How do you let go of a long-time employee?

Carefully, and directly. There are times in a business when someone who has been there from the beginning has to go, and there is no way around it. When that's true, be gentle and be plain, because passivity does far more damage than directness. Compensate fairly and help bridge the gap to whatever comes next.

Why do founders keep someone too long?

Fear, mostly. And passivity. And loyalty, which is the one that does the most damage because it looks like a virtue the whole time it's operating.

Founders who build something from nothing are usually intensely loyal people. It's part of why the early crew stayed through the lean years and part of why anyone believed in the thing at all. But loyalty clouds vision. It quietly reclassifies a business decision as a moral one, so that considering a change starts to feel like a betrayal, and once that happens the question stops getting asked.

So the years accumulate. The business changes shape three times. The person doesn't. And everyone can see it except the one person who has to decide.

I'd add one more, and this one is uncomfortable: sometimes the founder keeps the long-timer because the long-timer is the last person in the building who remembers when it was three guys and a van. That's not a staffing decision. That's company.

When is keeping them the right call?

Often. I want to spend real time here, because the internet's version of this conversation is far too quick to reach for the exit and the honest answer is that plenty of long-timers should stay.

Revenue per head is not the only metric that matters, and treating it as though it is will strip your company of things you cannot buy back.

Some people work slower and think better. They don't move as fast as they did at thirty and they will keep you from making a mistake that would have cost you a quarter. Wisdom shows up nowhere on a productivity report and it is worth more than most of what does.

Some people are the culture. They're the friendliest face in the building, the one a nervous new hire goes to on day three, the reason the room doesn't get mean when a job goes sideways. Research on group behaviour has found that a single person can meaningfully change how a team performs, in both directions, and most owners badly underrate the ones holding the room up. Lose that person and you'll spend two years wondering why the place feels different.

And some people know things nobody wrote down. Fifteen years of why-we-do-it-that-way lives in one head, and when that head walks out the door the business relearns it at full price.

If any of that describes the person you've been agonising over, the question probably isn't whether they stay. It's where they sit.

What does moving someone actually look like?

Get creative before you get final.

Sometimes it's an office job. The estimator who can't climb anymore but can read a set of drawings better than anyone you'll ever hire. The lead hand who's tired of the road and would be a very good scheduler. The person whose value was always judgment rather than output, in a seat that was measuring output.

This is where a lot of founders miss the obvious move, because they're thinking in terms of one ladder. The seat has outgrown the person, therefore the person goes. But you built the seats. You can build a different one, and in a business your size you have more freedom to do that than a corporation ever would.

That's the upside. Now the part that actually breaks these transitions.

The thing nobody warns you about

The hardest part of repositioning a long-timer is usually not the title, and it's usually not the money.

It's that they now have a manager.

For years this person walked into your office. They had you, directly, whenever they needed you, and that access was the real compensation. They were close to the centre. Then you grow, you put a layer in, and suddenly there is a person between them and you who is younger, who wasn't there for the hard years, and who now signs off on their work.

Nothing was taken away on paper. Everything was taken away in practice.

If you don't name that out loud, they will experience it as a demotion no matter what you call it, and the resentment will find a way out sideways. So name it. Say the thing plainly: this changes your relationship to me, and I know that's the part that stings. Then decide what you're going to do about it, because "nothing" is a choice with consequences. Some founders keep a standing monthly with their earliest people forever, for exactly this reason, and it is cheap at the price.

Openness, transparency, communication and humility go a very long way here. That's not a soft answer. It's the only one that survives contact with a person who gave you a decade.

If it does have to end

Be direct and be gentle. Both, in the same conversation.

The failure mode I see most is the founder who beats around the bush because he's trying to be kind, and ends up doing far more damage than a clear conversation would have. The person leaves the room genuinely unsure whether they've been fired. They tell their spouse something ambiguous. They spend a week in limbo. Vagueness under those conditions is not mercy. It's the founder protecting himself from a hard sentence and charging the cost to the other person.

Then compensate properly. If there's going to be a gap between this job and the next one, help them across it, and usually that means money. You can afford it more easily than they can, and you have known this was coming for longer than they have. Do the generous version. You will never regret it, and the rest of your team is watching how this gets done far more closely than you think.

One caveat worth the money it saves you: employment law on notice, severance and constructive dismissal varies enormously by jurisdiction, and it is least forgiving with long-tenured people. Get proper advice before the conversation, not after.

What is the loyalty actually costing?

You should know the number, because both directions are expensive and most founders only feel one of them.

