Essay · Applied for Operators

The Interview That Diagnoses the Founder

Autumn 2026 · Reading time · 18 minutes

An owner who wants to know what is wrong with her company does not need a consultant. She needs to sit through three of her own interviews with a notebook open, watching herself. What she designed into the room to protect her from the candidate is exactly the shape of what her company is protecting her from at scale.

Interviewing has always been a mirror the owner refused to look at. It is the room where she claims to be evaluating another adult, when in fact the entire structure of the conversation, the questions she asks, the answers she rewards, the silences she cannot tolerate, is a portrait of her own operating psychology. Every experienced hire I ever made taught me less about the hire than about the shape of my own bias in the moment. That is what makes the interview so useful. Not as a hiring instrument. As a diagnostic on the person conducting it.

The seat you are secretly evaluating

An interview looks like an evaluation of the person in front of the owner. It is not. The interview is, first, an evaluation of the seat the owner has designed, and only then an evaluation of the candidate. What the interview reveals about the seat is often more valuable than what it reveals about any candidate. A seat that cannot be described clearly in fifteen minutes will not be filled well by any hire, no matter how strong. A seat whose success criteria change depending on which team member is describing it will produce a person who cannot possibly satisfy the whole room. Owners who treat the interview only as a candidate assessment miss the diagnostic on the seat itself, and pay for the miss with the wrong person in a role that was never actually specified.

One test I now run before any hiring cycle is to describe the seat to a competent adult who does not work in the business and see whether she can, after five minutes, name three questions she would ask a candidate for it. If she can, the seat is well specified. If she cannot, no candidate should be interviewed until the seat is clearer, because interviews for unclear seats do not select for capability. They select for the candidate whose personality most flatters the owner’s current mood. That is a very expensive way to hire, and I have done it enough to recognize it in myself before I do it again.

Why the interview is a mirror

The candidate is a variable. The interview is a fixture. The same owner, interviewing ten different candidates in a season, will produce ten conversations whose surface differs and whose structure is nearly identical. The same three questions get asked with different wording. The same red flags get missed at the same point in the conversation. The same likable trait gets over-weighted because it flatters the owner’s own self-image. If you record an owner across a year of her interviews and lay them side by side, you are not looking at a series of assessments of ten different adults. You are looking at a study of one person’s judgment reproducing itself under slightly different lighting.

That is why the interview is a superior diagnostic to any survey the owner might complete about herself. A survey asks the owner to describe her thinking. The interview shows her thinking in motion, under time pressure, in a room where she has to make a decision an adult is waiting on. It reveals the actual instrument, not the description of it. Every operator who wants to know how she really evaluates people should sit for a full afternoon with recordings of her own interviews, no assistant summarizing them, no dashboard scoring them, just her, watching herself work. It is a corrective no external assessment can produce.

The room before the room

Long before the interview starts, the owner has already made half of the decisions that will shape it. The job description was written a certain way, the recruiter was briefed a certain way, the short list was assembled against a certain private image the owner has of who belongs in the seat. By the time the candidate walks in, the funnel has done most of the work. What the interview then does is either confirm or complicate the image the funnel already selected for. Owners who never study their own funnel behavior are astonished, year after year, at the sameness of the people they end up hiring. The sameness is not an accident. It was decided at the description stage and everything after was theater.

This is worth naming because it means the interview alone cannot correct for the funnel. If the owner wants a different quality of person in the seat, she has to change what happens before the interview at least as much as what happens during it. The interview will faithfully reproduce whatever the funnel delivered. So the diagnostic work has to widen. It is not just what you asked in the room. It is who you asked to be in the room. And who invited them. And under what description of the work. Every one of those decisions is also a portrait of the owner, and every one of them is edited by whichever tools she is using to source, screen, and schedule.

The signals a working owner watches for

There are a handful of signals that show up reliably in an owner’s own interview behavior and each one points to a specific pattern in how she runs the company. The first is talk ratio. If the owner is talking more than half the time, she is hiring for confirmation, not for capability. She is not evaluating whether the person can do the work. She is evaluating whether the person will accept her frame of the work. Companies run by owners with high interview talk ratios tend to have teams that are technically competent and strategically obedient. The obedience is not the team’s fault. It was selected for in the room.

