How do I know you won't just tell me I'm the bottleneck?

You don't, until you've seen a case where I was wrong. Here are two, both paid. In the first, a legal-tech founder's technical problems turned out to be exactly as bad as described. In the second, the founder corrected his own diagnosis by filling in my intake form, before I wrote a word. A diagnosis that can't come back wrong is worth nothing.

That's the short answer. Here's why it's the right question.

The question nobody asks out loud

Buying a diagnosis is a strange purchase. You pay someone to tell you what's wrong. You can't check the answer before you pay, and often not after. If the move works, maybe the diagnosis was right. Maybe anything would have worked.

Economists call these credence goods. Darby and Karni named them in 1973. Dulleck and Kerschbamer surveyed the research in 2006 under a title that gives the game away: doctors, mechanics, computer specialists. Their main finding is overtreatment. An expert who profits from the size of the prescription tends to find a bigger problem.

So a founder is right to ask something harder than whether I'm any good at this. Something like: you have a favourite explanation and it's about me, so what stops you finding it every time?

Fair. My work does point at the operator. Most of what I write says the problem you're describing sits downstream of a decision you haven't made. If I only published the cases that confirm that, you should discount all of it.

Here are the others.

The technical problem that was technical

A legal-tech company. About a dozen people, several years in. The founder described weak architecture, security exposure, and a data setup that survived on constant repair. The ask was an independent review.

Good conditions for the interesting read. Experienced founder, technical team, the whole problem framed as systems. The tempting move is to go a layer down and ask what the architecture talk is covering.

I looked. Nothing under it. The weaknesses were real. The security findings were real. The repair work cost what the founder said it cost.

So I wrote a review. Current state, target state, the sequence between them, the security findings.

What the evidence added was quieter than a reveal, and more expensive. Nobody there had both the time and the authority to install any of it. The freelancer lined up for the security work withdrew a month later; his main job kept eating the hours. So the recommendation came in two parts. Fix the architecture. Then get someone who can carry it. I introduced a candidate.

Insist on finding the founder inside that problem and you sell a reframe to a company that needed a document and a hire.

The founder who corrected himself

A referral-intelligence product, early, small. He bought a written second read on a plan he was already running. His stated bottleneck was demand generation.

The psychological version was easier to reach for here. Sales is the exposed work, and demand gen is what founders say when they mean they haven't been asking. I've written that post. It's often true.

His numbers went elsewhere. He had analysed 322 outbound messages. They produced 12 introductions, about 5%. Two of his six paying clients had already left, one saying flatly that he wasn't getting results. Nothing was segmented, so nobody could tell which wins came from which kind of client. Access explained why some prospects never bought. It didn't explain why the ones who paid walked.

So the stronger read was product delivery and thin proof. His refund guarantee looked like a patch over uneven outcomes.

Two things here are the point of the essay.

He got there first. Filling in the intake questions moved his diagnosis days before my answer arrived. A good second read sells the questions as much as the verdict.

Then he pushed back. He knew his block already, he said, and was running a plan against it, so what was left to buy. I could have manufactured some doubt. What I sent instead:

I don't want to talk you into doubting a plan you're already running.

Then I named the three possible answers in advance. You're right, keep going. Right area, wrong lever. Or the answers point somewhere else. And I asked whether it was still worth paying for with the first one on the list.

He bought it. He also told me afterwards, unprompted, that he wouldn't have bought from a stranger. That's a useful finding too, and not a flattering one.

Three questions to ask anyone selling you a diagnosis

Including me. None of them need you to know the answer already, which is the whole trouble with this purchase.

What would prove you wrong? A real diagnosis names what would kill it. The test I gave the second founder was to tag all 322 messages. A clear pattern among the 12 wins meant targeting. No pattern meant the product. Either result settled it. Watch for an answer that nothing could contradict. That's a worldview being applied to you.

Is "you're right, keep going" on the list? If it isn't there before you pay, you're buying a rewrite of your own explanation. Founders do arrive having named the thing correctly. Then you're paying for confirmation from someone with no reason to be polite. Worth something. Usually less than the seller admits.

What do you sell next? This is the credence-goods problem in plain form. Someone who diagnoses upstream of his own retainer has a standing interest in your problem being big. Most of them are honest. The incentive runs one way regardless, and you should price it in.

My answer: yes, there is more to buy. None of it is attached to the diagnosis. Most founders leave and do the work themselves, which is the common outcome and the right one. Some need a specialist I don't compete with, and I point them there. The ones who come back are buying against a different problem, the one that showed up once the first was handled. A founder I diagnosed last month bought an execution sprint the same week, for the thing his own diagnosis had exposed. That sequence I'll sell all day. Selling you a longer engagement because the diagnosis went well would put me straight back in the position this essay is about.

The boring answer

If every case confirmed that founders are the constraint, that would say more about my template than my perception. The method has to be able to come back boring.

Your architecture is bad, hire someone. Your product doesn't deliver yet, fix that before buying more reach.

Neither makes a good post. Both were right.

So when a founder asks how they know I won't find them at the bottom of their own problem: check. Ask what would falsify it. Ask whether keep going is on the list. Ask what comes next. Those work on anyone, including the ones who never write an essay like this.

---

The Founder Bottleneck Diagnostic is one 90-minute session, €750. You leave with the constraint named, the evidence behind it, and one concrete move. Sometimes that move is the one you were already planning. Coach, consultant, or diagnostician — which do you need? →