a builder's codex
codex · operators · Brian Halligan · ins_hire-spiky-not-un-weakness

Hire spiky candidates with real strengths and real weaknesses, not the un-weakness average

By Brian Halligan · Co-founder & former CEO, HubSpot; in-house CEO coach, Sequoia · 2026-04-28 · podcast · CEO craft, the LOCK(S) algorithm, distribution as the new moat

Tier B · TL;DR
Hire spiky candidates with real strengths and real weaknesses, not the un-weakness average

Claim

Hiring rubrics that reward "no major weakness" produce 3/4 candidates across every dimension and a low ceiling. Hiring rubrics that reward peaks — 4s and 2s — produce candidates with real strengths and real weaknesses. HubSpot's hit rate improved when they switched from un-weakness to spiky.

Mechanism

Capability is bottlenecked by the highest spike, not by the average. A team built of well-rounded 3s rarely produces breakthrough work. A team built of complementary spikes — each member a 4 or 5 in their lane and a 2 in another — covers the full surface through pairing. The hiring conversation also gets sharper because the rubric forces the team to name what specific spike they need, rather than rejecting candidates for soft "concerns."

Conditions

Holds when:

Fails when:

Evidence

"We used to hire candidates where everyone scored 3/4 — no major weakness. We switched to candidates with 4s and 2s — real strengths, real weaknesses. Hit rate improved."

Brian Halligan applied this at HubSpot across multiple hiring cycles before joining Sequoia as CEO coach.

— Brian Halligan on Lenny's Podcast, 2026-04-28

Signals

Counter-evidence

For roles where consistency is the value (compliance, finance ops, customer support), un-weakness wins. The spiky rubric is conditional on roles where peaks compound.

Cross-references

Open the interactive view → View original source → Markdown source →