In a startup's founder conversation, what is the founder actually evaluating in a candidate?

## The round's real function At a large company the first conversation is often a screen run by someone who will never work with you. At a seed-stage company the first conversation is usually with the person who decides, and it is doing three jobs at once: evaluating you, selling the company, and calibrating what the rest of the loop needs to test. Understanding that changes how you treat it. It is not a warm-up. ## Signal one: ownership under ambiguity Small teams do not have a backlog groomed into tickets. Work arrives as a customer complaint, a founder's half-formed idea, or a metric that looks wrong. The founder is listening for whether you convert that into a plan: what you would clarify first, how you would cut scope to ship something this week, what you would deliberately not build, and how you would know it worked. Candidates who describe work only as "I was assigned X and completed it" read as needing a scaffolding this company does not have. ## Signal two: specific interest Generic enthusiasm is transparent to a founder who has heard it fifty times. Specific interest looks like having used the product, having an opinion about a rough edge, understanding who the customer is and why they pay. This is cheap to acquire and rarely done. Note that the founder is testing *whether you did the reading*, not requiring you to flatter them; a well-argued criticism usually lands better than praise. ## Signal three: trust at close quarters In a nine-person company you cannot route around a difficult colleague. Founders probe how you handle disagreement, whether you say "I do not know", how you talk about previous employers, and whether you would be safe in a customer conversation. Some frame this explicitly as a values conversation; more often it is inferred from the whole exchange. ## What is not being tested This round is almost never an algorithm exercise. Preparing for it by drilling puzzles is the classic misallocation in small-company hiring, and it is doubly costly here because the preparation that *would* help, reading the product and thinking about how you would scope its problems, is cheap and takes an evening. ## The direction that gets forgotten The founder is also recruiting, and this is the only conversation in the loop where you can ask the person who knows: how the company makes money now, what the next two quarters must prove, how decisions actually get made, what the last engineer to join owns today, and how the team handles being wrong. Coming with no questions reads as low interest to a founder, and it also wastes your only high-bandwidth diligence window in a loop that gives you very few. ## How the round is recorded Often barely at all. There may be a paragraph in a shared hiring document, or the founder may simply tell the team "I liked them, let us do the paired session". Because the record is thin, the impression is the artefact. Two practical consequences: what you say in this conversation reaches the later rounds unevenly, so expect to repeat context; and a strong founder conversation cannot be relied on to carry a weak practical round, because that round produces the only concrete evidence anyone will look at. ## A worked example A founder at a nine-person seed-stage company opens their one-page hiring note with a line about what the first call is for: *we want to know what you would own in your first month*. One candidate answers by walking through how they would triage the product's current rough edges, name the one they would fix first, and say what they would need from the founder to do it. Another answers with a chronological résumé recap and a rehearsed line about wanting to work somewhere fast-moving. Both are competent engineers; only the first has demonstrated the thing that note was asking about. ## Practical preparation Spend an evening with the product and the market, write down three things you genuinely do not understand, and prepare a short account of a piece of work you scoped yourself, including what you chose not to do. Bring four questions. That is the whole preparation, and it beats a month of puzzle drills for this particular round.