The Riff Test: Hiring for Gut Instinct
Pay finalists to build something, then critique it and ask what they'd do with two more days
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 92%
Fried hires designers and product people by giving the last five finalists a paid, week-long project, then critiquing the result live — pushing back even on work he loves — and asking what they would do with a couple more days. The unrehearsed answer reveals whether the person has a gut full of ideas and the willingness to lean into it, which is what he is actually hiring for.
Origin
Jason Fried's own hiring process for designers and product people at 37signals; he personally hires designers/product while co-founder David Heinemeier Hansson hires programmers.
Core principles
- 01Craft first: the submitted work must be clear, well-executed and thoughtful, or nothing else matters.
- 02Ideas without execution is a no; execution without ideas is also a no.
- 03What you can't rehearse is what you learn from — riffing on the spot exposes the raw gut.
- 04Taste is diagnosable: what someone admires tells you the lens they see the world through.
- 05You are looking for synchronicity — the moment two people riffing lock into the same wavelength.
How to run it
- 1
Probe taste before skill
In interviews, ask what products they like, whose work they admire, what they have seen that they think is great — and then keep asking why. You are trying to get behind the words to where the opinion comes from.
Pro tip 'Whose work do you admire?' is more revealing than 'what can you do?' — and you don't care who they name, you care why.
Watch out Candidates can prepare polished answers. Chase the Genesis of the opinion, not the opinion itself.
- 2
Give the final five a paid week-long project
Take the last five finalists, pay them, and give them one week to produce a real piece of work. This is a work sample, not a puzzle.
- 3
Critique it live — even the work you love
Sit with the candidate and critique the submission, deliberately pushing back in certain areas even if you think it is excellent. You are creating live pressure, not scoring the artifact.
- 4
Ask: what would you do with another couple of days?
The core question. They have no time to prepare, so whatever comes out is the gut. You want to see that the submitted work was a slice of what they had in mind, not the total exhaustion of their ideas.
Pro tip The direction they go barely matters — what matters is that they go somewhere and are unafraid of saying things that may not make sense yet.
Watch out Some strong people genuinely need time to think. Don't treat one hesitant moment as disqualifying on its own.
- 5
Look for the riff, and for synchronicity
Start riffing back — 'what if we did that, what about this'. The signal is a resonance, two wavelengths vibrating together. When you feel it, you have something.
In the wild
Some candidates 37signals thought were really good submit a strong project, but when asked what they would do with two more days, they have nowhere to go. The submission was everything they had.
→ Fried reads this as an empty gut and passes, despite the artifact being good.
Other candidates arrive with a million ideas but the paid project comes back a mess — poor time management, focus on the wrong things.
→ Also a no. The bar is a clever design, well executed and clear, PLUS a live riff on where it goes next.
Common mistakes
Hiring for ideas alone
Fried explicitly does not want 'just ideas'. If the paid project comes back disorganized, the ideas are worthless — the person cannot ship.
Interviewing without a paid work sample
Without a real artifact to critique, there is nothing concrete to riff off, so the on-the-spot gut test has no ground to stand on.
Hiring the academic product person
Someone who reasons purely from process and evidence can be excellent elsewhere, but in a gut-driven, small, taste-led company they will not function.
Is it for you?
Best for
Founders and design/product leaders at small, taste-driven companies who hire a handful of people a year and need judgment, not credentials
Not ideal for
High-volume hiring, roles where output is standardized, or companies that cannot legally or practically pay candidates for trial projects
From the transcript
“they do a project for us at the end so the last like five finalists will do a project for them for us we'll pay…”
“want to see where people go with that to me that is where you can see they don't have any time to think so like…”
“like that's a clever design well executed thoughtful clear and here's where they could take it from here on out if they had a few…”
“when I hire designers I'm I'm curious about their taste I'm curious about what they like I'm curious about what other products they think are…”
From the episode
Jason Fried challenges your thinking on fundraising, goals, growth, and more