The First to Leave Are the Ones You Need: What the Downturn Reveals About Keeping Programme Talent
A downturn does not trap your best people; it tells them, in a language nothing else speaks so clearly, exactly how good they are.
The resignation nobody modelled
Three weeks after the redundancy round that everyone agreed had settled things, the lead solution architect on the programme that cannot slip hands in her notice. She has been the quiet spine of the delivery for two years — the one person who holds the whole design in her head, who the testers go to when the specification is ambiguous, who has never once been on the risk log because nobody thought to put her there. She is leaving for a competitor that, in the middle of the worst market in a generation, has found the money to hire her. The programme director is genuinely shocked. In every scenario he had war-gamed through that brutal winter, this was not one of them.
I have seen this exact sequence more than once this season, and the shock is always the same, and it is always misplaced. We had built our entire retention assumption for the downturn on a piece of arithmetic that felt obviously true and was quietly, dangerously wrong.
The comforting arithmetic that is wrong
The assumption runs like this. Nobody is hiring. The market has frozen. Therefore our people, whatever they feel about the cuts and the pay freeze and the doubled workload, have nowhere to go — and so they will stay. Retention, for the duration of the crisis, takes care of itself. We can turn our management attention to the things that are actually on fire and trust that the team, lacking exits, will still be there in the morning.
Every clause of that is true on average and false exactly where it matters. The market has indeed frozen — for the average performer. But hiring never stops for the exceptional, and it particularly does not stop in a downturn, because a downturn is the one moment a competitor can acquire talent it could never ordinarily prise loose. When budgets are flush, your best architect is unavailable at any price; her employer will match anything. When budgets are tight and the same architect is being asked to do more for less under a manager visibly rationing hope, she becomes, for the first time in years, gettable. The freeze that protects you against losing your average people is precisely the thaw that exposes your best ones.
A hiring freeze is not symmetric. It shuts the exit for the people you could replace and holds it open for the people you cannot.
Why the best go first
There is a deeper mechanism underneath the market one, and it is about how strong performers read a crisis. The instinct we assume — keep your head down, be grateful to have a job — is the instinct of someone who is not sure they could get another. It is not the instinct of your best people, and the reason is not arrogance. It is information.
“The strong performer and the weak one experience the same redundancy round completely differently: one sees a threat to survive, the other sees a signal about their own worth.”
When the cuts come and the strong performer is retained, kept back, leaned on harder — she learns something specific: that she is load-bearing. That the organisation cannot easily do without her. Most people, most of the time, have only a fuzzy sense of their own market value; a crisis resolves the fuzz into a hard number, because being the one who is protected while others go is the clearest possible statement of how the organisation rates you. And a person who has just been shown, unambiguously, that she is among the most valuable people in the building will, sooner or later, wonder what that value is worth outside the building — especially when the reward for it, inside, is more work and a frozen salary. A downturn does not trap your best people; it tells them, in a language nothing else speaks so clearly, exactly how good they are.
Meanwhile the people the assumption was really about — the ones who genuinely could not move — stay, as predicted. So the net effect of leaning on the freeze will hold them is a portfolio of people slowly enriched in exactly the wrong direction: the mobile and excellent leaking away, the immobile and average retained, and the crisis programme that depended on the first group discovering the loss one resignation at a time.
The objection: some attrition is healthy
The sensible objection is that this is over-wrought. Some attrition in a downturn is not a failure but a mercy — it does redundancy’s work voluntarily, at no cost, and even improves the average. You cannot retain everyone, the argument goes, and you should not try; a business shedding cost should welcome the people who leave of their own accord, because each one is a redundancy you did not have to make. Chasing every resignation with a counter-offer you cannot afford is how a cost programme quietly reflates its own payroll.
As a statement about the workforce in aggregate, this is correct, and I have no quarrel with it. Its error is one of resolution. It reasons about attrition as a rate — a percentage of headcount — when the thing that sinks a programme is not a rate but a concentration. Losing eight per cent of a two-hundred-person delivery is healthy if it is the right eight per cent and fatal if it is the wrong four people. The objection is right that you cannot and should not retain everyone. It is wrong to conclude from this that retention does not matter, because the task was never to retain everyone. It was to retain the handful without whom the rest cannot deliver — and undifferentiated attrition, welcomed as a mercy, is precisely how you lose them without noticing.
What actually holds people in a downturn
If money is off the table — and in a pay freeze it is — then the textbook levers of retention are off the table with it, and what remains is quieter and harder to fake. In my experience the things that hold a strong performer through a bad winter are not the things the engagement survey measures.
- Meaning under stress. A person will endure a great deal of overload if the work plainly matters and they can see that it does. The programmes that lose people fastest are not the hardest ones; they are the ones where the team can no longer tell whether what they are killing themselves to deliver still counts.
- Protection from the chaos. In a crisis the scarce commodity a manager controls is not money but shelter — absorbing the noise, the reorganisations, the daily lurches of direction, so the team can still do the work. The manager who passes every gust of anxiety straight through to the team loses the team.
- Honesty about the situation. Strong people can smell reassurance that the speaker does not believe. Told the truth — including I don’t know — they mostly stay and cope. Told a comforting story that later proves false, they conclude they cannot rely on what they are told, and a person who cannot rely on what they are told starts, quietly, to plan their exit.
- Being seen. The cheapest and most neglected lever of all. The architect who leaves has usually not been thanked in a year, because she never appears on the risk log and things that work attract no attention. Being load-bearing and invisible is a combination that, in a bad time, sends people looking for somewhere they are noticed.
Protecting the critical few
The practical task, then, is not a retention policy — policies are aggregate instruments and this is a concentration problem. It is a deliberate, named act of protecting a specific small group, and it starts with knowing who they are.
- Map the concentration before you need to. For the programme that cannot slip, ask a blunt question rarely written down: if any one person resigned tomorrow, whose departure would move the go-live date? On most large deliveries the honest answer is a list of eight or ten names, and it rarely matches the org chart. Until that list exists, retention is being managed blind.
- Put the critical few on the risk log — as risks. Key-person concentration is the largest undocumented risk on most programmes precisely because the people concerned make no trouble. Naming it converts an invisible dependency into something a director can actually manage.
- Spend the non-monetary currency deliberately on that list. Meaning, shelter, honesty, recognition — these cost no budget and are therefore assumed to be free and infinite. They are neither. Direct them, consciously, at the people whose loss you cannot absorb, rather than sprinkling them evenly and hoping.
- Have the honest conversation before the resignation, not after. Once the notice is handed in, the counter-offer is both too late and too expensive, and it corrodes everyone who sees it. The conversation that keeps a critical person is the one held while they are not yet looking — the one that tells them, plainly, that they are seen, that the work matters, and that the winter is understood to be hard.
None of this is generous. It is the opposite of generous; it is triage. It accepts that in a bad year you will lose people and concentrates a scarce, unglamorous set of managerial resources on the few whose loss would be strategic rather than merely regrettable.
The winter’s real lesson
The crisis will teach a great many organisations that their retention model was an accounting model — headcount in, headcount out, a rate to be managed — when it should have been a dependency model. The people who leave first in a downturn are not a random sample of the payroll, and they are not the weakest. They are the ones with the clearest view of their own worth and the shortest walk to a competitor who has just been handed a rare chance to hire them. Assume the freeze will hold them and it will hold everyone except the people the programme was actually relying on. The manager who understands that spends the bad winter not trying to keep everyone — an impossible and unnecessary goal — but making very sure that the eight or ten names who carry the whole thing make it through to spring.