The second signal is the treatment of silence. When a candidate stops speaking after a difficult question, some owners wait and some owners fill. The ones who fill are usually running businesses where uncomfortable truths never last long enough to be looked at, because the owner’s discomfort with silence in general is a real trait, not a specific one to interviewing. The ones who wait, even a beat past the socially comfortable, tend to run businesses where the team learns that a hard question can sit on the table without collapsing. The interview reveals which owner she is. The company mirrors it.

The third signal is the direction of the last question. An owner who ends every interview with a version of the same closing question, whether that is a warm one or a stress one, is running a company on a single emotional signature. An owner who varies the ending is running a company with room for different shapes of adult in it. Neither is correct in absolute terms. Both are diagnostic. What matters is that the owner see her own pattern and then decide, deliberately, whether it is the pattern she wants at scale.

The question the owner asks when she is uncomfortable

There is a fourth signal worth watching, and it is the one owners almost never surface on their own. It is the question the owner reaches for when she becomes uncomfortable in an interview. Every owner has one. Some retreat into a technical question they know they can grade. Some retreat into a warmth question that resets the emotional temperature of the room. Some retreat into a hypothetical that lets both parties leave the specific moment that just happened. None of these are bad questions in isolation. Their diagnostic value is that they are habitual. The owner is not choosing them for the candidate in front of her. She is choosing them for herself, because they are the ones that reliably lower her own discomfort. Notice which one you reach for, and you have found the emotional cost the interview is charging you that you had not admitted was being charged.

That cost is a signal about the whole business, not just the interview. Owners who reset with technical questions tend to run companies where difficult conversations get redirected into performance discussions. Owners who reset with warmth tend to run companies where hard feedback arrives late and softened. Owners who reset with hypotheticals tend to run companies where accountability is chronically abstract. None of this shows up on any org chart. All of it shows up in the retreat move the owner does not know she is making, and the interview is the smallest laboratory in the world for catching it in the act.

The candidate who comes prepared with a model

Every serious candidate now walks into the interview with a model in her pocket. She has rehearsed her answers with one, has stress-tested her weakness-question response, has drafted her closing thank-you before the meeting ended. This has not made candidates harder to read. It has made the reading different. A prepared candidate is not a dishonest candidate. She is a candidate whose surface is more polished than her surface used to be, and whose interior can only be reached by questions that cannot be pre-rehearsed. That is a change in the interviewing craft, not a threat to it.

The owner who is annoyed by prepared candidates is telling on herself. She has been relying on freshness of answer as a proxy for quality of thought, and freshness is no longer available. The owner who welcomes the preparation and designs around it is doing what an owner has always had to do: ask a question the candidate could not have anticipated the shape of, and then watch how the candidate thinks in front of her when no draft is ready. That is a harder question to design than it used to be. It is also a more useful one, because it is now the only kind of question that separates the person from the preparation.

The rubric that stops being useful

A rubric is a wonderful early instrument. It forces the owner to name in advance what she is looking for. It disciplines the conversation, gives structure to the debrief, and makes it possible to compare candidates on something other than mood. But every rubric has a shelf life, and the moment the owner starts finishing interviews with a full rubric completed and no genuine sense of whether the candidate is a fit, the rubric has begun to work against her. The rubric was a scaffold. Scaffolds have to be removed once the building is standing, or they become a substitute for the building.

The current version of this failure is a rubric auto-populated by a model that scored the transcript. The rows are filled, the summary reads coherent, and the owner has never actually thought about the candidate as a whole person. She has thought about the scores. This is a very specific way to hire the wrong adult with high confidence. The remedy is not to abandon the rubric. It is to insist that after the rubric is complete, the owner writes one paragraph, unaided, describing the candidate in her own words. If the paragraph is thin, the interview was thin, no matter what the scores say. That paragraph is the actual output. The rubric was the input.

The interview when the model is in the room

In many rooms now the model is not just in the candidate’s pocket. It is also in the owner’s. A transcript is running, a summary is being drafted, a scoring rubric is being suggested. Used carefully, this is an improvement. The owner is freed from note-taking and can watch the person in front of her. Used carelessly, this is a slow poisoning of the interviewer’s attention. The owner starts to interview toward the transcript. She asks the questions the summary will look coherent about. She stops asking the questions whose value is entirely in the room. Her instrument dulls in a way she cannot see, because the artifacts look better than ever.

