Skip to content
Playbook
IntermediateFounderEngManagerTechLead

Broken HR advice #6: 'Only hire 10x rockstars'

The 10x engineer is real and the 10x hiring strategy is broken. Why elite-only orgs underperform balanced ones.

7 min read
On this page
60-Second Summary
  • Individual performance variance in software is wide — but team output is bottlenecked by the worst-collaborating member, not lifted by the best one.
  • All-senior teams hit context-decoherence: too many decisions, no one to absorb routine work, ego friction on small calls.
  • Healthy teams are barbell-shaped: a few high-leverage senior contributors, a strong mid-level core, and 1–2 ramp-stage juniors.

Sutton & Allen's 1968 study about programmer productivity variance has been doing damage for almost 60 years. The variance is real. The strategic conclusion 'so hire only the top variance' does not follow.

The 10x myth, misread

  • All-senior teams produce more conflict per decision and slower decisions overall.
  • Senior engineers leave 18–22% faster from all-senior teams (DORA / Accelerate research) — they want leverage, not parity.
  • The 'rockstar' framing also tends to attract a specific personality profile that scales worse than a balanced team scales.

The barbell team

The empirical sweet spot for a 6–9 person product team: 1–2 staff/principal engineers, 3–4 mid-level engineers, 1–2 junior engineers being ramped, with a strong tech-lead manager. This composition outperforms an all-senior team on lead-time-for-changes and on retention in DORA's longitudinal data.

What 'rockstar' should actually mean

An engineer whose presence raises the performance of everyone around them. By that definition, most 'rockstars' aren't — they're high-output individuals who depress team output.

Written by Pawan Joshi.Sources cited inline.
First published 1 Oct 2024See site changelog →