Finding Great Candidates as an Unknown Startup

Build three channels that compound rather than one that scales: structured referral requests, direct sourcing from a maintained list of adjacent companies and communities, and public technical work that brings people to you. Compete on scope and decision rights rather than on cash, and be explicit about which one you are offering.

Channels that compound versus channels that scale

Most advice on how to recruit talent for a startup starts with channels. Start instead with what compounds, because an unknown company cannot outspend known ones and should not try.

A job board posting produces applications while it runs and nothing afterwards. It scales with spend and resets to zero when you stop. That is fine as a supplement and dangerous as a foundation.

Compounding channels get better with use. A referral network deepens as your team grows and as people you did not hire remember the conversation. A maintained sourcing list improves every time you add context about who works where and who is likely to move. Public technical work accumulates: something written a year ago still brings candidates today.

Measure qualified conversations by channel, not applications. Volume from a board can look impressive and convert at nearly zero for specialised roles.

Look where the work is visible

The candidates worth most are frequently not searching, which means profile-based search finds the wrong population. It finds people who maintain profiles.

Work-based search finds people through what they have produced: open source contributions, technical writing, conference talks, substantive answers in communities, and the people your own strongest engineers cite when explaining how they learned something. That last one is underused and unusually high yield, because it is a referral that does not depend on the referrer knowing the person personally.

Senior people leave for scope more often than money

The assumption that senior talent will not join an early-stage company usually reflects a targeting error rather than a market truth.

Many senior engineers at large companies are constrained by their inability to ship, not by compensation. Four approval layers, a quarter to get a change into production, and no contact with the people using what they build. If that is the constraint, an early-stage company is offering the solution rather than asking for a sacrifice.

Lead with it concretely. Not "you will have impact" but "you would own the billing system, decide its architecture, and ship to production in your first week." Specificity is the differentiator, because every company claims impact.

Brand is evidence of thinking, not culture claims

A candidate assessing an unknown company is trying to answer one question: are these people good at what they do? Culture pages cannot answer it because every company writes the same one.

Published thinking can. A written architecture decision, an honest post about something that went wrong, a role brief that names the hard parts of the job. All of it is verifiable in a way that adjectives are not, and it selects for candidates who care about the same things you do.

Be explicit when cash is below market

Candidates discover compensation eventually. Discovering it late, after investing in a process, produces withdrawal and reputational cost.

Saying it early costs you the candidates who were only ever going to decline, which is the point. Name the number, name what is above market instead, and let them decide. This converts better than a vague equity narrative, and it is the same behaviour that makes strong candidates trust the rest of what you tell them.

Common questions

How do I build a sourcing pipeline at scale?

Build three channels that compound rather than one that scales: structured referral requests, direct sourcing from a maintained list of adjacent companies and communities, and public technical work that makes people come to you. Job boards add reach but do not compound, so they should supplement rather than carry the pipeline.

How do I find exceptional talent that doesn't exist on LinkedIn?

Look where the work is visible rather than where the profiles are. Open source contributions, technical writing, conference talks, community answers, and the people your strongest engineers cite. The best candidates are frequently not searching, which means they are reachable through work rather than through profiles.

How do I find senior talent that's willing to join an early-stage company?

Target people whose current constraint is scope rather than money. Senior people at large companies frequently leave because they cannot ship, not because they are underpaid. Lead with the scope and the decision rights, and be specific about what they would own within a week of joining.

How do I build a strong employer brand with zero brand recognition?

Publish the thinking rather than the culture. Candidates evaluating an unknown company want evidence of how you make decisions, and technical writing, public architecture notes, and transparent role briefs supply that. Culture pages supply nothing they can verify.

What's a job description that actually attracts top talent to a seed stage startup?

One that names the problem, the constraint, and what the person will own, and is honest about what is unresolved. Strong candidates are attracted by a well-specified problem and repelled by a list of requirements. Say what is broken and what deciding it well would be worth.

How do I recruit when I can't offer top market salary?

Compete on the things large companies structurally cannot offer: scope, decision rights, direct customer contact, and speed from idea to production. Be explicit that cash is below market and name what is above market instead. Candidates discount vague equity stories and respond to concrete ownership.

How do I hire people who understand AI and organisational design?

Look for people who have automated part of their own job and can explain what they chose not to automate. That combination indicates judgment about where systems help and where they do not, which is rarer and more useful than tool familiarity.

How do I recruit a super IC when I'm a seed-stage startup?

Offer the thing they cannot get elsewhere: an outcome they own entirely, no approval chain, and a tooling budget. Seed-stage companies lose these candidates on process, not on pay. If your interview loop already feels bureaucratic, they will assume the job does too.

How do I hire someone who can be a super IC in 18 months?

Select for people who already scope ambiguous work without being handed a spec, then remove what blocks them from shipping end to end. Growing one internally has a higher hit rate than hiring one, because you are selecting on demonstrated behaviour rather than an interview signal.