Ask five skip tracing vendors how accurate they are and you’ll hear five numbers between 80% and 95%. All of them are technically true and none of them predict what your dialer floor will experience on Monday. This guide explains what the advertised numbers actually measure, which metric to demand instead, and how to test any vendor — including us — before you commit a budget.
Match rate measures effort, not usefulness
A “match” means the vendor returned a phone number for the record. It does not mean the number rings. It does not mean the person who answers owns the property. It does not mean the number is legal to dial. Match rate is a coverage statistic — and because returning any number counts, the incentive is to err toward returning something, which is exactly how wrong-party calls end up burning agent hours.
The funnel that actually matters
Between an address list and a conversation with the owner, every record passes four gates. Here is the arithmetic with sample numbers — illustrative math to show the shape of the funnel, not a measured benchmark of any vendor:
Illustrative example — the percentages below are hypothetical inputs for the worked math, not test results.
| Stage | Sample rate | Records left (of 10,000) |
|---|---|---|
| Phone returned (“match”) | 90% | 9,000 |
| Number is in service | 80% of matches | 7,200 |
| Belongs to the current owner | 70% of working | 5,040 |
| Safe to dial after DNC scrub | varies by market | the list your floor gets |
Two vendors can both say “90% match” while one delivers thousands more ownable conversations than the other, because the gates below the match line differ. That is why the only number worth negotiating on is the bottom of the funnel — cost per dialable, right-party contact. Our cost calculator does the arithmetic for any quote.
What moves accuracy in practice
- Name-matching to the deed. Tying the phone to the owner of title by name — not just to the address — is the single biggest right-party lever. It is the default on every list we ship.
- Ownership freshness. Homes change hands constantly; a trace against stale ownership reaches last year’s owner with perfect confidence.
- Dead-number replacement. No graph is perfect. The honest fix is replacing numbers that don’t ring instead of charging for them.
- Mobile-first output. Landline-heavy appends inflate match rates and deflate connect rates. Insist on knowing the mobile share.
The standard we think the industry should adopt
Published, methodology-documented accuracy tests — same file, several vendors, logged dialer outcomes — are vanishingly rare in this market, which is why advertised numbers float free of consequence. We’re preparing exactly that kind of documented head-to-head on our own data and will publish the results and method here. Until then, the honest answer to “how accurate is skip tracing?” is: test it on your records, and make every vendor earn the number they advertise.
Frequently asked questions
What is a good skip tracing match rate?
Advertised match rates of 80–95% are common across the industry, but the number is nearly meaningless on its own — it counts any phone returned for the record, including previous residents, relatives, and disconnected lines. Judge vendors on right-party contact from your own dialer logs instead.
Why do two vendors return different numbers for the same address?
Different underlying identity graphs, different ownership-record freshness, and different matching thresholds. A vendor tuned for coverage returns something for almost every record; a vendor tuned for precision returns fewer, better numbers. Neither philosophy is wrong — but they produce very different dialer days.
How do I test a skip tracing vendor properly?
Pull a few hundred records your floor has already dialed, run them through the candidate vendor, and compare against your logged outcomes: right-party connects, wrong parties, dead lines. A same-file test beats every advertised percentage — and any vendor confident in their data will happily run one. We will.