Finding the heirs a skip trace can't: what happened across 171 family members who never lived at the property

By: Jeremy Nash
Study completed:: 2026-07-20
Published:: 2026-10-01

Finding the heirs a skip trace can't

Across 171 paired cases where a property owner's immediate family member never lived at the owner's address, a traditional skip trace delivered that family member 0.6% of the time. A group search delivered them 83.0% of the time. The paired difference is +82.5 points (95% CI +75.1 to +88.8).

The question

An investor has a property owner's name, the property address, and the name of an heir, a sibling, or a spouse who is tied to that owner but has never lived at that address. Can a skip trace find that second person from those three facts?

A traditional skip trace query only has room for two of the three: the target's name and an address. It can't use the owner's name at all, because there's nowhere in the query to put it. The question this study answers is whether that missing third fact matters, and by how much.

How we tested it

We built 404 real trios (owner A, immediate family member B, property) from production data from an integration partner's property book, covering June 20 to July 17, 2026. Each trio passed four checks: A owns a property with a parseable address, A's record names B as a spouse, sibling, parent, or child, a full record exists for B, and B's address history does not contain A's property. A pre-declared subsample of 171 trios across 90 properties, drawn by clustering on property before any result was observed, was used for the paired comparison.

Both arms received the same three facts, with the traditional arm restricted to the two it can use:

Traditional skip trace: B's name plus A's property address.
Group search: a search of A and B together, with A's property as the input address.

"Delivered" means the system identified B as entry #1 for the traditional arm, or affirmatively matched B's record for group search. Ground truth is B's own vendor-issued ID, as recorded in A's own household relationship data, not an identification either method produced.

Results

Metric	Traditional skip trace	Group search
Returned nothing at all	91.2% (86.5–95.2)	0.0% (0.0–0.0)
Target present in results	0.6% (0.0–2.0)	87.1% (81.0–92.5)
Target delivered	0.6% (0.0–2.0)	83.0% (75.9–89.2)

Paired difference on delivered (group search minus traditional): +82.5 points (95% CI +75.1 to +88.8).

All intervals are cluster bootstraps over the 90 properties, not over the 171 trios, because trios sharing a property aren't independent observations. The wider, unpaired run across all 404 trios lands close to the same place: traditional returns nothing 92.6% of the time and delivers the target first 0.7% of the time.

One way to read the 171-trio subsample: group search delivered the target 142 times; traditional delivered it once. That comparison holds only for this specific case: immediate family who never lived at the property.

What this does and does not show

This is one tenant's property book, and it's predominantly Texas. Nothing here says how either method performs in other states or industries.

"Delivered" is checked against the owner's own vendor relationship record, not against field verification. It tells you the system named the record the data provider's own household data says is correct. It does not tell you a human confirmed that phone number reaches the right person.

The traditional arm is a single query shape: name plus address, run once. It is not a simulation of an investigator trying several queries or using a different tool. The claim here is about the query an investor can actually construct from the three facts they start with, not every query a determined human might eventually try.

The paired comparison (the 82.5-point gap) is only valid on the 171-trio subsample. It should never be read against the full 404-trio population, because that population wasn't searched by both arms.

The subsample's own fingerprint could not be re-verified after the fact: counts match exactly (171 trios, 90 properties) and the underlying 404-trio population re-verifies byte for byte, but the selection script that drew the subsample wasn't preserved, so its exact ordering can't be reproduced. That's a failure to verify the subsample's record-keeping, not evidence that the wrong trios were used.

Why it matters if you skip trace

If you work tax delinquency, probate, or pre-foreclosure files, this is the case you hit constantly: the person on the deed is dead, gone, or unreachable, and the person who actually controls the deal is a relative who's never lived anywhere near the property. A name-and-address lookup has no way to find that relative, because the address doesn't apply to them. Nine times in ten, that lookup comes back empty, and the file sits.

Group search works from the relationship instead of the address, which is why it can find someone a flat lookup structurally can't reach. If that's the file you're stuck on today, run a group search and see what it surfaces.

Data and reproducibility

171 trios across 90 properties (paired); 404 trios across 196 properties (traditional arm, unpaired).
Measured July 19–20, 2026, under a pre-registered blueprint committed before any result was observed.
Group search arm consumed 282 data-provider calls against a 500-call ceiling.
The 12 trios whose labels could not be uniquely pinned down by name and street were analyzed under both the most- and least-favorable assignment; the result did not move.
The full blueprint and results file, including the deviation log, are available on request.

Cite this: GroupSkip Research (2026). Finding the heirs a skip trace can't: what happened across 171 family members who never lived at the property. groupskip.com/research/finding-heirs.