The short answer: Finding the real problem in a business is usually fast, sometimes a single conversation. What costs money is the interval between knowing and being believed, because organizations act on evidence rather than on reasoning, and evidence only accumulates as the problem continues to fail. That interval gets paid for in cash, in years, or in the work of assembling proof, and shortening it is a different skill from being right.
I was sitting in a marketing meeting I had no business being in.
I had been hired as a contractor to build a mobile app. It was a small startup, and somewhere in the second month I got pulled into the commercial meetings, because the marketing plans and the software capabilities kept drifting apart and somebody had to reconcile them.
They had hired a premium agency for digital marketing. Social, SEO, technical SEO, the full package. Roughly twenty thousand dollars had gone out the door by that point.
They also had a target. Five hundred sign ups by the end of month three. Not purchases. Sign ups.
By the end of month two they had zero.
So I said the thing that was sitting on the table.
You were supposed to have five hundred sign ups by the end of the third month. You have none at the end of the second. How are you going to get five hundred when you have not got one?
The frowning faces
I want to be precise about what happened next, because it is the part that actually matters and it is the part people leave out of stories like this.
Nobody argued with the arithmetic. Nobody could. What I got was frowning faces, and a distinct sense that I had done something rude. The room’s reaction was not you are wrong. It was why are you interfering.
I said the honest thing, which was that it was not my business, but that red flags are red flags and these ones were on fire.
Then I tried to give them a way out. Forget the miss, I said. Worst case, you have achieved nothing so far. What happens from today? By the end of month three, are there indicators you will hit even a hundred, so that five hundred becomes reachable across months four and five?
There was no answer.
That silence was the actual finding, and it was worse than the missed number. A missed target has explanations. Slow starts happen. But nobody in that room could say whether next month would look different from last month, which meant they were not behind schedule. They were flying blind and calling it being behind schedule. From inside the plane those feel identical.
The part that cost twenty thousand dollars
Here is what I have thought about for years since.
I was right in month two.
Nothing could be done about it until month four.
The contract with the agency did eventually get terminated. What terminated it was not my argument. It was that two more months passed and there was still not a single subscriber, exactly as I had said there would not be.
Do the arithmetic on that. Twenty thousand dollars had been spent by month two. About forty thousand by the end. The two months between knowing and proving cost roughly twenty thousand dollars, and that money bought nothing at all except confirmation of something that was already known.
It is hard to convince people without evidence. That sentence sounds like a complaint. It is not. It is a description of how organizations work, and it is mostly correct behaviour. A company that reorganized itself every time a contractor had a hunch would be ungovernable.
But it has a price, and the price is not abstract. It is the burn rate of the problem multiplied by the length of the argument.
This is not a story about clever people and stupid people
The temptation here is to make the room look foolish. They were not. The founders had built the company. The agency was a real firm with real staff doing real work.
The thing they lacked was not intelligence. It was the ability to evaluate a domain they had hired out precisely because they could not do it themselves. They did not know how to set a KPI for it, so they could not tell a leading indicator from a lagging one, so they could not tell whether the answers they were getting were good answers.
The same gap that creates the need for an expert creates the inability to supervise one. That is the trap, and it is structural. It closes on the competent as reliably as on anyone else.
I have watched a version of this in every engagement I have taken since, and the shape barely changes.
Two years to change one small thing
A business owner I have known for years runs something that works, and has for a long time. What he sells is genuinely his own, which is both why people buy it and why he is the constraint on it.
The paid work came much later.
The persuasion took two years.
Two years before any of it began, I told him to change something small and cheap about how his business presented itself. He would not do it. He was successful already, so what exactly was the problem I was solving. That is not stubbornness so much as evidence: his business worked, and a thing that works is very difficult to argue with.
He arrived at the conclusion himself eventually, and arrived somewhere bigger than the small cheap change. Same conclusion I had offered. Two years of interval.
Nothing about my reasoning improved in that time. What changed is that he accumulated his own evidence.
How do you shorten the gap between knowing and being believed?
If the gap is the expensive part, then the useful skill is not seeing. It is compressing the distance between seeing and being believed. Four things have actually worked for me.
Build the evidence before you make the argument. In one engagement, the presenting complaint was that money owed was not coming in. I did not go and say the contracts were wrong. I found a pattern in the account data that should not have been there, and then walked it back through the contracts, the billing logic and the service procedures until the whole chain sat on paper and could be read by someone who had not found it. Departments that had never spoken to each other about this changed how they worked. None of that happens on an assertion.
Use their words, not yours. With that same owner, I did not tell him what was wrong. I asked him to describe how the business actually ran, wrote down what he said rather than what he concluded, and handed it back arranged in a way he had never seen it. He was not persuaded by me. He was persuaded by himself, which is the only durable kind.
Give them a forward test, not a verdict. The most useful thing I said in that marketing meeting was not this is failing. It was what would have to be true by month three for this to work, and is any of it true today. A verdict invites defence. A forward test invites a look at the instruments, and if the instruments are missing, everyone in the room finds that out at the same time without anyone being accused.
Say what it costs to wait. Not as a threat. As a number. Twenty thousand dollars for two months of confirmation is a sentence that changes a meeting, and I did not know to say it at the time.
The limit, stated honestly
I want to finish this properly, because there is a version of this essay that ends with the agency being fired and everyone learning something, and that version would be a lie.
After I finished the app, that company shut down. Not because of the agency, and not because of any single decision. The same class of mistake kept recurring until the money ran out.
Which means my finding was also a symptom.
I had correctly identified the nearest expensive instance of a defect that kept generating new instances. Removing one vendor did nothing about whatever had hired the vendor, approved the plan, and could not read the results. I was right, I was eventually believed, and it did not save the company.
That is the honest limit of the work. You can hand someone the right question. You cannot make them able to ask the next one.
So what
If you run something, the practical version of all this is short.
The problems that hurt you are not the ones nobody can see. They are the ones somebody in your building can already see and cannot get anyone to act on, because they have reasoning and not evidence, and by the time they have evidence you will have paid for it at the going rate.
Ask who in your organization has been quietly right for six months. Then ask what it would have cost you to find out sooner.
If you are in the middle of one of these and you want a second pair of eyes on it before more money moves, that is what I do for money, and I will tell you when the answer is that you do not need me.
Everything above is first hand. No client is named, and identifying details have been omitted or generalized. I have not found public research that measures the gap I am describing here, so I am not going to hang a citation on it and pretend otherwise.
Related reading
- They Already Owned the Fix: a company paying interest to cover a gap its own software could have closed.
- The Deal Was Already Lost: six months of a good argument, made in a room that was never going to decide.
- Most AI Automation Agencies Are a Scam: the same gap, in the specific case of buying automation you cannot evaluate.
- Mastering AI Agent Development: what it looks like to pick the problem before the tool.
- The Hardest Client I Ever Had: not a client case. The same gap between seeing something and acting on it, in the one situation where I was inside it.