📚 Google Hiring Algorithm
Core Lesson: People analytics
📋 Overview
| Attribute | Detail |
|---|---|
| Subject | Data Analytics |
| Core Lesson | People analytics |
| Source | HBS / Top MBA Case |
🕰️ Background
Google (under Laszlo Bock) famously used data to debunk their own hiring myths. They analyzed tens of thousands of interviews and found that typical industry tools—brainteasers, GPA from elite schools, and ‘intuition’—had nearly zero correlation with actual job performance. They shifted to a data-driven model based on structured interviews and ‘work sample’ tests.
❓ The Central Problem
How do you remove human error from the most important decision: who to hire? Google’s ‘People Analytics’ team proved that even ‘smart’ people are terrible at judging talent without data.
📊 Analysis
Findings: (1) Brainteasers (‘Why are manhole covers round?’) are useless—they only test for the skill of solving brainteasers. (2) Grades matter only for new grads; after 2 years, they have zero predictive power. (3) Consistency: Having the same questions for every candidate (structured interviews) is the only way to compare fairly. (4) The Rule of Four: After four interviews, the marginal return on added data (more interviews) drops to near zero.
🔑 Key Lessons
- Intuition is a poor predictor of job performance; structured data is superior
- Standardizing the interview process is the only way to remove ‘unconscious bias’
- GPA and prestige are ‘lazy proxies’ that data often debunks
- Optimization doesn’t just mean ‘better’—it means ‘faster’ (e.g., stopping at four interviews)
🎓 Discussion Questions
- If GPA doesn’t matter, why does Google still attract mostly elite school grads?
- How do you scale ‘structured interviews’ in a rapidly growing company?
- What are the limits of ‘People Analytics’ when hiring for high-creativity roles?
🔗 Connected Concepts
- Psychological Safety — Hiring for culture-add vs. culture-fit
- Google Project Oxygen — Companion people analytics case
- Regression Analysis — Correlating scores with performance