You send three thousand records to a skip trace provider. A large share come back with no phone number at all, and of the numbers that do come back, plenty are disconnected, wrong, or belong to someone who has never heard of the property. The instinct is to blame the provider and go shopping for a better one.
Sometimes that is the right call. More often the provider is returning what the input deserved, and switching vendors produces the same disappointing result at a different price. Here is how to tell which situation you are in.
Diagnose the Input Before Blaming the Output
Skip tracing is a matching process. The provider takes the identity you supplied and tries to find the same identity elsewhere in its data. Anything that makes that match harder reduces your hit rate, and most of those things are in the file you submitted.
Name formatting is the quiet killer. County records are inconsistent in ways that break matching: last name first, middle initials attached to first names, suffixes glued on, trust and estate language sitting in the owner field, two owners crammed into one string. A record reading "SMITH JOHN A & MARY TR" is not a person as far as a matching engine is concerned.
Address quality is the other half. Missing unit numbers, abbreviations, and unstandardized formatting all reduce match confidence. Running the file through address standardization before submission is unglamorous and frequently moves the number more than switching providers would.
The test that settles it: take a hundred records that failed, look at them by hand, and ask whether a human could identify the person from what you submitted. If the answer is often no, the problem is upstream.
Some Records Are Not Traceable, and That Is Correct
A portion of any list represents owners who genuinely cannot be matched to a phone number, and no provider will change that.
Entity-owned property is the largest category. An LLC, trust, corporation, or estate is not a person and has no personal phone number. Tracing them as individuals will fail every time. The correct route is a registered agent lookup or a filing search, which is a different process, and lists heavy in entity ownership will always look terrible on a raw hit rate.
Recently transferred property is another. If ownership changed in the last few months, the identity may not yet be associated with the address anywhere in the aggregated data. Very elderly owners, owners who have died with no probate filing yet, and people who have simply never had a mobile line in their own name round out the group.
None of that is fixable, and it is the reason a raw hit rate is a poor quality measure. Segment the failures first. Once entity-owned and recently-transferred records are pulled out, the remaining hit rate on genuine individuals is the number worth judging.
The Number That Actually Matters
Hit rate measures how often a phone number came back. It says nothing about whether that number reaches the right person, and those two figures can diverge dramatically.
A provider returning a number for nearly every record is not necessarily better. It may be returning low-confidence matches that a more conservative provider would have declined to guess at. Wrong numbers are worse than no numbers: they consume dial time, they annoy uninvolved people, and they create compliance exposure with someone who has no relationship to your business at all.
The measure to run your decisions on is connect rate to the intended party. It requires actually dialing a sample and recording outcomes, which is more work than reading a vendor report, and it is the only version of the number that predicts deals. As covered in the guide to bulk skip tracing, sampling two providers against the same records is the cleanest way to get a comparison that means anything.
When Switching Providers Is the Right Answer
Once the input is clean and the untraceable segments are excluded, a genuinely weak provider will show up clearly. Regional coverage varies more than most investors expect, and a provider strong in dense metro data can be noticeably weaker in rural counties.
Data recency is the other differentiator worth paying for. Some providers refresh continuously and some work from older snapshots, and that difference shows up as disconnected numbers rather than missing ones.
The practical approach is not loyalty to one vendor. It is knowing which of two performs better on the list types you actually run, and cascading: send the failures from the first provider through a second, since the overlap between databases is far from complete and the second pass frequently recovers a worthwhile slice at a fraction of the original volume.
Fix the Handling, Not Just the Data
A number of apparent hit rate problems are actually attempt problems. One dial to a mobile number is not a test of whether that number works. People screen unknown callers, and a single unanswered ring proves nothing about data quality.
Time of day matters, calling hours in the recipient's zone rather than yours matter, and repeat attempts across different days matter. A record marked bad after one attempt may be a perfectly good number attached to someone who was at work. That is the same persistence principle behind responding fast to inbound leads, applied in the other direction.
Equally, records that returned no phone are not dead. They are a direct mail list. Treating a failed trace as a failed lead throws away reachable owners for no reason.
The full context for how tracing fits the wider process is in the guide to skip tracing for real estate investors, and this channel sits inside the complete guide to real estate lead generation for investors.
Segment the failures before drawing any conclusion. Once entity-owned and recently-transferred records are set aside, what remains is the only population a provider can fairly be judged on.