People analytics
People analytics applies data analysis to workforce questions — who's leaving, what hires perform best, where the funnel breaks. The discipline behind serious HR decisions; AI is one tool, not the whole thing.
What questions people analytics actually answers
For SMBs, the useful and answerable questions are practical:
- Where are we losing candidates in the funnel? Stage-by-stage conversion data points at the leak.
- Which sources produce the best hires? Source-of-hire cross-referenced with first-year retention.
- What’s the time-cost of our hiring process? Hours per hire by role, panel, and stage.
- Who’s leaving, and at what tenure? Attrition by year, by team, by hire-source.
These don’t need AI. They need clean data and a willingness to look at it.
Where people analytics breaks
Three failure patterns:
- Vanity dashboards: 20 charts, no decisions. The team builds and admires; nothing changes.
- Over-fitting to noise: an SMB with 30 hires/year does not have statistical power to draw most conclusions. Pattern-matching on small samples produces false signal.
- Privacy creep: employee monitoring (keystroke logs, productivity scores) marketed as “people analytics.” Different beast; high GDPR risk; demoralizing.
The honest SMB scope
For a 50-200-person company, people analytics is well-scoped at:
- Three or four hiring-funnel KPIs.
- One quarterly retention review.
- One annual deep-dive on a specific question (e.g., “why did sales hires turn over more this year?”).
More than that requires headcount the data doesn’t justify.