A floor manager who types “homeowner phone list” into a search box is not shopping for a vendor. They are shopping for a file — one that loads on Monday, rings on Tuesday, and reaches the person who owns the house often enough that the reps stop asking who they are talking to. Vendors come later. The file is the thing, and the file has a spec.
This page is that spec: what a dialer-ready homeowner file has to carry on every row, what the word “match” has to mean for the phone on it, and how to check both on a sample before any money moves. It is written for the person who will load the file, not the person who will sign for it.
Disclosure: Scout Data is our product, and homeowner phone files are what we sell to call centers. The spec below is the one we build to, and it is also the one we would hand a floor manager buying from anyone else.
The row
Everything a dialer needs, and everything a setter needs, fits on one row. If a column is missing the floor will improvise it — from an area code, from a guess, from nothing — and the improvised version is where the compliance and connect problems come from.
| Column | Why it is there | Send the file back if |
|---|---|---|
| Owner name, as on title | The opener names them; right-party is judged against it | Names are blank, initials only, or “Current Resident” |
| Property address and ZIP | The call is about this house; the ZIP sets the clock | ZIPs have lost their leading zeros; addresses unstandardised |
| One primary phone | The dial | Several numbers per row and no ranking between them |
| Line type on every number | Dialing mode and consent rules differ for wireless | Blank or “unknown” on more than a sliver of rows |
| DNC status and the scrub date | Proof the registry scrub ran, and when | A yes/no flag with no date behind it |
| Owner-occupied flag | The person who answers can say yes | Absent — you will be dialing landlords about tenants’ roofs |
| The fields that set the pitch | The opener’s fact: last permit, build year, panels, sale date | Nothing on the row but a name and a number |
| Timezone | The calling window runs on the homeowner’s clock | Absent — derive it from the property ZIP, never the area code |
| Source and batch stamp | So this file can be judged apart from every other file | Absent — add it yourself at import, before anything else |
Two of those rows deserve a second look. Line type is a compliance input as much as a performance one, because the rules for automated dialing to a wireless number differ from the rules for a landline, and a file that leaves the flag blank has handed you the problem. And the timezone row is there because area codes stopped meaning where someone lives when numbers became portable: a homeowner in Phoenix with a 917 mobile is a 9:30pm call if the dialer trusts the prefix. You know where the house is. Use it.
What “match” has to mean
Every vendor will quote a match rate, and every match rate is true. The number counts rows that came back with a phone on them. What it does not say is how the phone got there, and there are three ways, which produce three different floors.
- Associated with the address. The number is one the data has seen alongside this street address — a previous owner, a tenant, an adult child who once had a package delivered. High match rates, working lines, and a steady stream of “they moved.”
- Associated with a person at the address. The number belongs to somebody linked to the house, chosen without checking whether that somebody is the owner. Better, and still the source of most wrong-party dials on a rental-heavy file.
- Matched by name to the owner on title. The deed says who owns the house; the number was resolved to that person, by name, and then checked to be live. Lower match rate on paper, because some owners genuinely have no reachable number — and the only one of the three where a connect is usually the right party.
The word on the invoice is the same in all three cases. The floor experiences them as three different products, and the gap shows up in one place: the share of connected calls where the person who answered is not the owner named on the row. Skip tracing accuracy covers why the match rate cannot tell you which one you bought, and skip tracing for solar and roofing teams covers how the third kind of append is built.
How to check it on a sample
Before the purchase, not after. This is a pre-purchase check of the match, and it takes an afternoon; the fuller post-delivery audit — four tests on a sample you drew — is in how to audit a lead vendor and belongs on the first paid file.
- Draw the rows yourself. Fifty to a hundred, at random, from a sample built to your real spec and geography. Not the vendor’s showcase rows.
- Check twenty names against the county. Assessor and recorder sites are public. Owner on the row versus owner on the record: a mismatch share above a handful is a file matched to something other than the deed, before a single dial.
- Dial them on a normal shift with a normal rep, and log three outcomes as three codes: right party, wrong party, dead or non-residential. “Not interested” is not one of the three; it is a right-party outcome.
