You are closing deals yourself and it is eating the week. The obvious fix is a salesperson, and a good one is available: nine years at a company you recognise, a book of contacts, keen to join something early.
Nine months later they are gone and everyone calls it a bad hire. The seat was built to fail before anyone sat in it, and the next person will fail in it too.
Why these three models
The decision is whether to hire your first salesperson now, and what to build first if not. The features that fire are a candidate pool whose availability carries information, a failure rate that is knowable and never looked up, and a structure that produces the same outcome whoever occupies it.
Three lenses. Selection produces an equilibrium answer about who is actually available to you. Base rates produce a cycle answer about how often this works and why. System traps produce a complex answer about the structure, which is the only one of the three that tells you what to do instead. The third is where the article ends up, and it is not the one founders expect.
1. Who is available, and what that tells you
Start with a question that feels ungenerous and is simply informative: why is this person available to you.
Excellent enterprise salespeople at established companies are expensive, compensated on a book they have already built, and rarely leave for a pre-product-market-fit company at a fraction of the earnings. That is not a comment on anyone’s character. It is what the market for that skill looks like, and it means the pool available to you is selected before you start interviewing.1
The selection runs in identifiable directions. Some candidates are available because they want to build rather than execute, which is exactly what you need and is rarer than the CVs suggest. Some are available because their last number was not made, which may be their fault or their territory’s and is worth establishing. And some are available because they have never sold without a brand behind them and do not yet know how much of their result was the brand.
The last group is the largest and the most expensive mistake, because their record is genuinely strong and it was earned in conditions you cannot offer. A salesperson whose meetings were taken because of the logo on their email will find that the same email gets no reply from your domain, and neither of you will have predicted it.
The pattern has a local shape worth naming. A salesperson who spent six years at a bank, a telecom or a large distributor in Kampala or Nairobi carries a record built on meetings that were granted to the institution. Procurement officers returned their calls because of who was calling. The same person at a two-year-old company finds that the introduction no longer travels ahead of them, and the part of the job they never had to do, getting into the room at all, is now most of the job. Their closing skill may be real and entirely intact. It is simply downstream of a step they have never had to perform.
The diagnostic question is direct and most candidates answer it honestly. Ask what they did when a prospect had never heard of their employer. If the answer is a story with specifics, you have found someone who has sold without cover. If the answer is that it did not really come up, you have learned what their record measures.
That question has an obvious weakness, which is that a prepared candidate can supply a story. Close it cheaply by asking for the name of that prospect and whether you may call them. The request is reasonable, it is rarely refused by somebody describing something real, and the hesitation is itself the answer. A candidate who has actually sold cold will usually volunteer the harder version unprompted, including the deals that did not close, because to them it is a war story rather than a test.
The contact book needs its own test, because it is the asset both sides overvalue. Contacts transfer as introductions, not as deals. In a market where the buying side is small enough that everyone knows everyone, an introduction is worth a first meeting and nothing past it, and the book was built while they were selling something else at a price point that may not be yours. Ask for five names and what each of those five currently buys. If they can answer, the book is real and partly transferable. If the answer is a category rather than a purchase, you are buying a phone list.
There is a cheaper way to run all of this than a salary. Engage one experienced seller part time, or on commission against a small retainer, for a defined ninety days against a named list of accounts. It costs a fraction of a hire, nobody reads it as a demotion, and it produces the one thing an interview cannot: evidence of whether that person can get a meeting on your name rather than their last one. Some strong candidates will decline the arrangement, and the decline is information too.
2. How often this works, and why it does not
The second lens supplies the number that should govern the decision, and the number is not good.
First sales hires at pre-repeatability companies fail often, and the failure is concentrated in a specific cause rather than spread across many. Ask any group of founders one stage ahead of you: of their first sales hires, how many were still there at eighteen months, and what happened. The answer is a base rate you can collect in an afternoon and it will be worse than your expectation, because your expectation was formed from this specific promising candidate rather than from the class.2
You are forecasting from the person in front of you, which is exactly the method that cannot see the pattern.
