Should a seed stage startup hire an in-house recruiter?

Usually not until hiring is continuous, meaning roughly six or more engineering roles a year. Below that volume you are paying a salary to keep a pipeline warm that there are not enough open roles to justify, and the work collapses into coordination a founder could do in an hour a week.

Are engineering recruiting agencies worth the fee for startups?

Only when the introduction is one the founders could not have made themselves. If the person was reachable through the existing network, the fee bought nothing. If the seat is on the critical path and the candidate was out of reach, the fee is small next to the cost of the role staying open.

How can you tell whether a recruiting firm is good?

By how few people they put in front of you. A short list with specific reasons for each person is a judgment someone made. A long list of profiles is a search handed back for you to filter, and the filtering was the work the fee was supposed to cover.

For founders

The best way to hire engineers at a seed-stage startup

There are four ways to fill an early engineering seat and they’re good at different things. How they compare, where each one fails, and where we fit.

Zelcast

At seed stage there are four real options: hire from your own network, source cold yourself, hire an in-house recruiter, or pay someone with relationships. Your network is the best and it runs out fastest. Cold sourcing works and it costs weeks of founder time. An in-house recruiter only makes sense once you’re hiring continuously. Paying for relationships pays off only when the introduction is one you couldn’t have made yourself.

Start with your own network, and know when it is empty

The first two or three engineers at most good companies come from people the founders already know, a pattern the Y Combinator library has documented across a lot of batches. It’s the highest conversion, the lowest cost, and the best information you’ll ever have about a candidate, because you’ve seen them work.

It also runs out fast, and founders keep drawing on it after it is dry. The signal is when the names you are considering are third order, people your friend thinks might be interested. At that point you are cold sourcing, only with worse targeting.

Before you move on, ask the strong engineers around you who the best person they’ve worked with is, whether or not that person is looking. That question refills the list once.

Cold sourcing

It works, and a lot of good teams got built this way. The cost is weeks of founder time, because senior engineers are employed and buried in identical messages, and reaching enough of them to produce three real conversations takes volume.

Choose this when you have more time than money, when your product is something engineers already admire, or when the role is specific enough that the target list is short and findable. Know the going rate before you start, which levels.fyi will tell you, and read why the messages go unanswered before you send five hundred. Avoid it when the seat is on the critical path for the next two months, because it is the slowest route on the page.

An in-house recruiter

The right answer once hiring is continuous. If you are filling six or more engineering seats a year, someone whose full time job is your pipeline will beat every alternative on cost per hire and on candidate experience.

At seed stage with one or two roles a year it’s usually wrong. You’re paying a salary to keep a pipeline warm that you don’t have the volume to justify, and a recruiter without enough requisitions ends up doing coordination work a founder could do in an hour a week.

Paying for relationships

This is the case for an agency, and it’s a narrow one. What you’re paying for is someone to reach a person you couldn’t reach yourself, and to be right about whether that person is any good.

There is one test, and it applies to us as much as to anyone else: was the introduction one you could have made on your own? If yes, you overpaid regardless of the fee. If no, the fee is small next to what the seat costs while it stays empty.

The failure mode to watch for is volume: a firm that sends ten profiles has described a search and handed you the filtering they were paid to do. The useful version sends a small number of people with a real argument for each.

This is what we do: we work only with startups under a hundred people, the fee is paid when someone is hired, and there’s no charge to talk or to give us a brief.

What we’d recommend

Exhaust your network properly, including the referral question above. It costs nothing and beats everything below it.

If the seat is not urgent and your product is a draw, source cold and accept that it takes five to seven weeks.

If the seat is on the critical path, borrow someone’s relationships, whether that is an investor who will make introductions or a firm that does this for a living, because what you’re buying is speed.

Hire a recruiter in house once the volume is real.

Common questions

What is the best way to hire engineers at seed stage?: Your own network first, including asking good engineers who the best person they have worked with is. After that the choice is between cold sourcing, which costs weeks, and paying for relationships, which costs a fee.
Should a seed stage startup hire a recruiter?: In house, usually not until you are filling six or more engineering roles a year. Before that the volume does not justify the salary.
Are recruiting agencies worth it for startups?: Only when the introduction is one you could not have made yourself. Apply that test to us as well as to anyone else.
How do I know an agency is any good?: By how few people they send. A small number, each with a real argument, is a judgment, and ten profiles is a search you’re being asked to filter yourself.