A Startup Doesn’t Need the “Best” Candidate. It Needs the Right Risk.

Most hiring processes are built to minimize risk. Add another requirement. Look for someone who has already done the exact job, at the exact scale, ideally in the same vertical. Prefer a recognizable logo on the CV. The more boxes a candidate ticks, the safer the decision looks — and for a company with an established playbook, that instinct is usually right.

For startups, it breaks down. A candidate who has done the exact job before may have done it in an environment so different that almost nothing transfers. A candidate with an obvious gap on paper may have precisely the capabilities the company needs this quarter.

So the useful question for a founder is rarely "Who has the strongest profile?" It is: "What are we betting on with each person, and which of those bets can we actually afford to lose?"

A 40-person startup with 14 engineers is hiring its first Head of Engineering.

Candidate A spent seven years at a large tech company and managed 80 engineers at serious scale. She has already seen the problems the startup hopes to have in two years. But she has never recruited without a dedicated talent team, never built process from nothing, and hasn't been in the codebase putting out operational fires in years.

Candidate B leads 12 engineers at an early-stage company, hired most of them personally, and works comfortably in ambiguity. He can do this job at this exact size — month one looks like month zero. But he has never managed 50+ engineers or built a middle-management layer.

Treating Candidate A as automatically more senior — and therefore safer — misses what's happening. The choice isn't senior versus junior. It's context risk versus growth risk, and those are not the same price.

Ben Horowitz made the underlying point in The Hard Thing About Hard Things: the work itself is different. In a mature organization an executive inherits people, processes and an operating system that already exists. In a startup, a substantial part of the job is creating those things.

Jack Altman, co-founder and then-CEO of Lattice, offers the practical version: hire someone who will be exceptional for the next 18 to 24 months, not the next five or ten years. He warns specifically against hiring "too big" — an impressive operator whose style belongs to a stage you haven't reached. Startups do need people who can grow ahead of them. But there's a meaningful difference between hiring one stage ahead and hiring three.

And it's worth knowing what the losing side of that bet costs. A Center for American Progress analysis pooling 30 case studies put the typical cost of replacing an employee at around 21% of annual salary — but for senior and executive roles, as high as 213%. Roughly two years of compensation to undo a single hire, and that's only the measurable part. It doesn't count the quarter the team spent building toward a direction the company later abandoned.

Hiring is uncertain, so everyone leans on shortcuts: a tier-1 logo for quality, industry experience for ramp speed, years for competence, a previous title for leadership. These aren't worthless — they're compressed information, and that's useful when you're screening 200 profiles. The failure mode is that they survive long after you've decided to stop using them.

A 2024 study by the Burning Glass Institute and Harvard Business School's Project on Managing the Future of Work looked at large US employers that had publicly dropped college degree requirements. Across 11,300 roles, hiring barely moved — the net effect came to fewer than one in 700 hires. Different world from yours, same mechanism: removing a requirement from a job description is much easier than removing it from the mental model used to evaluate candidates. In a startup, where a small number of people make most hiring decisions, those assumptions carry even further.

Startup hiring has its own entrenched shortcuts: "ex-Google," "previous exit," "fintech background," "managed 100+ people." Each can be informative. None is evidence that this person can solve the specific problem the role exists to solve.

Across our last 30 closed searches — roughly the past six months — exactly half of the candidates ultimately hired would have been screened out by the client's own original brief. Not a revised brief: the one written before the search started, must-haves intact. And there's no pattern in which requirement gave way — the vertical in one search, years of experience in another, team size in a third. The problem isn't that founders pick the wrong filters. It's that filters get written before anyone has seen who's actually available.

That doesn't mean the brief was wrong. It means the search produced better information than the company had when the brief was written.

Stop sorting finalists into pros and cons. Sort their gaps by type, because the types carry different costs — and show up on different timelines.

Capability risk. The candidate has never demonstrated a skill the role fundamentally requires — hiring a VP of Sales to build an enterprise motion when they've only run inbound transactional sales. The expensive one. It rarely resolves on its own.

Context risk. The capability exists but was built elsewhere — a director from a large company moving to a 40-person team. The question is whether the skill transfers when the supporting infrastructure disappears. Tends to surface early, which makes it easier to watch for.

Growth risk. The candidate has succeeded at your current stage but not the one you're moving toward. The question is whether they can grow fast enough to deliver the outcomes you're hiring them for, and how much support that will take. Unlike context risk, this one is deferred — which is exactly why it gets underweighted in the debrief.

Learnable gaps. A different stack, an adjacent vertical, unfamiliar tooling. When the underlying capability is strong, these are often the cheaper risks on the list — and they routinely consume the largest share of debrief time.

Risk is never priced in isolation — it depends on who else is in the building. A first-time VP of Engineering is a reasonable bet when the founder is a seasoned technical leader who can backstop the first hard call, and a serious gamble when nobody on the leadership team has built a software organization before.

So the question is not only "What is this candidate missing?" but "Do we already have that capability somewhere, and is that person actually available to absorb it?"

Most interview processes assess candidates broadly, even when the hiring team already knows where the uncertainty is. The central risk then stays untested until after the person joins — by which point you've paid recruiting costs, relocation, a notice period, and two months of a role sitting half-filled.

Once you've named the gap, design one step of the process around it:

None of this eliminates the risk. It converts it from a guess into a priced bet.

Before extending an offer, make the hiring team write down answers to five questions. Twenty minutes, and it surfaces the disagreements that would otherwise show up in month three.

Question four is the one teams skip, and it's the one that matters. A risk you can't absorb isn't a bet. It's a hope.

Fill one memo out per finalist and read them side by side. The purpose isn't to decide who has fewer weaknesses. It's to see which weakness costs most if the assumption behind it turns out to be wrong — which often produces a very different answer from asking who looks strongest overall.

Get the template. We've put this into a one-page memo you can fill in — the five questions with space for two finalists side by side, the four gap types as a reference, and prompts for the parts teams tend to rush. Reply to this email or write to a.sekachev@clever-recruiting.com we'll send it over.

None of this means brand and pedigree are always traps.

If you're raising in six months, a recognizable name on the leadership page does real work in a data room. If you're selling into enterprise, a VP who has closed those accounts brings a network you can't build in a year. If the founder has no functional background in the area, a seasoned operator with a known playbook is often the lower-risk choice — because there's nobody to catch the first-timer when they fall.

The point isn't that unconventional candidates are better. It's that "obviously safer" is a conclusion, not a starting assumption — and it should survive the same five questions as everyone else.

The strongest startup hire is rarely the candidate with no visible gaps. It's the one whose strengths match what the company needs now, and whose gaps represent risks the organization has named, tested and is equipped to carry.

At Clever Recruiting, this is how we think about finalist decisions: not eliminating uncertainty, but making the trade-offs explicit before the offer is signed. If you're weighing a hire like this right now, we'd be glad to talk it through.

Subscribe if you'd like them in your inbox. No generic hiring tips — just the trade-offs, failure modes and decisions that shape real searches.