Cutting the engineering hire was the right call because nine months of runway gave us time to prove what each customer could return, while another salary would have made a weak unit-economics answer more expensive.
In April 1970, Apollo 13 was far from Earth when an oxygen tank failed. The crew had a carbon-dioxide problem, and the square lithium hydroxide canisters available in the command module did not fit the round openings in the lunar module. NASA engineers in Houston worked with the materials the astronauts had on board and designed an adapter. The outcome was still uncertain when they began. NASA’s Apollo 13 mission archive documents the improvised fix and the crew’s safe return.
That was the shape of my Monday, at a much smaller scale. I had started the day ready to send an offer letter. We needed more engineering capacity. The backlog was real. A product decision had been waiting too long, and every customer conversation seemed to produce another request that sounded reasonable in isolation.
Then the investor questions changed.
They had spent earlier conversations asking about growth. This time they wanted to know what each customer cost us to acquire, support, and keep. They wanted to know which work was repeatable, which work depended on founder time, and what happened to our margin when a customer asked for a change that only one engineer understood.
The offer letter stayed unsent.
The hire would have solved the visible problem
A good engineering hire can remove pressure quickly. More hands means faster fixes, shorter waits, and fewer late nights where the founder is trying to translate a customer problem into a product decision.
That was exactly why the hire felt easy to justify.
But the backlog contained two different kinds of work. Some items would make the product more useful for customers we already understood. Others were requests from prospects whose willingness to pay was still vague. Hiring before separating those categories would have turned uncertainty into payroll.
The important question was not whether we had enough work for another engineer. Any early-stage company can produce enough work for another engineer. The question was whether the next engineer would help us learn faster than they increased our monthly commitment.
For the next nine months, preserving runway mattered more.
Customer cost became the operating metric
A customer can look like revenue while quietly consuming the company around it. The warning signs are familiar: a long onboarding process, custom data preparation, repeated support from the same people, and requests that make sense only for one account.
None of those are reasons to reject a customer automatically. They are reasons to price and scope the work honestly.
We started putting the full cost beside the contract value. How much founder time did this account require? Did the implementation create a reusable product capability? Could the workflow run without someone manually checking every exception? Would the customer renew at a price that covered the work?
Those questions changed the roadmap. Some requests moved forward because they appeared across several conversations. Others stayed on a list, even when a prospect wanted an answer that week.
That discipline connects with the lesson in AI demo validation: What two real retailer trials taught Ama about runway. A working demo can earn attention. It does not prove that a buyer will support the cost of delivering the product repeatedly.
Capacity follows a repeatable decision
Cutting the hire did not mean deciding engineering was less important. It meant treating the hire as a consequence of evidence, rather than a way to create it.
We kept building. We also made smaller decisions that protected the runway: fewer one-off commitments, clearer boundaries on pilots, and more direct conversations about what a customer would pay for after the first useful result.
That changed how I evaluated every request. “Can we build this?” became “Will this reduce the cost of serving the next customer, or add another exception we carry ourselves?”
The distinction matters across Accra, Berlin, London, and the US. A contract from a larger market can look like validation from a distance. It may still pull a small team away from the workflow that local customers need, or demand a level of support the current product cannot carry. The decision deserves the same scrutiny as a hire.
Keep the materials that get you home
Apollo 13’s engineers did not get to order the ideal component. They had to work with what was already inside the spacecraft and make it function under pressure.
A founder facing a runway decision has the same constraint. The answer may be a new hire later. Today, the materials on hand are the customers, product usage, team capacity, and cash already in the business. Put them together clearly before adding a fixed cost.
The offer letter can wait until the work has a repeatable reason to exist. This week, take the three customers asking for the most attention and write down their contract value, the time they consume, and the product work they create. The pattern will tell you more than the backlog does.
Comments
No comments yet.