Ask that question carefully, because the first answer you get will mislead you. Founders describe a failed sales hire in personal terms almost every time: they were not hungry, they did not fit, they came from a big company. Those explanations arrive first because they cost nothing to produce. Push past them with a second question. What did the person spend their first ninety days doing, and who decided it. If the answer is that they were left to work it out, you have the structural version, and it will come back that way from most of the founders you ask.
One trap remains in the collection itself, and it is why the number people repeat is too kind. The founders you can easily ask are the ones still operating. Companies that were killed by this mistake are not in the room to answer, and the founder who is now employed somewhere else is not at the event where you are asking. So go and find two of those deliberately. They are not hard to identify, they are usually willing to talk, and their account carries the part the survivors have smoothed away.
Then decide in advance what the number would change, because a base rate collected after the offer is drafted is decoration. Write the rule down now. If fewer than half of the first hires you hear about lasted eighteen months, you will run the one-page test below before making an offer to anybody. A threshold agreed in advance is a decision. The same threshold discovered afterwards is something you will explain around, because by then you will have met the candidate and they will have been impressive.
The consistent cause, once you ask what happened, is not competence. It is that the founder had not established a repeatable motion, so the new hire had to discover the buyer, the message, the objection set and the qualification criteria at the same time as carrying a number. Discovery and execution are different jobs with different feedback loops, and the second is what the compensation plan was written for.
That reframes the decision. The question is not whether this candidate is good. It is whether the job exists yet, and that is answerable without reference to any candidate at all.
3. The structure that produces the failure
The third lens explains why replacing the person does not fix it, which is the observation that saves you the second failed hire.
Certain configurations produce the same outcome regardless of who is inside them.3 This one has a recognisable shape. The hire is asked to invent a motion, which takes quarters and produces no revenue early. They are measured on revenue from month three, because that is what a sales compensation plan does. The absence of early revenue is read as underperformance. The response is pressure, which pushes them toward whatever closes fastest rather than toward the repeatable segment. That produces a few unrepresentative deals, which teach the company the wrong thing about its market, and then they leave.
Everyone in that sequence behaved sensibly. The founder measured the thing the role is normally measured on. The salesperson optimised the number they were paid for. The structure did the rest, which is why the second hire into the same seat produces the same result and gets the same explanation.
The exits from this structure are specific rather than motivational. Separate the two jobs: hire for discovery with a discovery scorecard, meetings held with the right buyer type, objections documented, segments tested, and put revenue targets on a later horizon. Or do not hire yet, and instead close ten more deals yourself with the explicit goal of writing down what repeats. Or hire a second closer into a motion you have already proven, which is a different and much easier role to fill and to measure.
Expect the candidate to resist the discovery scorecard, and understand why before you concede. A good seller reads a role with no revenue target as a role with no upside and no exit story, and they are not wrong: a year of documented objections is hard to put on a CV. So pay for discovery like discovery. A higher base, a defined bonus on the milestones themselves, and a written commitment that the revenue plan starts when the motion is agreed rather than on a date. If you cannot afford that structure, you cannot afford the hire, and the middle option was the answer all along.
The scorecard itself does not need to be elaborate. Four lines will do: meetings held with the buyer type you named, objections logged in the prospect’s own words, segments tested and eliminated, and one written description of the motion updated monthly. Then review it monthly with the seriousness you would give a pipeline review. A discovery target nobody reviews is a revenue target with a delay on it, and everyone in the arrangement works that out by month two.
Watch for the specific tell that you are inside this structure rather than merely having a slow start. It is when the explanation for the absent revenue changes every month and each version is plausible. A record that has never been tested against a bad quarter tells you almost nothing about the person, and a record produced under conditions nobody has characterised tells you nothing about the role either.4 Two quarters of a first sales hire is not evidence about them, it is evidence about whether the job existed.
The middle option is the one most founders reject and it is usually correct, because the thing being bought with the hire is time, and the thing actually missing is knowledge that only the founder is positioned to acquire.
What the three say together
- Ask why this candidate is available, and what they did when a prospect had never heard of their employer.
