Every time a strong engineer resigns, there is a temptation to reach for the simplest explanation: they got a better offer, it was about the money, nothing we could have done. Sometimes that is genuinely true. Far more often, by the time someone is sitting across from me with a resignation letter, money is just the thing that finally made an already-made decision easy to act on. The real reasons left months earlier, and they left quietly.
If you want to keep your best people, you have to understand the actual mechanics of why they go, which are almost never the ones written on the exit form.
They stopped learning
Good engineers are motivated by growth more than by almost anything else, including, within reason, money. When the work becomes a treadmill of similar tickets, when every problem looks like one they have already solved three times, the best people feel it first and most acutely. They do not complain about it in standup. They update their CV on a Sunday evening and tell no one.
The signal to watch is not unhappiness, which is loud and easy to spot. It is boredom in your strongest people, which is quiet and easy to miss because they are still delivering. A bored A-player is a flight risk dressed as a reliable performer.
They lost faith in the decisions
Senior engineers can tolerate a great deal, but not the steady accumulation of decisions they believe are wrong and cannot influence. When the same architectural mistakes get repeated, when their warnings about technical debt are heard and then ignored, when the roadmap keeps lurching for reasons no one explains, they reach a conclusion that is fatal to retention: this place is not fixable from where I sit.
People rarely leave because of one bad decision. They leave because of a pattern of feeling unheard, which teaches them that staying is pointless.
The dangerous part is that this conclusion is silent. By the time someone has decided the organization will not listen, they have also decided it is not worth telling you they reached that decision.
What actually keeps them
Over the years, a handful of things have reliably kept my best people engaged, and none of them are foosball tables.
- Real problems. Hand them the hard, ambiguous, important work and genuinely trust them with it.
- Visible impact. Make sure they see how their work changed the business and the customer, not just the codebase.
- Honest growth conversations. Not an annual review, but an ongoing, specific dialogue about where they are heading and whether they are getting there.
- Protection from nonsense. Shield them from the organizational churn and politics that drain the energy they would rather spend building.
The truth about compensation
I am not naive about money. If you pay meaningfully below market, all the interesting work in the world will not hold people, and you will deserve to lose them. Compensation has to be fair and competitive; it is the floor. But above that floor, more money is a surprisingly weak retention tool for engineers who are bored or who have lost faith in the direction. I have watched people turn down significant raises to leave, and accept lower offers to join somewhere they believed in. Money keeps people who would otherwise stay; it rarely keeps people who have decided to go.
The conversation to have before it is too late
The retention conversation that actually works is the one you have six months before anyone is thinking about leaving, not the panicked counteroffer after the resignation. I ask my senior people directly, in calm moments: what would make you want to leave here, and what would make you want to stay? Then I listen without getting defensive, because defensiveness teaches them never to answer honestly again.
The answers are almost never primarily about money, and they are almost always things I can act on if I hear them early enough. The counteroffer is a confession that you stopped paying attention. The early conversation is the actual work of leadership.
Conclusion: a failure of attention, not of markets
You cannot keep everyone, and you should not try; some departures are healthy for everyone involved. But losing a great engineer to preventable neglect, to boredom you could have relieved or frustration you could have heard, is a failure of leadership, not an inevitable fact of a competitive market. Pay attention early, listen without flinching, and you will keep the people who were never really for sale in the first place.