Case study — GE's vitality curve, 1981–2015 (and why every imitation failed)

If Netflix's culture deck is the most copied HR document of the last 20 years, the GE vitality curve is the most copied of the prior 20. Both produced the same lesson: technique without context destroys value.

What the system was

Each year, managers ranked all employees. Top 20% got disproportionate raises, equity, and promotions. Middle 70% got development. Bottom 10% were managed out. Welch credited it with making GE the world's most valuable company in the late 1990s.

Why it worked at GE

• Conglomerate of independent BUs — low collaboration needs, ideal for tournaments.
• External labor market deep enough to absorb cuts and supply replacements.
• Compensation skewed enough to make the top 20% prize worth the contest.
• Strong middle management bench — calibration was rigorous, not theatrical.

Why imitations failed

• Microsoft: high collaboration between product teams broke under stack-ranking; engineers gamed ratings, knowledge-sharing collapsed.
• Amazon: forced ranking + 'hire-to-fire' rotational programs created burnout and turnover.
• Yahoo: lawsuits from terminated employees challenged the methodology.
• Every imitation underestimated calibration cost and overestimated managerial accuracy in distinguishing top 20% from top 30%.

Lessons + discussion questions

1. When do tournament structures work? When do they break? (Reference tournament theory.)
2. What is the cooperation cost of your performance system?
3. Is your bottom-N driven by accountability or by ritual?
4. What replaced calibration discipline when you simplified the process?
5. What is your version of GE's 'PD@GE' continuous-feedback shift?