Reassigned Numbers: The TCPA Trap Insurance Agents Miss

Most agents I talk to have their consent story straight. The lead came in through a form, the form had the right language, the timestamp is in the CRM, and the number was scrubbed before the first dial. Good.

Then they ask the question that unravels it: is the person who answers that phone today still the person who filled out the form?

Because consent under the TCPA belongs to a person, not to a string of ten digits. Phone numbers get disconnected and handed to somebody else constantly — people switch carriers, drop a line, lose a job phone, pass away. When that happens, the consent you carefully documented stops covering the number you are about to dial, and nothing in your file tells you.

This is the exposure nobody warns new agents about, and it lands hardest on exactly the lists agents like to work: aged leads, old CRM records, and anything you bought secondhand. Here is how the mechanic works, what the FCC database does and does not do for you, and the workflow that keeps a stale number from becoming a demand letter.

What a reassigned number actually is

A reassigned number is a phone number that was permanently disconnected by one subscriber and later issued to a different one. It is not a wrong number in the typo sense, and it is not a disconnected line. It is a live, working number in a stranger's hand that used to belong to your prospect.

From your side of the call, nothing looks different. The number rings. Somebody answers. Your CRM shows a lead record with a name, a form submission, and a consent timestamp. The only signal that anything is wrong is the human voice telling you they have never heard of you — and by then the call already happened.

The FCC built an entire national database around this problem, which should tell you how routine it is. Numbers churn. A lead list that was accurate the week it was generated gets less accurate every month it sits, and the decay is invisible from the inside.

Why your consent record does not travel with the number

The TCPA restricts certain calls unless you have consent from the called party. Courts and the Commission have consistently treated the called party as the current subscriber or customary user of the line — the person who actually picks up — not whoever the caller believed they were reaching.

Follow that through and the conclusion is uncomfortable but simple. If the number was reassigned, the person answering never gave you consent, never filled out your form, and never heard of your agency. Your consent record is a genuine, well-kept document about a different human being. It is evidence of diligence. It is not a defense to that particular call.

This is also why “but I had consent” is a weaker answer than agents expect. Nobody is disputing that you had consent. The dispute is about whose.

The Reassigned Numbers Database, and what it really gives you

The FCC created a national Reassigned Numbers Database — run by the North American Numbering Plan administrator and reachable at reassigned.us — to give callers a way to check before dialing. Carriers report permanent disconnects into it. You query it.

The query is not “who owns this number.” It is narrower and more useful than that. You submit a phone number plus the date you obtained consent, and the database tells you whether that number has been permanently disconnected at any point since that date.

The safe harbor is the reason to bother. If you queried the database before the call, the database returned no, and it turns out the number had in fact been reassigned, you are protected from liability for that specific call — because you did the check and the database was wrong.

Read the boundaries of that carefully, because they are tight:

• A “no data” response is not cover.The safe harbor attaches to a clean “no,” not to the act of asking. If the database has nothing, you are back where you started.
• It only helps if you had consent to begin with. The safe harbor addresses reassignment specifically. It does nothing for a call you never had permission to make.
• You have to query before the call, not after. A retroactive check is a research project, not a defense.
• You need the consent date to be real. The query is anchored to it. A guessed or backfilled date produces an answer that means nothing.

That last one links this whole topic back to record-keeping. If your lead records do not carry an honest consent timestamp, you cannot run a meaningful query at all — which is one more reason the consent date belongs in a structured field, not in a notes box.

Where this bites agents: the aged list

Fresh leads carry very little reassignment risk. A form filled out this morning is almost certainly still in the same hands this afternoon. Risk accumulates with age, and it compounds quietly.

So the profile of a high-risk list is easy to describe. Leads six months to several years old. Records that have been sitting in a CRM since a campaign you stopped running. Lists bought secondhand, where the consent date belongs to somebody else's form and you are trusting a spreadsheet column. Any file where the consent date is missing entirely.

