“Hire slow, fire fast”

A few days ago, working with an EMBA group, one participant came out with the phrase “hire slow, fire fast,” which he had heard in another session of the same program, and asked me point blank what I thought of it.

With all due respect for other people’s opinions, which are as legitimate as one’s own, I have mine. And beyond the opinion itself, what really matters is the reasoning that leads us to it. So I am going to try not to stop at “I disagree with the phrase,” but to unpack what sits underneath it, where it comes from, what it was written against, and why it has ended up meaning something different from what it meant at the start.

It is worth adding that the phrase travels well beyond that classroom. I have heard it from more than one person I respect, always said in good faith and almost always backed by some real experience. And my disagreement is not only about the tone, which is the first thing people tend to hold against it.

Is “hire slow, fire fast” a good idea?

The phrase measures two different decisions with a variable that does not fully explain either one. Hiring well depends on judgment and on defining the role beforehand, not on time spent. In a dismissal, what should be fast is the execution once the decision is made, while the diagnosis should not be judged by how long it took but by how accurate it is: whether it was documented and whether it was shared with the person affected.

Where the phrase comes from, and what it was written against

The phrase has no conclusive author, and it is worth presenting it as what it is: a management aphorism, born in the startup and entrepreneurship world, that has settled in through repetition in talks and handbooks, with almost nobody stopping to check whether it holds up.

The most concrete documented trail leads to Chuck Sekeres, founder of Physicians Weight Loss Centers. The source is Dan S. Kennedy, who says he heard it at one of his own seminars and wrote it down on the spot because it struck him as particularly sharp (his own account, with no independent confirmation, and one that does not prove Sekeres coined it either). Jeffrey J. Fox used it in 2002 as a chapter title in How to Become a Great Boss, and Greg McKeown consolidated it in an article published in Harvard Business Review on March 3, 2014. It also circulates attributed to Jack Welch, with no documentary support whatsoever, so it is better left where it is.

Authorship aside, what matters is the brief. McKeown’s article deserves to be read before it is argued with, because it does not say exactly what the slogan has ended up meaning. His target was fast-growing startups, and the behavior he described was precisely the opposite of what the phrase prescribes: they hired quickly, pushed by the urgency to grow, and then took forever to part with whoever was not working out. He called the accumulated result of that dynamic a “bozo explosion.”

His defense was not an economic one either. He argued that reversing the order was kinder in the long run, and he anticipated the obvious objection by acknowledging that the principle could sound harsh at a time of high youth unemployment. The author himself felt the need to argue that this was not cruelty.

The advice changed audience

Here is the shift that strikes me as relevant, and the one that explains why the phrase grates on me so much outside its original context.

McKeown was writing for a very specific profile: the founder who hires in the heat of the moment, with the vacancy burning a hole in his hands, and who then avoids the difficult conversation for months. For that person and at that moment, “hire slow, fire fast” is sound advice, because it pushes in exactly the direction they do not move in on their own.

Those same five words land differently in an established organization, with an HR department, quarterly targets, pressure on headcount, and managers who are evaluated on cost. The words are identical, but whoever hears them there was already leaning that way, so instead of correcting their course they confirm it, and with a reputable endorsement attached.

This advice was written for organizations that hired quickly and took forever to correct course. Repeating it where dismissals already come easily corrects no bias at all.

This is the classic problem of any model applied outside the context it was built in, something we take as read in other fields and keep tolerating in people management with remarkable cheerfulness.

There is one more detail that usually gets lost along the way. McKeown tied his argument to Reed Hastings’s “talent density,” and the Netflix model is a complete package: pay well above market, high demands, and generous severance when someone leaves. Anyone who keeps the demands and drops the pay and the severance is quoting a doctrine they are not applying.

Hiring and firing are not measured by the same yardstick

The phrase treats hiring and firing as symmetrical operations that a single stopwatch can be applied to, one counting down and the other counting up. They are not, and I think that is where its construction fails.

In hiring, time does not determine the outcome. A slow process with improvised criteria produces bad hires just as efficiently as a rushed one. What separates a good process from a bad one is defining the role before seeing anyone, the structure of the assessment, a plurality of judgment, and the honesty to admit that none of the candidates seen is the right fit (probably the most uncomfortable decision in the whole process). None of that is measured in weeks.

In a dismissal there are two clocks that the slogan melts into one: the diagnosis and the execution.

The diagnosis should be neither slow nor fast, but rigorously documented and, above all, shared with the person affected while they can still do something with that information.

The execution, once the decision is made, should be fast, clean, and properly compensated. “Fire fast” speaks about the execution and says absolutely nothing about the diagnosis, which leaves out the part of the process that determines whether the decision is sound.

That the execution should be fast is no minor concession, by the way. Dragging out an exit that has already been decided is not mercy: the person senses it long before anyone says it out loud, the team does too, and every extra month in a job where they do not work is a month they are not in one where they might. There is also a perfectly legitimate institutional version of that speed, the probationary period, designed precisely so that both sides can part early and at low cost when the fit does not hold up.

