Guide
Skip tracing accuracy and match rates explained
Why every skip tracing vendor claims 90%+ match rates — and why the number you actually get is almost always lower.
Every skip tracing vendor's marketing page eventually gets to a number. "90% match rate." "95% accuracy." "Industry-leading hit rates." Pull back the curtain and these numbers usually mean different things at different vendors, measured on different lists, against different definitions of "match." Here's how to read them honestly.
Match rate vs. response rate vs. accuracy
Three numbers get conflated in vendor marketing:
Response rate — what percentage of your input records got any data back from the API. The vendor returned something, even if the something is a phone number from 2018 that's been disconnected.
Match rate — what percentage of input records got back at least one verified, currently-valid contact attribute. This is the number that should matter to you.
Accuracy — when the API returns a phone number, what percentage of those numbers actually reach the right person on the first try.
A vendor can advertise a 95% response rate that translates into a 65% match rate and a 50% accuracy on dial. The marketing is technically true and operationally useless.
Why match rates depend on the list
Match rate is mostly determined by the input quality, not the vendor. The same vendor will get wildly different numbers depending on the list:
- Recent property records with full names + current addresses → 75–85% match
- Old probate records with name + 5-year-old address → 50–60% match
- Common-name list with no disambiguators → 30–50% match
- Foreign-language name list → much lower, often 20–40%
If a vendor brags about a 90% match rate without saying what list they tested on, they're either tested on a curated easy list or they're inflating the response rate.
How we measure ours honestly
We test on real anonymized customer lists across the categories above. On typical real estate investor lists from county records (absentee owners, tax delinquent, probate), we see 65–80% match rates with at least one verified mobile per matched record. On collections and process-server use cases the range is similar. On B2C enrichment with stronger inputs (name + email), match rates run higher.
None of those numbers are 90%+ and we don't claim they are.
How to test a vendor for yourself
- Pick a real list of yours, 500–1,000 records, representative of your normal use case.
- Send the same list through 2–3 vendors in parallel.
- For each vendor, calculate: response rate, match rate (at least one phone or email returned), and overlap with the other vendors.
- Ideally, run a small dial test on the matches and measure first-attempt connect rate. This is the only number that translates directly to revenue.
- Pick the vendor that wins on cost-per-useful-match, not cost-per-API-call.
Why pay-per-match billing matters here
Vendors that bill per API call have no economic incentive to maintain accuracy. Their revenue is the same whether the number is good or dead. Vendors that bill per successful match are forced to care about quality because misses are free. Skip Trace API uses pay-per-match for exactly this reason. Tracerfy and several others bill per attempt.