None of that means aged records are off limits. Working your own old leads is one of the better uses of a dialing hour, and I have written a full 30-day plan for reviving an aged list. It means the re-scrub step at the front of that plan is not optional busywork. The older the list, the more of its value depends on cleaning it before the first dial rather than discovering the problems one angry call at a time.

The wrong-number call is a compliance event

Here is the habit that costs agents more than any database subscription. Somebody answers, says “that's not me, you have the wrong number,” and the agent marks it no contact or bad number and moves on. Three weeks later the cadence brings it back around and calls the same stranger again.

That second call is worse than the first in every way that matters. The first was a mistake. The pattern is what supports a willfulness argument, and willfulness is what turns $500 per call into $1,500. It is also precisely the fact pattern that professional TCPA plaintiffs are looking for, because repeat dials to a non-consenting party are easy to document and hard to explain.

So the rule is short. A wrong number is a hard stop:

1. Confirm briefly, then get off. You are not qualifying this person. You are not asking if they might also be interested. Apologize and end the call.
2. Suppress the phone number, not the lead record. The same number can sit on three lead records from three sources. Suppress only the record you were dialing and the other two will ring through.
3. Give it its own disposition. A wrong number is not a no-answer and not a not-interested. It needs a code that blocks future dials, which is worth checking against your disposition set.
4. Do not resell, retrade, or recycle it. The number does not become clean because time passed or because it moved to a different campaign.

Signals in your own data that a number went stale

A database subscription is the formal answer. Your own call history is the free one, and it is usually earlier. Patterns worth watching:

• The name never matches. Two or three separate connects where the person gives a different name than the record. That is not a bad list, that is a reassigned line.
• The voicemail greeting names someone else. Cheap, reliable signal. Note it and stop.
• The area code stopped matching the address. Not proof on its own — people move and keep their mobile numbers — but combined with anything else on this list, it is worth a check.
• A long silence followed by a live answer. A number that went to nothing for months and then suddenly connects has a plausible explanation, and it is rarely that your prospect came back.

These only help if somebody is looking. That is a case for reviewing dispositions during your monthly pipeline audit instead of letting the same record cycle through the dialer indefinitely.

A workflow that fits a real dialing week

You are not going to run a manual database query on every record before every dial. Nobody does. What works is tiering the effort by how much risk the list actually carries.

1. Capture the consent date as a field. Not a note. A date field, populated at import, for every record. Without it nothing downstream works.
2. Fresh leads: dial normally. Standard DNC scrubbing and quiet-hours checks. Reassignment risk on a same-week lead is negligible.
3. Anything past roughly 90 days: check before the campaign, not before each dial. Run the batch at the point you decide to work the list. One pass, one record of results, filed with the list.
4. Purchased or inherited lists: check every time, no exceptions. You did not collect the consent, so you cannot vouch for the date. Treat the whole file as unverified until it clears.
5. Keep the query results. The safe harbor is a record, not a habit. If you cannot show the query and its response, you did not query as far as anyone else is concerned.
6. Wire wrong-number dispositions to suppression by phone number. This one is free and prevents more repeat calls than everything above it combined.

What software should carry

The judgment calls here are yours. The mechanics should not be something you remember at dial 80 on a Friday afternoon. Whatever you dial with should be handling four things on its own:

• A structured consent-date field that survives CSV import instead of getting flattened into a notes column.
• Suppression keyed to the phone number, applied across every record that number appears on, the moment a wrong-number or DNC disposition is set.
• A hard block on the next dial to a suppressed number — not a warning banner you can click past at speed.
• An exportable log tying each dial to the scrub and consent state at the time it was placed, so your defense is a file rather than a memory.

Notice that three of those four are things you can have today without any subscription. The database query is the formal safe harbor. The suppression plumbing is what actually stops the second call to a stranger, and the second call is the one that gets expensive.