What to do with headcount you did not ask for

It is a strange thing to be the silver lining.

Tech has cut more than 124,000 jobs in the first seven months of this year, already past the total for all of 2025. Fintech accounts for roughly 9,706 of those, and the names on the list are ones everyone knows.¹ Meanwhile the money in this sector moved toward payments infrastructure, where UK vacancies are forecast to rise about 14 percent this year while consumer neobanks shrink.² I am standing on the side of that line where the team grows. That is position more than virtue, and it is worth saying before anything else.

The easy move from here is to walk into a room and say we have hit the ceiling of what we can do with who we have. Give me more hands. The ceiling is real, but more hands is not necessarily what lifts it, and the req is not free to the organization we are in. Somebody else’s opportunity funds it.

Opportunity is the scarce resource right now, not headcount. The hires rate fell to 3.1 percent in February, the lowest since January 2011 and matching the worst month of 2020. Quits dropped under three million, the lowest since August 2020.³ Read those together and you get a workforce that cannot move. The engineers on my team cannot leave to grow. Whatever growth they get, I have to make it where they stand.

So the first question about a new role is not who fills it. It is whether the work in front of us needs more output from the people we have or genuinely needs someone new. The tells are not subtle once you look for them. Review latency climbing. Dependencies stalling between people rather than inside the work. Seniors spending their week unblocking instead of building. When that is the picture, another pair of hands joins the queue rather than shortening it.

The question I forget to ask is the one behind it. Is the goal itself too broad to be executed by the hands I am asking for? Some work does not need more people. It needs somebody to say out loud that we are trying to do three things and can properly fund one.

I start with who is closest to becoming a multiplier, and what it would take to get them there. If the honest answer is that two people are ready for more scope and the work is coordination bound, then another hire is redundant headcount wearing the costume of investment.

Most of the time I do not get to ask the question. A reorg moves someone onto the team. A contingent worker shows up against a budget line I never saw. Treating that as the problem is just joining the inertia. Who arrives is not mine, but the critical path is, and nobody has to approve how I use it. That is what decides what an arrival becomes. Some improve throughput from the start. Some continue to do assistive work off the critical path until they show me otherwise.

The risk on that second track is stranding people without meaning to. If you are never on the critical path you cannot demonstrate that you belong on it, and a temporary weighting quietly becomes a permanent one. So the work off the critical path has to be somewhere people can build a case for themselves, not somewhere they wait.

What I look for on it is evidence the feedback landed, and I watch for it in my own one-on-ones before anywhere else, because that is where it surfaces soonest. I see a change before the team feels it, and the team feels it before the organization records it. I am the shortest loop that can verify anything, and organizational impact is the slowest signal and the easiest to counterfeit. So I am not waiting to hear that somebody has become visible. I am watching whether the thing I said last month changed what they did this month.

We calibrate monthly, three of those, faster when the answer is obvious, and the question does not stay open past a quarter. Three is a ceiling, not a quota. There is no number of laps somebody owes me before the answer is allowed to be yes. The point is to find a fit, not force one at the organization’s expense.

I write that rule down because I have broken it. The last person who arrived on my team by reorg sat inside that open question for three quarters, and not because the answer was unclear. Closing it meant reopening a decision the reorg had already made, and I did not believe I had the standing to reopen it. So I called the delay fairness and let it run. The cost did not land on me. It landed on the person who spent three quarters being assessed for a role we both understood they were not going to get, and on the team that quietly worked around it.

What eventually closed it was not a decision I made. The fit arrived from a direction I had not been assessing. They turned out to be unusually good at jerry-rigging coordination across dependencies that report nowhere near me, which is a real skill, and one the team needed more than the thing I had spent three quarters measuring them against. So I stopped measuring them against it. They are still on the team.

Then I had the conversation I should have opened with. You have been doing this by instinct and nobody has ever named it as yours. Own it deliberately for two quarters and it stops being a habit and becomes a thing you can point at, on your resume and in the next conversation about your level. Not a promotion, and I said as much rather than dressing it up as one. A track record where there was only an aptitude, which was the honest thing I had to offer. That is a better ending than the delay deserved, and I do not get to count it as vindication. The three quarters were still mine.

Absorbing somebody is not free, and the people who pay are the ones already here. Somebody senior takes on the arrival I did not pick, and in a market where nobody can quit, that coaching is the most real growth I have left to offer after titles and transfers have gone quiet. I say so directly instead of letting them discover it as unpaid overhead. Here is the scope, here is what it builds toward, and here is the part that matters most. If we brought in the wrong person, I want to hear that from you, and saying it will not count against you.

The obvious objection is the correct one. Scope is not a promotion. Somebody who has carried expanded responsibility for eighteen months without a title has been paid in experience for work the company priced at zero, and calling that development does not change the arithmetic. So I stopped presenting scope as its own reward. The people I was asking wanted to know how it served the objectives they already held, not what it might trickle down into later, and that is a fair question with a concrete answer. So the scope has to be work that genuinely needs doing, and then we work out together what doing it will teach them. Usually it is managing expectations, or delegating work they would rather keep doing themselves. Those are capabilities that stay with the person whether or not the budget ever opens, which is the only reason I think they are worth asking for. And the delegation cannot only run toward the new engineer. It has to run to established peers too, because delegating to somebody who can push back is the harder version and the one that actually builds the skill.

That is a smaller offer than a promotion. It is the one I can make good on, because it does not depend on a budget I do not control.

I want to be careful about how far I push that, because it is an answer to a narrow question. It works when the coaching is real work somebody wanted anyway and an arrival happens to be the occasion for it. It does not answer the harder version, which is what you owe a senior engineer who is stuck, knows they are stuck, knows that you know, and has correctly worked out that scope is not currency. I have made that offer to somebody in that position and watched it land as one more thing being asked of them. I do not have a clean account of what to do there yet, and I would rather say so than let a paragraph about arrivals pass itself off as the answer.

It also expires. A quarter of this is development. A year of it is the same arrangement with better vocabulary, and if I still cannot tell them when that changes, the honest move is to say plainly that scope stopped being enough, rather than let it roll into another quarter as though nobody is keeping count.

The same honesty is owed upward, and it is cheaper to give in advance. So I forecast the dip instead of absorbing it quietly. Ramping multipliers costs throughput for a quarter or two before it returns any, and saying that number out loud in advance is the only thing that makes the next conversation about how I grow start from a better place. Predicting a shortfall has bought me more latitude than surviving one in silence ever did.

What I told my team was narrower than they deserved and truer than what I wanted to say. We are growing because the money moved toward the part of the stack we happen to sit on, not because we earned it while other teams did not. I cannot get anybody promoted this year. What I can do is make sure the work you take on while that is true is work you would have chosen anyway, and tell you plainly when something is not mine to change instead of letting you assume I chose it.

That last part took me longest. I thought naming the limits of what I could do would read as weakness, so I let people believe the constraints were preferences. They were always better at telling the difference than I gave them credit for.

If you are growing a team right now while your industry contracts, I want to know what you told your people about why. Not the version you gave your skip. The one you said out loud to somebody who could tell you were uncomfortable saying it.

Footnotes

1. Layoffs.fyi, 2026 Tech Layoffs, tracked through July 2026. Fintech share reported by IBS Intelligence, FinTech sector faces rising layoffs as firms shift to AI. ↩

2. Morgan McKinley, UK fintech hiring set to rise 14% as payments and infrastructure lead the shift away from neobanks, 2026. The vacancy forecast is UK data. ↩

3. Indeed Hiring Lab, February 2026 JOLTS Report: Stuck in Neutral, March 2026. ↩