What Running an Office Taught Me About Leading People

For a period of my career I carried final technical responsibility for everything that left an office — every piece of work, reviewed and approved — while at the same time being accountable for that office's revenue against target, its recruiting, its training, its budget and its overheads. Nobody warns you that those are two different jobs, or that they disagree with each other daily.

Most of what I understand about leading people came out of that period, and almost none of it resembles what leadership books describe.

Hiring is a forecast made with almost no information

You meet someone for an hour, perhaps twice. From that you must predict how they will behave in February when they are tired, under pressure, and nobody is watching.

I stopped looking for brilliance fairly early. What I look for now is narrower: whether the person can describe something they got wrong without arranging it into a story where they were secretly right. That one signal predicts more than any qualification on the page, because it tells you whether they will tell you about a problem while it is still small.

The second thing I look for is whether they ask about the work or about the conditions. Both are legitimate. But someone who asks nothing about the actual work has told you what they will optimise for.

Training is the only real lever

If the output is not good enough, there are only three explanations: the person cannot do it, the person does not know how you want it done, or the person is not able to care. The middle one is by far the most common and by far the most fixable, and it is the one managers most often skip past on their way to concluding the first.

Training is also the least glamorous thing on any manager's list, and the first thing dropped when the month is busy. That is precisely backwards. The hour you do not spend explaining the standard is repaid in review time for the rest of the year — you simply pay it in small instalments where it is invisible, instead of once where you would have noticed.

Quality control is a relationship, not a checklist

Reviewing another person's work and returning it is one of the most delicate things in professional life. Done badly it teaches people to hide problems from you, which is the exact opposite of what a review is for.

Two things made it work. The first was separating the standard from the person out loud, every time, until it was obvious I was not keeping score. The second was being consistent about what I would and would not accept, because inconsistency is what people actually resent. Nobody minds a high bar. Everybody minds a bar that moves depending on the day.

The target changes how you listen

Here is the honest part. When you are answerable for revenue, you begin to hear operational conversations differently. Someone explains that a piece of work needs more time, and part of you is doing arithmetic while they speak.

You cannot switch that off, and I am suspicious of managers who claim they have. What you can do is notice it, and refuse to let it decide alone. The discipline is to make the quality decision first and then deal with the consequence to the number, rather than making one decision that quietly serves both and pretending it was about quality.

Every time I got that order wrong, it cost more later than the time I had saved. Not occasionally. Every time.

The thing nobody tells you about being liked

You will want to be liked. It is a normal thing to want and there is nothing shameful in it.

But being liked and being trusted are different, and they are sometimes in tension. People like a manager who lets things go. People trust a manager who is predictable — who says the same thing in the room as outside it, who does not have favourites, and who delivers difficult news early rather than letting someone discover it themselves.

Trust is worth more, and the good news is that it is largely mechanical. Say what you will do. Do it. Say when you cannot. Repeat for two years. That is most of it.

What carried across

I run businesses rather than an office now, but the transferable part turned out not to be any technique. It was the habit of holding two responsibilities at once and refusing to let the louder one make every decision.

Quality and the number are permanently in tension. Anyone who tells you a framework resolves that has not held both. What resolves it is judgement on the day, and judgement is just the accumulated memory of the times you got it wrong.

Letting someone go, and what it actually costs

This is the part of leading people that nobody prepares you for, and the part where I see the most avoidable damage done.

The mistake is almost never the decision itself. It is the timing. Managers know months before they act, and spend those months hoping the situation will resolve on its own. It rarely does, and the delay is not kindness — the person can feel the change in how they are treated long before anything is said, and that limbo is worse than the conversation.

The delay is also expensive for everyone else. A team knows when someone is not carrying their share. If nothing happens, the message received is not that the manager is compassionate. It is that the standard is negotiable, and that the people doing the work properly are absorbing the difference. That costs you the ones you least want to lose.

What I learned to do instead was to make the problem explicit far earlier, in plain language, with a specific description of what would need to change and by when. Sometimes it changed things, which is the best possible outcome. When it did not, the eventual conversation surprised nobody, and that is the most humane version available.

I would not describe any of it as easy, and I am suspicious of managers who do. But the cost of doing it late is paid by people who did nothing wrong, and once you have watched that happen you stop deferring.

Delegation is a transfer of judgement, not of tasks

For a long time I thought delegating meant handing over work. It does not, and that misunderstanding is why so many capable managers stay exhausted.

Handing over a task means the person does the thing and returns it to you for a decision. You have moved the labour and kept the judgement, which means you are still the bottleneck and they have not grown. Do that for a year and you have a team of people waiting for you.

Real delegation is handing over the decision, with the boundary stated: here is what you can decide alone, here is what you check with me, here is what is never yours. Most managers never draw that line explicitly, so their staff guess — and people guess conservatively, which means they escalate more than necessary and everyone concludes they lack initiative.

Once the boundary is stated, something uncomfortable follows. They will make decisions you would not have made. Some will be worse than yours and it will cost something. That cost is not a failure of delegation; it is the price of it, and if you are not willing to pay it you have not delegated, you have supervised. It is also the habit that makes a business capable of running without you possible at all.

What I got wrong about motivation

I used to think motivation was something a manager supplied — through encouragement, through energy, through making the work sound important. I put real effort into it.

What I observed over time is that people arrive motivated far more often than they arrive unmotivated, and that most of what a manager does is either preserve that or spend it. Unclear priorities spend it. Work that gets redone because nobody defined it properly spends it. Watching someone else get away with less spends it fastest of all.

So I stopped trying to add motivation and started looking for what was draining it, which turned out to be a much more productive question and a considerably less flattering one, because the answer was frequently something I was doing.

The one thing that genuinely adds rather than preserves is letting somebody see the consequence of their own work — the client's reaction, the number moving, the problem that did not happen. Most staff are kept at a distance from that, and it is the cheapest thing a manager can give.