Startup culture has plenty of commandments: sell the vision, move fast, break things.
I had invested in a founder who did none of them. Two years in, he had never oversold his product, never locked a customer into a deal they’d regret, never kept money he hadn’t earned.
Yet, he was running out of cash. His company had less than two years of runway left and a pricing model that gave most of their revenue back. The bank said he had time. The math said he didn’t.
He emailed his investors with the subject line “positioning.” The tone was controlled. Its subtext was not.
He’d attached a timeline, a runway projection, and a sentence he probably rewrote six times: I’m trying to avoid the fatal pinch. Right now we are default dead.
Two replies came back within hours. His lead investor sent a positioning template, blanks for target customer, pain point, and differentiator. Another sent three clean bullets: hit a clear ARR mark in six months, narrow the ICP until it hurts, keep burn low until the go-to-market motion proves out. Both were smart. Both were right. Both could have been sent to any founder in any category without changing a word.
Good advice travels well because it costs nothing to give. The positioning template, the three bullets — none of it asked anyone to sit with the mess. It did not ask them to hear a founder explain, for the fourth time, why his pricing model made sense. It did not ask them to watch him defend a structure you suspect existed because he could not bring himself to charge what the work was worth. It asked them to be right. It did not ask them to be there.
I canceled my afternoon and drove to the office.
I had no framework, no agenda, no template. I had questions. Because nothing that fit in three bullets was going to change the direction of his company. The problem sat one layer deeper. His plan rested on a belief he had never examined. No one around him had stayed in the room long enough to make him name it.
The first thing he pulled up was a dashboard.
Each customer paid a few hundred dollars a month. At the end of the billing cycle, the company backed out actual usage — enrichment costs, tokens, compute — and sent the rest back. In a strong month, a customer might use a fraction of that. In a slow one, almost nothing. The balance returned to them.
He walked me through it the way an engineer walks you through an architecture diagram: here’s the logic, here’s the elegance, here’s why it works. And the logic was sound.
He’d learned, painfully, that customers churned when they paid a flat fee and saw no activity. A monthly charge with nothing to show for it bred resentment. A retainer where unused funds were returned bred trust. Customers stayed. They didn’t question the invoice. They expanded when results arrived, sometimes months later. One customer came back after nine months of modest activity and signed a contract worth multiples of the original — because a single campaign had quietly generated a pipeline no one expected.
He had data for all of this. Expansion rates. Retention curves. A customer who had started small and scaled into a major annual contract after one closed deal traced back to his platform. The math held on every decision he’d made. The ledger accounted for everything except whether the company would be alive to collect.
That was the other math. He had less than two years of runway, and actual monthly revenue was whatever usage happened to be. Some months it barely cleared compute. He knew this.
He would stare at the numbers, update the forecast, then move to the next slide and explain how usage would catch up once certain features shipped. The forecast always improved. The revenue never did. He was thinking clearly about a version of the company that no longer existed.
I told him he had too much integrity.
He did not argue. He did not push back. He went quiet. He had come from a career that prized precision above all else. Rigor, never overpromising — these were ingrained in his identity. Telling him his integrity was a liability was like telling an architect his foundations were too deep.
Then he said: “Okay, but can I keep going?”
And he did. He pulled up the competitor who had taken the opposite path — a well-funded company that sold hard, locked customers into annual contracts, marketed a vision the product could not deliver. Within eighteen months the reviews circulated.
Customers felt trapped. The press noticed. He had watched this happen in real time, and it confirmed every instinct he already held: do not oversell, do not overpromise, do not take money you have not earned.
He walked me through the customer who churned when there was no visible activity for a month. He showed me the one who stayed because the refund proved good faith. He cited the difference in retention between flat-fee billing and the retainer — his model, the one he had built through trial and error and first-principles thinking over two years in a market that punished dishonesty.
Every data point was real. Every justification was earned. And every one of them was now a wall.
The problem was not the evidence. The problem was that the evidence all pointed backward. His best customer had expanded into a major annual contract — but only after months on a modest retainer. His competitor’s implosion proved that overselling destroys trust. But trust alone, without the nerve to charge for it, does not keep the lights on. His refund model reduced churn — and reduced revenue to a level that could not sustain the team.
I asked him what it costs to hire a sales rep. Seventy, eighty thousand fully loaded — before they close a single deal. His product had already justified the comparison. He was pricing it like a tool when his customers should have been comparing it to a hire.
He heard it. I could see him hear it. But he had spent two years building the price.
There is a phase of company building that rewards exactly what this founder had mastered.
At pre-seed, before the product works, before the revenue exists, the founder’s only asset is their own clarity. You sell what does not exist yet — not by lying, but by articulating someone’s problem in terms they had not found for themselves. They give you their time, their data, eventually their money, all on the strength of your word. The founders who succeed there tend to share a trait: they refuse to promise what they cannot deliver. It is what makes them credible. It is what earns them the first check.
The problem is that the trait does not know when to expire.
The same refusal to oversell that made him trustworthy at pre-seed was now preventing him from pricing the product at what it would be worth in six months. The same instinct that kept him from locking customers into contracts they would regret was now keeping his revenue at a number that could not support the team. The same honesty that had earned him the first check was now making it impossible to earn the next one. The skill that built the company had become the constraint on the company.
The pattern is common. Not the founders who are reckless — those are easy to spot and easy to advise. The ones who sit across from me and make perfect sense. And yet, they are stuck — not because they are wrong, but because they are right about something that no longer applies. The scar tissue from the last phase becomes the armor that prevents growth in the next one.
Startup wisdom tells you to hold strong opinions weakly. But that advice assumes you can identify which opinions to hold weakly. No one tells you what to do when the opinion has years of evidence behind it, when every customer interaction reinforced it, when the competitor who did the opposite blew up in public. You do not hold that opinion weakly.
You hold it like a load-bearing wall.
At some point, the wall stops holding up the building. It starts keeping you inside it.
Weeks later, he sent me a message. His demo page had wrong answers now.
Where the old form had tried to convert every visitor into a customer, the new one tried to talk them out of it. Every field had a wrong answer. Every wrong answer told the prospect, politely, to leave.
He had built a page that rejected people. The same founder who had spent two years refunding money to keep customers comfortable was now telling prospects to leave before they had spent a dollar.
The pricing changed too. He stopped selling the product as a tool and started positioning it as what it had always been: labor replacement for a function his customers could not staff. The customers who had been using the product the way it was designed were now the only customers he wanted. The real business was not in the dashboard. It was in the headcount the dashboard could replace.
But none of that is what stayed with me.
What stayed with me was a message he sent the morning after our session. He said the workshop had not pushed him in a new direction. It had articulated what he had been trying to do for twelve months but had not been able to name. He had never given himself permission to sell the future he was building rather than the present he could prove.
He called it a positioning problem. I think it was something closer to permission. The product had not changed. The team had not changed. What changed was his willingness to let go of the belief that had kept him honest and small in equal measure.
It was the same integrity. The company wasn’t.