How these benchmarks are built
Every number on this site is an anonymized, cross-brand median computed from Triple Whale's customer base. Here's exactly where the data comes from, how we protect each brand's privacy, and how every metric is defined.
Where the data comes from
Benchmarks are aggregated from thousands of ecommerce brands connected to Triple Whale. The raw grain is one row per shop, per day, per channel for ad metrics, and per shop per day for order and retention metrics. Shops are classified into cohorts by industry, annual revenue band, average order value band, and ad channel (Meta, Google). All monetary values are normalized to USD before aggregation.
How we anonymize it
No individual brand, store name, or account is ever exposed. We only ever publish aggregate statistics across a cohort, and we enforce a k-anonymity floor: a cohort must contain at least 50 distinct brands before any number is shown.
If you slice a cohort so narrowly that it falls under that floor, we automatically roll up to the nearest broader cohort and tell you we did — we never display a statistic derived from a handful of stores.
Why the median, not the mean
Every headline value is the cohort median (the 50th percentile) — the value a typical brand actually sees. Ecommerce performance is heavily skewed: a few huge accounts or a few near-zero ones would drag a mean away from reality. The median is robust to those outliers, so the benchmark reflects the middle of the pack rather than the loudest brands. The percentile bands (10th, 25th, 75th, 90th) describe the full spread so you can see where any result lands.
Freshness
A daily pipeline recomputes every cohort over a rolling 90-day window. Each page shows the date of its most recent refresh. Because the data updates daily and reflects a 90-day trailing window, small day-to-day staleness is expected and harmless.
Metric definitions
Every metric, its plain-English meaning, and the exact formula we use.