Methodology
How Pebbles.fyi keeps data realistic.
We focus on typical compensation ranges, not the extremes.
Typical pay framing
Every public view highlights the median, the 25th/75th percentiles, and the 10th/90th percentiles. This keeps the focus on what most companies pay rather than the top-of-market outliers.
Outlier detection
Submissions are screened using a robust median absolute deviation (MAD) check. Values outside a conservative band are flagged and excluded from aggregates. We never publish raw submissions.
Privacy and safety
We do not collect names, emails, exact teams, or offer letters. IP and device fingerprints are stored only as salted hashes for rate limiting.
Sample size thresholds
Aggregates are only shown when at least 5 submissions exist for the bucket. Confidence levels are based on sample size: 5-14 (Low), 15-49 (Medium), and 50+ (High).
Cost-of-living adjustment
We show an optional COL-adjusted median using a local index (US=100). The adjusted value is computed as median total divided by (COL index / 100). If we do not have a location index, we say so instead of guessing. The current map is a small placeholder until we wire in official RPP data.
Compensation math
Totals are annualized in USD. Bonus and equity are self-reported and can vary. Sign-on is included in total annualized to keep the math simple and comparable.