Clarity free-tier sampling bias when 100k session cap is reached mid-month

Sampling behavior at the monthly session limit

Microsoft Clarity's free tier captures up to 100,000 sessions per month per project before probabilistic sampling activates. Heatmap aggregation continues across all sessions, but session replay availability becomes incomplete once the cap is exceeded.

Unresolved behavior

Documentation does not specify whether the sampling algorithm applies uniformly across all traffic sources or introduces bias toward newer sessions when the cap is reached mid-month. This ambiguity affects trend analysis reliability during capped periods, as analysts cannot determine if observed patterns reflect true traffic distribution or sampling artifacts.

Constraints:

• Free tier only; no paid tier exists to increase caps or extend retention as of 2024
• 30-day recording retention; 12-month heatmap retention
• Sampling indicator shows a percentage in the dashboard when approaching the limit
• Behavior may change without notice; no SLA or export guarantees

Questions

1. Does Clarity's mid-month sampling distribute probability uniformly across all sessions, or does it weight recent sessions more heavily?
2. Can the sampling ratio be inferred from the dashboard percentage indicator, or is that metric itself derived from the sampled subset?
3. Are there documented workarounds to maintain consistent trend analysis during capped periods without upgrading (since no paid tier exists)?