The rule I use for myself, and recommend to any owner who asks, is that the model is present as a stenographer and never as a co-interviewer. Its job is to catch what I miss and to give me the transcript later. It is not permitted to shape the question I ask next. When I have broken this rule, and I have broken it, the interview has always weakened. The person in front of me deserves the full weight of my attention, uncoached by a tool that is optimizing for a summary I will read afterward. The interview is the last room in the business where the owner and the adult across from her should be alone with each other, even when the transcript is running.

The debrief that does the real work

If the interview is diagnostic, the debrief is where the diagnosis has to be honored. Most debriefs collapse into consensus quickly, because consensus is comfortable and disagreement is expensive. That collapse is where good hires are lost and questionable hires are approved. The debrief I now insist on has a specific shape. Each interviewer writes down her assessment before hearing anyone else’s. The written assessments are shared simultaneously. The discussion begins from the disagreements, not from the consensus. When there are no disagreements, I am suspicious. Real adults evaluating the same candidate almost always see slightly different things, and if they do not, either the interviewers are too similar or the assessment is too shallow.

In the current season, the temptation is to let a model synthesize the assessments and offer a recommendation. Do not accept the synthesis as the decision. Read it, use it to sharpen your own thinking, and then set it aside. The recommendation the tool produces will be well-formed and it will be the wrong kind of well-formed for a hiring decision. Hiring is not a scoring exercise. It is a judgment about whether an adult you will now be responsible for is the right person for a seat you will now be responsible for filling well. Judgments of that shape do not survive being outsourced, no matter how fluent the outsourcing has become.

The exercise I ask owners to run on themselves

The exercise is short. Take three interviews from the last quarter, ideally recorded. Watch them back with the following in front of you and nothing else. First, mark your talk ratio in five-minute buckets. Second, count every time you filled a silence longer than three seconds. Third, write down the last question you asked in each interview. Fourth, note the candidate you decided to hire and the moment in the recording you can now see you had decided. Fifth, look at how many questions you asked after that moment and whether their character changed.

What emerges from this exercise is not a rating of your interviewing. It is a portrait of your executive attention. It shows the shape of what you find easy to look at, what you defend yourself from, what you close off prematurely, what you invite in when you feel safe. Every owner who has done this exercise honestly has come back and made two adjustments to the interview design and one adjustment to how she runs a meeting that is not an interview at all. The interview is the smallest room in the company. It is also the most concentrated. What the owner does inside it, she does at scale outside it. Diagnose the room and you have diagnosed the company. That is the interview’s highest use, and it has nothing to do with hiring.

The interviewer the tools cannot replace

A number of platforms now offer to conduct the first-round interview on the owner’s behalf. The candidate speaks to a screen, the screen scores the responses, the owner gets a report. For some roles, at some volumes, this may be efficient. For any role the owner cares about, it is a false economy. The first-round interview is where the owner most needs to see the adult in front of her, unmediated by scoring, and where the candidate most needs to see the actual human she may be about to work for. Outsourcing that meeting saves time and quietly reduces the accuracy of every subsequent decision, because every round after the first is now being built on a foundation the owner never actually inspected.

The exception, and it is a real one, is a very high-volume top-of-funnel screen for structured skills that can be assessed cleanly. There the tool is fine. Everything above that layer, and certainly every conversation with a candidate the owner would consider hiring, has to include the owner’s own attention in the room. Attention that has been delegated to a tool is not the owner’s attention. It is the tool’s attention with the owner’s brand on it, and candidates can feel the difference from the first thirty seconds of the second-round conversation, when they realize that whoever the tool reported them to has not actually met them yet.

A closing note to owners who are hiring this quarter

If you are about to run a serious hiring cycle in the next ninety days, do not spend the weekend refining your questions. Spend it watching yourself. Do the exercise above with three past interviews. Take the two adjustments it produces and put them in the design of the next cycle. You will hire a different person than you would have hired without the exercise, and the person you hire will be closer to the person the seat actually needs. Every serious upgrade to my own hiring has come from watching my own tape, not from adding a smarter question to the script. The candidate is the object. The owner is the instrument. Instrument work is the highest-leverage work an owner can do the week before an interview cycle, and almost no owner ever schedules it.