bulb had turned on over my head, like in cartoons.

I was discussing some performance issues in our application with the team, and I had laid out what I thought were the things we could do on our own and what we would probably need to lean on our platform support team to help out with.

I then roughly described how I thought things could play out depending on whether the platform team prioritized our concerns or not, and how that might impact what we did next.