{
  "schema_version": "1.1",
  "id": "amazon-100ms-latency-1-percent-sales",
  "slug": "amazon-100ms-latency-1-percent-sales",
  "url": "https://groundedkit.com/claims/amazon-100ms-latency-1-percent-sales/",
  "api_url": "https://groundedkit.com/api/claims/amazon-100ms-latency-1-percent-sales.json",
  "claim": "Every 100ms of latency cost Amazon 1% of sales",
  "commonly_cited_as": "\"Amazon found every 100ms of latency costs 1% in sales\"",
  "verdict": "unsourced",
  "verdict_definition": "No primary data found; the trail ends in a roundup, a disowned quote or a circular citation.",
  "confidence": "medium",
  "rating": {
    "value": 1,
    "best": 5,
    "worst": 1
  },
  "origin": {
    "publisher": "Greg Linden (ex-Amazon) blog post and slides, 2006, about a pre-2002 A/B test",
    "title": "No published data",
    "year": "2006 (anecdote about pre-2002 test)"
  },
  "quote": "In A/B tests, we tried delaying the page in increments of 100 milliseconds and found that even very small delays would result in substantial and costly drops in revenue.",
  "method": "No dataset, sample, test length or significance was published; the '1%' wording comes from a later presentation slide, as traced by the source.",
  "drift": [
    "Precision: 'substantial drops' became an exact 1% per 100ms",
    "Date: 20+ year-old test quoted as current"
  ],
  "interest": "n/a",
  "safe_rephrasing": "An early-2000s internal Amazon A/B test reportedly found small page delays reduced revenue; no data was published.",
  "sources": [
    {
      "title": "Niels-Ole Kühl: Amazon 100ms = 1% revenue? (2018 trace)",
      "url": "https://niels-ole.com/amazon/performance/2018/10/27/100ms-latency-1percent-revenue.html",
      "primary": false
    }
  ],
  "kind": "stat_check",
  "as_of": "2006 (anecdote about pre-2002 test)",
  "last_checked": "2026-10-16",
  "stale": false,
  "date_reviewed": "2026-10-16",
  "license": "CC BY 4.0",
  "license_url": "https://creativecommons.org/licenses/by/4.0/"
}