How the Best Startups Hire Without Headhunters

Startups that hire well without headhunters keep three things in-house — the brief, the bar and the close — and buy the rest deliberately, role by role. They have not banned agencies. They have stopped using them for work that repeats. The reason is not fee arithmetic: your first 20 hires set the standard every later hire is judged against, and if an agency made those 20 calls, that standard lives outside your company and walks out when the contract ends.

I ran a recruitment agency before I built Pickr, and a good share of my clients were Series A companies. The mandates I was proudest of were the ones nobody in-house could have run: a niche market I had spent four years mapping, a replacement we could not advertise while the incumbent was still in the chair. The ones I am least proud of were the repeats — the third mid-level account executive at the same company, at the same fee, because nobody there had built the process the first two should have taught them.

Why your first 20 hires matter more than the fee

A Series A company often makes 15 to 30 hires in the 12 months after the round, and the round itself usually buys 18 to 24 months of runway. Hiring is the largest line in that budget, and those people are not just headcount. They become the interview panel for everyone who comes after them.

Hire 30 is assessed by hires 6, 11 and 17. Whatever standard produced those three propagates through the company for years, through people rather than through documents. That is the mechanism nobody puts in the business case, and it is why the build-versus-buy decision at Series A is a different decision from the same one at 400 people.

A contingency recruiter screens out most of a pipeline before the founder sees anyone. That work is real and often good. But the reasoning behind it never enters your company. You inherit the outcome — five people on a shortlist — and none of the judgement that produced it. Two years later someone asks why the engineering bar slipped, and there is nothing to look at: no written reason for any rejection, no record of what strong looked like in month 4 versus month 18.

There is also an incentive worth naming plainly, and I say this as someone who was on the other side of it. A contingency agency is paid when the person starts, not when the person is still the right hire 18 months later. That is the contract, not a moral failing. But the candidate being optimised for is the one who passes your interview, and that is not the same as the one who is still right at Series B.

What a founder should never outsource

The brief. A job description is a marketing document. A brief is a decision document: the three to five things this person must be able to do in their first year, the evidence that would prove each one, and what you are willing to trade away. Only the founder knows which trade-off is survivable at this stage — whether you can take someone who has never managed if they have shipped the hard thing. Hand that over and you get a wish list back, then six weeks of shortlists against it. The mechanics are in how to write a job brief.

The bar. The bar is not a document, it is a set of worked examples: this candidate got an offer and here is the specific evidence, this one did not and here is exactly why. Two sentences per decision, written at the moment the decision is made, not reconstructed in a quarterly review. Do it for 20 hires and a new engineering manager can read what you actually valued. Skip it and every new interviewer invents a bar and calls it instinct.

The close. At Series A the founder is the product. A candidate leaving a bigger salary and a safer logo is betting on you specifically, and no external recruiter can make that argument on your behalf. I have watched good agency processes fall apart in the final week because the founder stayed out of the last conversation on the grounds that they were paying someone else to have it.

None of that is free, and the cost is not money. Two founders running 15 hires in a year will each spend something like 6 to 8 hours a week on hiring, and in Austria and Germany, where notice periods of 1 to 3 months are normal, the calendar stretches further than anyone plans for. That time is the real price of keeping the work inside, and it is the line nobody writes down.

What a Series A startup should build and what it should buy

Keep with the founders	Worth paying for at this stage
The brief and which requirements are tradeable	Market mapping in a function or country where you have no network
The written reason behind every yes and no	One confidential or genuinely scarce senior search
Final-round judgement and the close	Overflow in the month after the round, when eight roles open at once
The relationship with everyone you rejected	Scheduling and coordination once interviews run daily

The rule under the table: buy network and capacity, build process. A VP of Engineering search where the qualified pool across Austria and Germany might be a couple of hundred people, most of them not looking, is a network problem, and networks take years to build. A first hire in a country where you have no employer brand and no feel for the salary bands is the same problem wearing a different hat. The second backend engineer and the fourth account executive are neither. Paying a fee to solve a process problem means you never build the process, and the same invoice comes back next quarter.

The option most founders never consider sits between the two columns: embedded capacity instead of per-placement fees. Someone running your process, in your system, against your brief, whose output stays with you when the engagement ends. That is the case for recruiting as a service over contingency — you are buying hours and a process rather than renting judgement you never get to keep.

When to stop paying agency fees and hire your first recruiter

Two signals, and both are arithmetic rather than sentiment.

The fee run-rate crosses a recruiter. Contingency fees run 20 to 30 percent of first-year salary across most of the European market. Three hires in a quarter at a mid-market salary, at 25 percent, is a fee run-rate that annualises past a fully loaded recruiter, a recruiting system and a year of job advertising combined. The crossing happens quietly, because fees arrive one hire at a time and never look like a salary line. Plenty of companies spend more on agencies over 12 months than a recruiter costs while telling themselves they cannot afford one. The full arithmetic and the sequence for building the function are in the in-house playbook.

The same roles keep repeating. If you have paid a fee twice for the same title, the third one should be yours. By then you have two briefs that worked, two panels who know the role, and — if you kept the records — the candidates who lost narrowly to the last two hires and are probably still in the market. Almost nobody goes back to them, because in most systems the archive is unreadable six months later.

How to keep the hiring standard inside the company

Pickr is the AI-native recruiting platform that scores every candidate on evidence of skills rather than keyword matches, and feeds what happened to the people you actually hired back into how the next candidates are evaluated. For a company making its first 20 hires, three consequences matter more than any feature list.

Interviews are transcribed and scorecards arrive pre-filled with evidence mapped to each criterion, so an interviewer edits a draft the same afternoon instead of facing an empty form days later — the only realistic way founder-led hiring produces a written bar at all. Interviewer and hiring-manager seats are free, which matters here because everyone interviews; the moment feedback costs a licence, someone caps the seat count and the evidence goes back into hallway conversations. And candidates are scored on demonstrated and adjacent skills rather than exact keyword matches, which matters most when you have no employer brand and cannot afford to filter out the person whose CV uses different nouns for the same work. What that looks like at this stage is set out on the page for startup hiring teams.

Two honest limits. The outcome loop needs outcomes, and in year one you have none: candidates are scored against the brief you wrote, not against your own hiring history, because that history does not exist yet. And if you are making three hires in the next 12 months, none of this pays for itself — a shared document and the discipline to write two sentences per decision will beat any system, including mine. I built Pickr, so discount these two paragraphs accordingly. Everything above them stands without it.

Candidate data is hosted in Frankfurt, Germany, Pickr is built in Austria, it is GDPR compliant with a data processing agreement included, and personally identifying information is redacted from AI prompts by default. Pickr can connect to your current system read-only for a free recruiting audit before you commit to anything, and import your hiring history if you move across. If that audit says your problem is an unowned stage or a brief nobody ever agreed on, no agency fixes that either.

The takeaway

Agencies are worth paying for network and capacity, and a waste of money for process. At Series A the risk is not the fee, it is that the first 20 evaluation decisions — the ones that set the bar for the next hundred hires — happen outside the company and leave no trace inside it. Keep the brief, the bar and the close. Buy the rest when it genuinely is a network problem, and stop the moment the same role comes round a third time.