- Collect the base rate from three founders one stage ahead: how many first sales hires lasted eighteen months, and what happened.
- Decide whether the job exists. If the motion is not repeatable, you are hiring for discovery and must measure it as discovery.
- If you hire, write the scorecard before the offer: meetings with the right buyer, objections documented, segments tested. Revenue later.
Where they disagree
The selection argument and the structural argument point to different remedies.
Selection says the pool is the problem, so solve it with search: pay more, look longer, find the rare person who has sold without cover and wants to build. Structure says the seat is the problem, so a better occupant changes little, and the money spent on a longer search would be better spent on the founder closing ten more deals.
The reconciliation depends on how far along the motion is. If you can already describe the buyer, the message and the top three objections, the job exists and the search argument governs: hire well, because the structure is sound and the variance is in the person. If you cannot, the structure argument governs and no candidate resolves it. Most founders believe they are in the first case, and the test is not an opinion: write the buyer, the message and the objections on one page in ten minutes. If you cannot, you are in the second case.
What none of them contain
None of the three prices what founder-led selling costs you elsewhere. The time going into deals is not free, and a founder who closes ten more deals personally has not built product, not raised, and not hired anyone else. The recommendation to keep selling is expensive and this article does not pretend otherwise.
None of them handles the market where the founder cannot sell. In some sectors the buyer will not meet a founder, or the sale requires a licence or a relationship the founder does not have. The advice to do it yourself first assumes you can, and where you cannot the correct move is a partner or an agent rather than an employee.
And one property the ensemble will not produce: a discovery hire changes what you learn. The founder selling learns things about the product that get built. A hired discoverer learns the same things and reports a summary, and the difference between experiencing an objection and reading about it is most of the value.
The one action that survives the ignorance: take ten minutes and write your buyer, your message and your top three objections on one page. If you cannot, you do not have a job to hire for yet, and the next ten deals should be yours.
Who has to move
The founder, and mostly by not hiring yet. The instinct when selling is consuming the week is to buy time, which is the correct instinct applied to the wrong constraint: the missing thing is knowledge rather than hours. The cheapest first test is the one-page write-up, which costs ten minutes and settles the question without any reference to a candidate. If the page is easy to write, hire with confidence and pay for someone good. If it is not, you have found the actual work.
Sources and notes
- Jeffrey Carpenter and Andrea Robbett, Game Theory and Behavior, MIT Press. Adverse selection, in which the population that accepts a given set of terms is determined by private information the other side holds, is developed in the chapters on asymmetric information. Used in section 1 for the claim that the candidate pool available at your stage is selected before the interviewing starts. The related point about attributing a record to a person rather than to their circumstances is a separate error and is treated in the same volume alongside the discussion of noisy signals.
- Sanjit S. Dhami, The Foundations of Behavioral Economic Analysis, Oxford University Press. Base rate neglect, and the tendency to substitute a vivid individual case for the underlying frequency, sit in the part on bounded rationality. Used in section 2 for the direction of the error only. No failure rate is quoted for first sales hires, because no reliable published figure exists at this stage and geography, which is why the article sends you to collect your own from three founders rather than citing one.
- Donella H. Meadows, Thinking in Systems: A Primer, Chelsea Green Publishing. The system traps, and the general finding that a structure produces characteristic behaviour regardless of which individuals occupy its positions, are set out in the chapter on system traps and opportunities. Used in section 3, including the corollary that replacing the person leaves the structure intact and reproduces the outcome.
- Michael J. Mauboussin, The Success Equation: Untangling Skill and Luck in Business, Sports, and Investing, Harvard Business Review Press. The difficulty of separating an individual’s contribution from their circumstances in a short record supports the caution in section 1 about a candidate whose results were produced under a strong brand. The specific application, that the record measures the conditions as much as the person, follows from the luck-skill framing rather than being a stated result.
A note on a number this article does not give. There is no revenue level at which the first sales hire becomes safe. The threshold is whether the motion is describable, not how large it is, and a company at a larger number with no repeatable motion is in more danger rather than less, because it can afford the mistake twice.
Joshua Agonya Pi’Rwot, Founder.