Gallup puts the cost of replacing an employee at one-half to two times their annual salary once you count recruiting, lost productivity and the time it takes someone new to become useful. For a long-tenured person with a decade of institutional knowledge, the top of that range is optimistic. Churn is not free and founders who fire quickly rarely price it.

But the other side has a number too, and it's harder to see because it doesn't arrive as an invoice. It's the work that quietly routes around the person. The projects you don't take. The younger employee who is doing part of that job unofficially and will leave within a year because there's no room above them. The standard everyone else adjusts to, because the whole team can see what's tolerated.

That last one is the real cost. Keeping someone in a seat they cannot hold is a decision, it is visible, and everybody reads it. What they read is what the standard actually is around here, and no amount of talking about excellence overrides it.

Where does the architecture come into it?

Under almost every one of these situations is a seat that was never designed.

The person didn't fail. The role changed underneath them, four times, without anyone deciding it had, and nobody ever wrote down what the job became. So there was no moment when the gap became visible and fixable. There was just a slow accumulating sense that something wasn't right, and then a crisis.

This is the same structural problem that shows up when you can't find and keep good people, viewed from the other end. When roles live in the founder's head, nobody can grow into one on purpose and nobody can be measured against one fairly.

The fix is upstream of any conversation about a specific person. It's deciding what this business actually is now, who it serves, what work it does and doesn't take, and what each seat has to be able to hold as a result. That's a positioning decision before it's a staffing one, and it's most of what we work on with founders in exactly this spot. Until the business is decided, every personnel question is a coin flip dressed up as judgment.

One more thing about the hard version

The first volunteer I ever had to let go was a young adult in a youth ministry I led. I was friends with his parents. He wasn't a bad person and he hadn't done anything scandalous; he simply could not read what a room required, and no amount of coaching had moved it.

I called him in and worked through it. I felt cruel doing it, and I still remember the feeling.

At the end he said, "Well, I know I'm a good leader. I guess I'm just not quite right for this youth ministry." And I sat there realising he hadn't heard any of it.

I bring it up because of what I did with that feeling afterward, which was to second-guess myself for a long time. Feeling terrible about letting someone go is not evidence that you got it wrong. It's evidence that you understood what it meant. Those are different, and confusing them is how founders end up doing this two years too late, gently, badly, and at everyone's expense.

Putting it to work

Before you decide anything about a specific person, answer these on paper.

What does this seat need to hold in twelve months? Not what it holds today. Write the job as the business will need it next year, and be specific about the decisions it owns.

Which of those can this person do, and which can they learn? Split the list honestly. Most people can learn more than founders expect and fewer things than they hope.

Is there a different seat? Look at what they're actually good at rather than what their title says. Wisdom, culture, institutional memory and client trust are all real assets that a different role could use properly.

If they moved, what would they lose? Almost always the answer includes access to you. Plan for that specifically or the move will fail for a reason nobody names.

If it has to end, what does generous look like? Decide the number when you're calm rather than in the room, and get proper legal advice for where you operate before you say anything.

Adapt all of it. A ten-year lead hand in a trades business and a founding associate in a professional-services firm are not the same problem, even though the feeling in your chest is identical.

Sources

Frequently asked

How much notice or severance does a long-time employee get?

That depends entirely on your jurisdiction and their contract, and long tenure usually increases the obligation substantially. Get advice from an employment lawyer where you operate before the conversation happens. Treat anything you read online, including this, as general context rather than an answer.

Should I offer a demotion instead of letting them go?

Offer a different role, and frame it as a different role rather than a step down. It works when the new seat genuinely uses something they're good at. It fails when it's a holding pattern you invented to avoid a decision, because they'll know, and so will everyone else.

How do I tell the rest of the team?

Briefly, respectfully, and without a narrative. Say that the person has moved on, thank them for what they contributed, and say what happens to the work. Detailed explanations invite debate and make the next person wonder what will be said about them.

What if they're a family member or a close friend?

The conversation is the same and the preparation has to be better. Separate the roles explicitly when you sit down, decide in advance what you want the relationship to be afterward, and get a neutral third party involved if the business is large enough to justify it.

Am I keeping this person out of loyalty or out of judgment?

A useful test: if this person applied for their current role today, with everything you now know, would you hire them into it? If the answer is no and you can't name a different seat that fits, you have your answer, and the delay is costing you both.

Ready to look at the architecture honestly?

Book the conversation. We'll tell you what we see, and whether the work we do fits where you are.

Book the conversation →