Truth Is Not Trouble. Truth Prevents Trouble.

The crisis that nearly took the company down arrived at 7:12 in the morning, disguised as a routine email.

In the book, David has just put his bag on the chair and taken a sip of coffee that is still too hot when he notices the subject line.

Safety Deviation Notice — MicroBatch 14B.

He frowns, more from curiosity than concern. MicroBatch 14B is a minor production lot for a mid-tier client in Southeast Asia. Not a flagship product. Not a high-risk environment. Barely a footnote in a vast supply chain.

The message is from a compliance officer at an overseas vendor — crisp, clinical. A deviation has been detected in a component. Nothing severe. Enough to “flag potential irregularities under specific environmental loads.”

He reads that sentence slowly.

It is the kind of deviation companies usually smooth over with technical language and a quiet fix, hoping nobody looks too closely.

The kind that, ignored, might never harm anyone.

Or might one day destroy the whole thing.

Something feels off to him. The report is too clean. Too ready.

So he opens the internal system and searches the quality logs.

And there it is.

A small entry, timestamped three weeks earlier, from a junior engineer named Ravi. Concise and unmistakably clear: possible anomaly detected, requires further review, risk uncertain.

It had been tagged for escalation.

And then left untouched.

The vendor's email wasn't the first alert. It was the second.

The Second Alert Problem

I want to be precise about what went wrong there, because the obvious reading is the wrong one.

Nobody buried Ravi's note. There was no cover-up, no manager deciding the risk was inconvenient. It was, in the book's phrase, deprioritised into invisibility — the quiet oversight that happens in complex organisations when no one means harm and everyone assumes someone else will follow through.

That is the version I have seen far more often than the dramatic one.

In more than two decades I have been in a handful of genuine cover-ups. I have been in a great many rooms where the information existed, somebody had written it down, and it simply never travelled.

And bad news, it turns out, is measurably worse at travelling than good news.

Psychologists have a name for this. In 1970, Sidney Rosen and Abraham Tesser published a study in Sociometry on what they called the MUM effect — the reluctance to communicate undesirable information.

Their experiments were simple. Give people a message to pass along. Some good, some bad. Watch what happens.

People passed on good news readily. With bad news they were slower, more hesitant, and more likely not to pass it on at all.

A follow-up field study made it less abstract. Among applicants to a disability aid agency, the time taken to communicate the decision was longer when the decision was to deny aid than to grant it.

Same information. Same effort to send. The only difference was who it would upset.

Later researchers took this into organisations and found exactly what you would expect. Work on troubled software projects found the MUM effect reduced communication from subordinates to supervisors precisely when those projects were going badly — which is to say, precisely when the communication mattered.

Read that again, because it is the uncomfortable part.

The worse the news, the less likely it is to reach you.

Your information quality does not degrade gently as problems get more serious. It degrades fastest exactly where you need it most.

What Ravi Was Actually Afraid Of

When David calls him in, Ravi arrives nervous — shoulders tense, fingers working at the seam of his notebook. He is shown his own note on the screen, and his eyes widen.

“Sir, I didn't want to create panic. I flagged it, but then I wasn't sure if it was a false alarm. I didn't want to get anyone in trouble.”

Listen to what he is actually saying.

He is not saying he didn't notice.

He is not saying he didn't care.

He is saying he ran a calculation, and the calculation told him to stop.

And here is the thing — his calculation was correct. In most organisations I have worked in, it would have been the right read.

Consider the asymmetry he was facing.

Raise the alarm and be wrong: you are the engineer who caused a scramble over nothing. That story follows you. People remember it in performance conversations.

Stay quiet and be wrong: the failure is systemic, the blame is diffuse, and nobody can point to the moment you decided.

One risk is personal and immediate. The other is collective and deferred.

A junior person weighing those two outcomes is not being cowardly. They are being rational about the incentives you built.

Which means the escalation problem is rarely the thing we name it as.

David's answer to him is the line that gives this piece its title:

“You did exactly what this company needs. Truth is not trouble. Truth prevents trouble.”

What matters is not that the sentence is inspiring. It is that it was said before anyone knew whether Ravi was right.

At that moment the anomaly was still unconfirmed. It might have been nothing. Ravi might have been the engineer who caused a scramble over nothing.

He was thanked anyway.

Reward the Flag, Not the Outcome

That is the whole mechanism, and it is narrower and more practical than a culture programme.

Separate the quality of the decision to escalate from the accuracy of the alarm.

They are different things and most organisations conflate them. We praise the person whose warning proved correct and quietly file the person whose warning proved unfounded under excitable. Do that twice in view of a team and you have taught everyone the real rule, whatever the posters say.

Three things I have found actually change the numbers.

Thank people for flags in public, at the moment of flagging. Not after vindication. The audience for that thank-you is not the person who raised it — it is everyone watching to see what happens to them.

Close every loop, visibly. Ravi's note was tagged for escalation and then nothing happened. Nothing is the most corrosive response available to you. A flag that disappears teaches a more powerful lesson than a flag that gets rejected with a reason, because rejection at least proves someone was reading.

Make escalation traceable rather than personal. If raising a concern means walking into someone's office, the cost is social and the junior person pays it. If it means an entry in a system that someone is accountable for reviewing, the cost drops close to zero. David only found the note because such a system existed.

The Cheapest Sentence in Leadership

There is a version of this article that ends by telling you to build psychological safety, and I have read enough of those to know they change nothing on Monday.

So here is the narrow version.

Somebody in your organisation is currently holding a piece of information they are not sure about.

They have half-decided it is probably nothing.

They are weighing what it costs to be the person who was wrong.

They are almost certainly weighing it correctly.

You cannot fix that by asking people to be braver.

You fix it by making the flag cheap and the silence expensive — and by saying the sentence early, before you know who was right.

Truth is not trouble. Truth prevents trouble.

It costs nothing to say when the news is good.

It only counts when you say it to someone who has just made your week harder and might turn out to have been wrong.