NNaval
← All frameworks
Leadership

The Random Interview Test

Let a candidate interview any teammate at random and judge whether they would still join.

Difficulty
Easy
Time to result
~days to results
Steps
5
Confidence
82%

The mechanism is a two-sided audit disguised as a recruiting offer. You tell the candidate to walk into the team room, pick anyone at random, pull them aside for thirty minutes, and interview them. If they are not impressed, they should not join. Because the pick is random rather than curated, you cannot stage the encounter with your strongest person, which is exactly what makes it informative. The output has two parts. The candidate gets an unfiltered read on the team and self-selects, which saves you a bad mutual match. More usefully, at the moment you make the offer you will instinctively flinch at one name in the back of your mind. That flinch is the real signal: it identifies the person who is capping the team's ability to mutually impress each other, and therefore the person to address.

Origin

Extracted from the Naval podcast episode Curate People, where it is described as one good test for maintaining the bar across the first cohort of directly managed hires.

Core principles

  • 01The best people only want to work with the best people.
  • 02Working alongside someone below your level is a cognitive load, not just an annoyance.
  • 03A team that mutually impresses each other is self-motivating.
  • 04Your flinch is data: the person you hope is not picked is the person to let go.
  • 05The bar applies hardest to the people the founder directly manages.

How to run it

  1. 1

    Make the offer unconditionally

    Tell the candidate they may walk into the room, pick anyone at random, and interview that person for thirty minutes. Do not pre-select who they meet.

    Pro tip The randomness is the whole test; a curated panel measures nothing.

  2. 2

    Give them permission to walk

    State plainly that if they are not impressed by whoever they pick, they should not join. This makes the offer credible rather than theatrical.

    Watch out If you hedge the instruction, the candidate reads it as a sales tactic and the signal is lost.

  3. 3

    Catch your own flinch

    As you make the offer, notice which teammate you quietly hope does not get picked. Write the name down immediately.

    Pro tip Do this before the candidate answers, while the reaction is still instinctive.

  4. 4

    Act on the name

    Decide whether that person can be moved into a role where they clear the bar or whether they need to be let go. Do not leave the name unresolved.

    Watch out Keeping that person is what prevents the team from becoming a group that all wants to impress each other.

  5. 5

    Re-run the test as the team grows

    Apply it again with each new cohort of people you will directly manage, since the bar drifts silently as headcount rises.

In the wild

The flinch that names the problem

A founder offers a strong engineering candidate the random interview. The candidate agrees, and in that instant the founder realises they are hoping the candidate does not pick one particular teammate, someone hired quickly during a crunch. Nothing about that teammate's performance had triggered a formal review, but the founder's own reaction has now made the judgement explicit. The founder treats the flinch as the finding rather than the candidate's eventual verdict.

→ The team's weakest link is identified in seconds by a test that was ostensibly about the candidate, not the team.

Common mistakes

Curating who the candidate meets

Selecting your strongest teammate for the conversation converts a diagnostic into a sales pitch and destroys the signal the test was designed to produce.

Rationalising the flinch away

Founders explain the flinch as a bad week or a communication style mismatch. The instinct is usually a compressed judgement built from months of observation.

Is it for you?

Best for

Founders and early leaders building the first 10 to 40 hires who need an honest, fast read on current talent density.

Not ideal for

Large organisations with formal HR processes where a candidate cannot be given free access to interview staff at random.

From the transcript

“walk into that room where the rest of the team is sitting, take anyone you want, pick them at random, pull them aside for 30…”

“When you do that test, you will instinctively flinch at the idea of them interviewing randomly a certain person that's kind of in the back…”

“the best people truly only want to work with the best people”

From the episode

Curate People

→