- Compute two shares. Wrong party as a share of connects, and dead as a share of dials. The first tells you how the file was matched; the second tells you how old it is.
- Check the flag. On connects, ask or look up whether the line was a mobile, and compare against the line-type column. A flag that is wrong on a quarter of rows is worse than no flag, because the floor will trust it.
- Run your current file alongside. Same rep, same afternoon, same codes. A wrong-party share means nothing on its own and everything against the list you are already paying for.
The fields that set the pitch
A name and a live number will get you a connect. They will not get you past “why are you calling me,” and on a homeowner file that question is answered by the property columns, not the script. Which columns depends on what the floor sells:
- Solar. Whether panels are on the roof now — from imagery, not from the absence of a permit — plus utility territory, an estimated bill, and roof age. The opener is about the bill; the suppression is about the panels.
- Roofing. The date of the last roofing permit on the home, and whether the street sat inside a hail or wind footprint, with the event date. The setter says “the last permit we can see is from 2009” and asks if that is right.
- HVAC. The last mechanical permit, what it installed, and the home’s heating fuel — because a heat-pump sentence lands in Maryland and dies in Michigan.
- Any of them. Sale date, because a new owner has no contractor yet; build year, because a home under about 25 is plausibly on original equipment; and owner-occupied, always.
What each of those fields actually is — recorded, licensed, or modeled, and what a blank means — is in homeowner data fields explained. The short version: a field the vendor modeled is a guess dressed as a measurement, and the reps will find out at the door of the conversation.
Replacement and refresh: get them in writing
Phone data decays from the day it is delivered. Numbers disconnect, port, get reassigned; owners sell. A file that is right on delivery is a file that is drifting by the second month, and the two terms that decide what the drift costs you are whether a dead number is replaced or simply yours, and how often the file is re-verified against current ownership. Neither is on the rate card. Both go in the order.
Two more terms belong beside them: household-level deduplication before delivery, so the same parcel does not arrive twice under an owner’s name and a trust’s; and acceptance of your suppression file — prior opt-outs, customers, the last three orders — before the file is built, so you are not paying for rows you will discard. The routine that keeps a file dialable after it lands is in list hygiene for call centers.
Where the free test fits
Everything above is a check you can run on any vendor, and we would rather be checked than believed. For a floor already dialing homeowners, the fastest version is the free 1,000-record test: tell us your market and who you sell to, and we build a thousand rows to this spec — one active mobile per home, matched by name to the owner on title, scrubbed before delivery — and ship the file within a day. Dial it three times over a week or two on your own dialer, next to your current list, and we put the connect and wrong-number rates side by side on a call. No card and no minimum; if it does not beat what you are dialing now, keep the rows.
Frequently asked questions
What is a homeowner phone list?
A file of residential properties where each row carries the owner of record and a phone number for that owner, usually with property fields attached — build year, last permit, whether panels are on the roof — so the call can be about the house. The phrase covers everything from a compiled consumer file with a homeowner flag to a file built from deeds and permits; what separates them is how the phone got onto the row.
How many phone numbers per record should a homeowner list have?
One that reaches the owner beats five that might. A row with several numbers and no ranking is a row where the vendor did not decide, and the deciding gets done by your reps, one dial at a time. If a file does carry more than one number, it should say which is primary, what line type each is, and how each was tied to the owner — and your dialer should count attempts per household, not per number.
What does “owner-matched” mean on a phone list?
That the phone was tied by name to the person on the deed, not to the address. An address-matched number belongs to whoever the data associates with the house — a previous owner, a tenant, a relative — and it produces a working line that reaches the wrong person. An owner-matched number was resolved to the owner of record first and then to a phone. Both count as a “match” on a vendor’s report.
How do I check a homeowner phone list before buying it?
Draw fifty to a hundred rows yourself, check the owner names against the county record, dial them on a normal shift, and log three outcomes separately: right party, wrong party, dead number. The wrong-party share of connects is the number that tells you how the file was matched. Run your current list on the same afternoon so there is something to compare against.