{
  "slug": "coefficient-of-variation",
  "term": "Coefficient of variation",
  "aliases": [
    "CV"
  ],
  "definition": "The coefficient of variation expresses standard deviation relative to the mean, usually for ratio-scale data with a meaningful zero.",
  "note": "The sample CV is s / mean, where s uses the n - 1 variance denominator. It is dimensionless; multiplying by 100 expresses it as a percentage for positive means. A zero mean makes the ratio undefined, and near-zero means make it unstable. [nist-cv]\n\nSynthetic latencies [1, 2, 3] ms have sample s = 1 ms and CV = 1/2. Adding 100 ms to each observation leaves s unchanged but gives CV = 1/102. Converting milliseconds to microseconds preserves CV; adding a delay does not. Lower relative spread alone is not lower latency. [nist-cv] [hesela-spread]\n\nCV summarizes spread relative to a mean, not the standard error of that mean or an upper latency percentile. Comparisons still require matching measurement boundaries and request populations. [nist-cv]",
  "related": [
    "median-absolute-deviation",
    "quantile",
    "tail-latency",
    "standard-deviation"
  ],
  "citations": [
    {
      "id": "nist-cv",
      "title": "NIST Dataplot: Coefficient of Variation",
      "url": "https://itl.nist.gov/div898/software/dataplot/refman2/auxillar/coefvari.htm",
      "accessed": "2026-11-19"
    },
    {
      "id": "hesela-spread",
      "title": "Hesela: robust and relative latency spread, fixed arithmetic and reproducible code",
      "url": "https://hesela.com/analyses/storage-latency-coordinated-omission/#robust-relative-spread",
      "accessed": "2026-11-19"
    }
  ],
  "status": "research",
  "advantages": [],
  "limitations": [],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-11-19"
}
