Recognition that costs nothing and recognition that costs something

For a long time I thought I was good at recognition because I said thank you a lot. Public praise in the team channel, a mention in the weekly update, a shout in the all hands. I was generous with it and people seemed to like it.

Then an engineer told me in a 1:1 that she had noticed something. The people who got the most public praise were not the people getting promoted. The praise was real, and it was also, in her words, "what you give people instead of the thing they actually want."

She was right, and it took me a while to admit it.

Two currencies

Recognition comes in two currencies and they are not interchangeable.

The first costs nothing: a specific thank you, credit given in front of the right audience, a note to someone's skip level. It is cheap, it should be frequent, and its value comes entirely from being specific and deserved. "Great job on the launch" is worth almost nothing. "The rollback plan you wrote is the reason last night's incident took twenty minutes instead of four hours" is worth a lot, and it was free.

The second costs something: scope, title, compensation, the visible project. This currency is scarce and everyone knows it. It is the one people plan their lives around.

The mistake I made was paying in the first currency when the second was owed. The engineer who has been operating a level above her title for a year does not need another shout in the channel. She needs the promotion, and every public thank you that arrives instead of it is a reminder that I have noticed and not acted.

Make the cheap one specific

Since the free currency is only worth anything when it is specific, I stopped doing broadcast praise entirely. No "thanks everyone for a great sprint." Instead, one message per week, about one person, describing one thing they did and why it mattered. I write it in the team channel and I send a copy to their skip level.

That single habit did more for morale than any program I ran, and the reason is that people could see the criteria. The recognized work was always concrete, and it was often work nobody else had noticed: the person who cleaned up the flaky tests, the reviewer who caught the migration bug, the engineer who rewrote the onboarding doc because the new hire got lost.

Spend the expensive one honestly

For the currency that costs something, my rule is to never let it be a surprise in either direction.

If someone is on track for a promotion, they know it, they know what evidence is still missing, and they know the cycle it is likely to land in. If they are not on track, they know that too, and they know why. The promotion itself, when it comes, should feel like the confirmation of a conversation that has been going on for months.

When I cannot deliver the expensive currency, because the budget is frozen or the level is capped, I say so directly, and I say what I can do instead: a bigger project, a title change without the band change, a conversation with my director about the timeline. People can live with a delay they understand. They cannot live with one they have to guess about.

What I got wrong about rewards

I used to think recognition was about making people feel good. It is actually about making the criteria visible. Every time a manager recognizes something, the team updates its model of what gets rewarded here. If the recognition goes to the loudest person, or the one on the flashiest project, the team learns that. If it goes to the person who made the migration boring, the team learns something else.

That is why I stopped praising outcomes and started praising the behavior behind them. Not "the launch went great" but "the launch went great because you cut scope in week two when the rest of us wanted to push." The first teaches nothing. The second teaches everyone what I want more of.

During the rollout

When we moved to an AI first way of working, the first engineers who shared prompting patterns in the team library got named in the weekly note. Within a month the library had thirty entries. I did not spend anything. I just made it clear, once a week, that sharing what you had figured out was the thing this team valued. That is the cheap currency doing exactly what it is for.