Free interactive tool
Match Rate Calculator
The match-rate calculator computes a visitor-identification match rate the way you would defend it in a renewal meeting: identified sessions divided by a denominator you explicitly chose and labeled (all sessions, filtered human sessions, target-market sessions, or high-intent sessions), then adjusted for false positives by multiplying by the precision of a ground-truth sample — because the usable match rate, not the raw one, is the number your alerts and routing rules actually run on. It also sizes your next precision sample, so you know how many identified sessions to score before the estimate is worth quoting. Installing the tag first? The platform install guides give the exact code location and the gotcha for every major platform.
The formula
raw match rate = identified sessions ÷ denominator sessions
precision = confirmed ÷ (confirmed + wrong)
usable match rate = raw match rate × precision
next sample size ≈ 1.96² × p(1−p) ÷ e² (95% confidence,
corrected for the identified-session pool)Sessions marked plausible-unverified count in neither the confirmed nor the wrong bucket — they are tracked separately, and when they exceed half the sample your ground-truth lists are too thin to score the tool yet. The denominators and the sampling method come from the methodology guide.
Run your numbers
After excluding bots, monitors, employees, and existing customers. The working number for operations.
From your own analytics property, not the vendor's dashboard.
Sessions the tool attributed to a company, on the same page scope.
Travels with the quote, e.g. “four-week average”. A single week is weather, not climate.
Random sample scored against ground truth you already own.
The rest of the sample counts as plausible-unverified, tracked separately.
Used only for the sample-size line: 95% confidence, corrected for your identified-session pool.
- Raw match rate (identified ÷ D2)
- 30.0%
- Precision (confirmed ÷ scored)
- 80.0%
- Plausible-unverified share of sample
- 20.0%
- Usable match rate (raw × precision)
- 24.0%
- Sessions to sample next cycle
- ~237
Match-rate worksheet — visitorops.com/tools/match-rate-calculator Denominator: D2 — filtered human sessions (20,000 sessions) Identified sessions: 6,000 Raw match rate: 30.0% Precision sample: 50 sampled · 32 confirmed · 8 wrong · 10 plausible-unverified Precision: 80.0% Usable match rate: 24.0% Window: four-week average Next sample: ~237 identified sessions for ±5 percentage points at 95% confidence Quote as: "24.0% usable on filtered human sessions (D2), four-week average"
The pre-filled numbers are a labeled hypothetical example, not a benchmark — this site deliberately quotes no “good” match rate. Pull your denominator from your own analytics property and score the sample against ground truth you own (form fills, customer lists, your own employees) before quoting any number internally.
Worked example (hypothetical numbers)
A team counts 20,000 filtered human sessions (D2) over four weeks, of which the tool identified 6,000 (hypothetical numbers, not a benchmark). Raw rate: 6,000 ÷ 20,000 = 30.0%. They sample 50 identified sessions against ground truth: 32 confirmed, 8 wrong, 10 plausible-unverified. Precision: 32 ÷ (32 + 8) = 80.0%. Usable rate: 30.0% × 80.0% = 24.0% — the number to quote is “24.0% usable on filtered human sessions, four-week average,” not “we identify 30% of traffic.” For a ±5-point margin at 95% confidence, the next cycle needs about 237 sampled sessions from the 6,000-session identified pool.