"I can build, but nobody buys."
That sentence is the defining problem of the AI era for technical people. Building used to be the moat: if you could ship, you were rare. Now anyone can ship a working product in a weekend — which means the bottleneck has moved, for everyone, to the same place: strangers becoming customers. And most of us never trained for that at all.
It is not a personality flaw
Technical people tell themselves a specific story: I am not a salesperson. Not pushy enough, smooth enough, extroverted enough. The story survives because it is flattering — it frames the gap as character rather than skill, and you cannot be blamed for your character. But watch what it protects: as long as selling is a personality you were not born with, you are excused from learning it, and free to do the thing that feels productive instead. Which is build more.
The reframe that changes everything: the pushy salesperson is not the skilled version of selling — it is the failed version. Pushing is what people do when they have skipped every earlier step and are trying to force a stranger through a wall. Skilled selling is closer to what you already do when you debug: observe, hypothesise, test, fix the actual fault.
"One more feature" is procrastination with a clean conscience
When the product is not selling, building feels like the responsible response. It has visible output, it is within your control, and nobody rejects you while you do it. That last property is the tell. Adding a feature is the only response to "nobody buys" that involves zero contact with the people not buying — which is why it is the most popular and the least effective. The information you are missing is not in your codebase. It is in their heads, and there is no API for that.
What selling actually is
A stranger becomes a customer along a path with fixed stages: they have to want something (market), see why you specifically (position), find the trade obviously worth it (offer), understand you instantly (message), encounter you at all (attention), have a route from curious to committed (journey), and be listened to rather than pitched at (conversation). None of those requires charisma. Every one can be diagnosed.
The trap: the stages fail silently and point the blame downstream. Weak positioning presents as "a marketing problem". A missing journey presents as "bad luck with conversions". So builders fix the visible symptom while the actual fault sits untouched two stages earlier.
Do this today
Write down, as a falsifiable statement, why you believe people are not buying — the way you would write a bug hypothesis. If you cannot finish the sentence with something you have tested against real potential buyers, that is the diagnosis: you are debugging without logs. Your next unit of work is not code — it is five conversations.
Originally published as a field note at strangertosold.com. If you want logs instead of theory: the free 3-minute Sales Bottleneck Finder — 18 questions, no email — names the weakest of the nine stages in your sales system.
Top comments (1)
Five conversations is the right unit of work, though the sourcing decides what ends up in the logs. Anyone you recruit for a call answers the question you asked. Someone who posted the complaint last Tuesday wrote it down before you existed, so the wording isn't shaped by an interviewer sitting there. I'd read those first, take the phrasing verbatim, then spend the calls testing a hypothesis instead of hunting for one. We built Viewfy around that ordering.