The Trust You Spend Without Noticing

Every time you override a team's concern about supportability, you make a withdrawal from an account nobody is tracking.

The withdrawal feels like nothing at the time, which is precisely the problem. Somebody says "this is going to be painful to maintain," you say "let's get it out and revisit," and the meeting moves on in about eleven seconds. Nobody records anything. Everyone is being reasonable. You genuinely are trying to move fast in a quarter where moving fast is the whole assignment.

But the account is real, and it has a balance, and the people who understand the balance best are the ones who never see the statement.

The sequence is always the same

I've watched this happen in three organizations now and it runs the same way every time, in the same order, at roughly the same speed.

First they stop raising concerns. Not out of resentment — out of arithmetic. Raising a concern costs about twenty minutes of social capital and has, in their experience, changed nothing. So they run the expected value and go quiet, and from where you sit this looks like alignment. It looks like the team finally getting on board.

Then they stop caring. This is the dangerous middle, because output doesn't drop yet. They still ship. They just stop doing the extra thing — the unrequested refactor, the "actually, have we thought about", the quiet Sunday-evening fix that used to keep things from breaking on Monday. Your systems cannot see this at all.

Then they leave. And here the story gets written down in a form that guarantees nobody learns anything: they left for money, or for a title, or because the market got hot. The exit interview says "new opportunity." Your retention dashboard says the number went up. Nothing in that dashboard says we spent this person and then they ran out.

You don't lose trust in a moment. You spend it in eleven-second increments over eighteen months.

What it cost to teach me this

My first job out of school was at DTCC, working on mainframes, and I did not understand what an IPL was.

An IPL — an initial program load — is what the machine does when it restarts, and the timing of it matters enormously. I was rushing at the end of a day, I made a change without properly understanding when that restart would happen, and overnight I broke the system — which is a sentence that took me years to be able to say without flinching, because at the time I genuinely believed my career had ended in a single afternoon.

It cost just over $942,000 in SEC fines, a number I could recite in my sleep for about a decade afterwards, and which I still reach for whenever somebody tells me a junior person's mistake was unforgivable.

I was then interviewed, in sequence, by every level of leadership in that organization, ending with the CIO. It is the worst professional week I have ever had, and do you know what I remember most clearly? The temperature of the room.

Here's the part that matters. My VP was a man named Tony Dolan, and behind closed doors — I only learned this later — he fought for me. What he said was roughly this: he's a bright kid, and that million-dollar mistake is going to be a great investment in his time here. I think he's going to be a net positive.

Nobody had to do that. Firing me would have been the obvious, defensible, entirely reasonable call, and no one would ever have questioned it.

I have now spent twenty years being right about things partly because a man decided my worst day was a deposit rather than a loss. That's what the account looks like when somebody is paying into it instead of drawing it down. It's not a perk and it isn't softness — Tony got a person who would run through walls for that organization for years, at the price of tolerating one genuinely expensive mistake.

Yikes, though. Nine hundred and forty-two thousand dollars is a lot of tuition.

Make the withdrawal visible before you make it

You are going to override people sometimes. That's the job, and a leader who never overrides anyone isn't empowering a team, they're abdicating.

The fix isn't to stop deciding — it's to stop deciding invisibly.

When you override a supportability concern, say the whole sentence out loud, in the room, in front of the person whose concern it was:

"I'm choosing speed here. I know that means Priya carries this, and I know that's a real cost. We're doing it anyway, and here's what I'll do about it."

Then do the thing you said. Three options, all of which are cheap:

Book the repayment now, with a date. Not "we'll revisit" — a calendar entry in eleven weeks with the specific work named on it. Vague debt never gets repaid, and everyone in the room already knows that.

Take something off them. If you can't, say so plainly rather than pretending the capacity exists. "I can't take anything off you this quarter and I know that's not good enough" is a sentence that builds more trust than any amount of pretending.

Put it in their review, in writing, that quarter. The single largest driver of quiet resignation is invisible work at review time. This is the cheapest fix in the entire essay and almost nobody does it.

The number to watch

How many concerns were raised in your last three planning meetings?

Not how many were accepted, which is a different and less interesting number, but how many were raised at all. Go back through your notes and count them, and if the number is trending toward zero while your velocity looks fine, you are not looking at a healthy team. You are looking at a team in phase two, and phase three is a resignation you will attribute to the market.

The healthiest teams I've worked with are noisy at exactly this moment: before the build, about supportability, in front of the person who can decide. Silence there is not consensus. Silence is the balance running low.

What to stop

Stop counting an override as free just because nobody argued with you about it.

Stop reading a quiet planning meeting as an aligned one, because what else would phase two look like from where you sit?

And stop treating trust as a thing you either have or lack. It's a balance, you're moving it every week in both directions, and the only real question is whether you know which way it went.

Do this in your next planning meeting

Count the concerns raised and write the number in your notes, then do it again next month and see which way it moved.

Then, the next time you override one, say the whole sentence out loud — the choice, the cost, and the name of the person carrying it.

Who on your team stopped arguing with you this year, and have you been reading that as progress?