What was done with those months

When you reconstruct from the inside an exit that the organization experienced as fast, the pattern that often emerges is the opposite one. The manager had been living with the doubt for months. There were conversations in committee, consultations with HR, cost calculations and, in some cases, a discreet search for a replacement. The only genuinely fast part was the final stretch.

This deserves a qualification, because these things are rarely that clean. People in that situation usually sense something: teams are porous and the atmosphere communicates a good deal more than anyone says out loud. But sensing is not the same as having information you can work with. Perceiving that something is not quite right is one thing, and knowing exactly what is expected of you, by when, and against what criteria you will be measured is another thing entirely.

And here comes the question I find the most uncomfortable of all, far more than the one about timelines. If the organization spent those months deliberating about someone’s suitability, why did it not spend that same time exploring whether that suitability could be developed?

Because that is the prior obligation, and it gets skipped with notable frequency. Before writing off a professional, you have to have tried to turn the situation around, which means telling them clearly what is not working, agreeing on what has to change, examining whether the problem lies with the person or with the fit (the direct manager, the team, responsibilities that drifted without anyone formalizing it, a training gap that is perfectly solvable) and giving them a real window to move. If after all that the exit still comes, it will come on a different footing, and with the person having had room to change the outcome.

The usual discussion is how many months went by, when what really matters is what was done with them. If the exit happens anyway, the difference between one column and the other is that in one of them the person had room to change the ending.

Letting the months go by debating internally whether someone is good enough, without having checked whether they could be, is a shortcoming in the job of whoever is managing, and it deserves to be called by its name before we start arguing about how quickly the exit is executed. What was a long process for the organization is usually compressed, for the person on the receiving end, into one short meeting, and what is clearly open to reproach in that mismatch has to do with the conversations that were never had along the way.

So a good part of what we label firing fast is, looked at rigorously, a silent slowness followed by an abrupt ending. The ending hurts, of course. What stops the person from having done something about it is the silence that precedes it.

“Hire slow” does not hold up either

It is worth being just as demanding with the other half of the phrase, because rejecting speed as a criterion and accepting it in reverse at the same time would be fairly incoherent.

Slowness in hiring has real and well-known costs. Solid candidates have alternatives and drop out of the process while they wait for an answer that never arrives. The vacancy costs money every week it stays open, and that cost is often not calculated with the same diligence as the salary of whoever is going to fill it. Coherence and consistency, two qualities that are not always present.

And there is an even more uncomfortable phenomenon: many slow processes drag on simply because nobody will decide. They pile up interviewers with no shared criteria, repeat rounds because nobody wants to sign off, and confuse the number of stages with the depth of the analysis.

A well-run process can be resolved sooner simply because whoever is running it knows what they are looking for, so the duration ends up being an indicator of how clearly it was framed in the first place.

Why the phrase sticks

Once stripped of its context, the slogan could have collapsed on its own. It has not, and I think there is a fairly prosaic reason: speed is the only dimension of the process with a number attached to it.

Time to fill is measured everywhere. It shows up in dashboards, it gets compared across quarters, and it is used to evaluate the recruiting team. Quality of hire at twelve or eighteen months is measured almost nowhere. You would have to define it precisely, wait more than a year to be able to read it, and accept that the result calls into question whoever made the decision. None of those three things is comfortable, and the third a good deal less so than the other two.

When only one of two dimensions has a metric, the organization ends up optimizing that one. And a phrase that turns precisely that metric into a management principle has a very smooth road ahead of it.

How far this argument goes

It would be dishonest to take it further than it goes. In my previous article I argued that a dismissal is usually the bill for a hire that started badly, and I still hold that. But “usually” is not “always,” and it is better to draw the line yourself before somebody else draws it for you.

There are exits that have nothing to do with a selection error. The job changed in content, the company moved into a different stage, the role disappeared along with a product line, the fit broke when the direct manager changed, or the personal circumstances of whoever held the job shifted. An organization’s needs change and are sometimes temporary, and claiming that every early exit reveals a hiring mistake is a hindsight bias that also loads onto whoever did the selecting a responsibility that is not always theirs.

The hiring-error hypothesis deserves to be the first one examined when somebody leaves too soon, even though that examination may perfectly well end up ruling it out.

What I would put in its place

If it had to be condensed into something memorable, which almost never helps and is where most of these aphorisms come from, I would prefer something that did not talk about clocks.

None of the four fits into five words, and that is probably their main commercial defect.

Speed, at both ends of the phrase, shows up on its own when everything else is done well. When a selection process is well defined it can be resolved sooner without anyone having to suggest it. And when it drags on forever, before blaming the market or the calendar it is worth reviewing the job ad we published, the profile it describes, and whether that profile really matches what the role needs. It’s not me, it’s you.

When a working relationship is managed through honest conversations, the exit, if it comes, should surprise nobody (although some people still manage to be surprised) and could be executed in a few days without feeling like an ambush.

An organization that needs to remind itself to fire fast is usually one that arrived late to several earlier conversations. That is where it is worth looking, and not at how long it takes to execute the last of them.