Lead Disposition Codes for Insurance Agents: The Setup That Keeps Your Pipeline Honest

Nobody gets into insurance because they love taxonomy. But the two seconds after you hang up — the moment you tag what just happened — is the single highest-leverage habit in a phone-based practice, and it is the one almost everybody does badly.

I have watched agents run 200 dials a day and disposition maybe half of them. I have also watched agents build a 34-code system so elaborate they had to think about it, which is worse, because a code set you hesitate over is a code set you eventually abandon. The pipeline then becomes a pile: every lead looks equally cold, so you start at row one every morning and re-dial the same disconnected numbers you re-dialed last week.

This is the setup I would give a new agent on day one. Ten codes, what each one should trigger automatically, the three that carry legal weight, and a monthly audit that takes ten minutes.

What a disposition code is actually for

Most agents think of dispositions as record-keeping. They are not. A disposition is an instruction. When you tag a call “No Answer,” you are telling your system: try this number again, at a different hour, in about two days, and do not count it against the lead. When you tag it “Do Not Call,” you are telling it: suppress this number forever, and log that I did so.

That reframe changes the design rule. Every code must map to exactly one next action. If two codes produce the same next action, you do not need both — you need one code and a note. If a code produces no next action, it is decoration.

The ten codes

These cover essentially every outcome an agent working final expense, mortgage protection, or IUL lists by phone will hit. Names matter less than the behavior behind them — use whatever vocabulary your team already speaks.

Ten codes. Every one produces a different tomorrow. That is the whole test.

The three distinctions agents blur — and why they cost money

1. Not Interested is not Do Not Call

This is the expensive one, and it runs in both directions. Agents who collapse the two into a single “dead” bucket either torch thousands of workable leads, or — far worse — leave genuine do-not-call requests sitting in a pile that gets re-imported next quarter.

A prospect saying “not right now, money is tight” is a soft no with a shelf life. A prospect saying “take me off your list” is a legal instruction. Under the TCPA, a company-specific do-not-call request must be honored and kept indefinitely, and calls made after it carry statutory damages of $500 to $1,500 per call. Those two outcomes cannot share a bucket. If you are still building out the rest of your suppression discipline, the four different scrubbing jobs are worth understanding before you dial another list.

2. No Answer is not Bad Number

A number that rings twelve times is a prospect who was busy. A number that plays three tones and dies is a number you paid for that does not exist. Same non-conversation, completely different meaning.

Blurring them corrupts the only metric that tells you whether your list is any good. If 18% of a batch dispositions as Bad Number, that is a conversation to have with whoever sourced it — but only if you can see the number cleanly, separate from the ordinary no-answers that are just part of dialing.

3. Contacted is not the same as Reached

“Contacted” should mean you spoke to the person on the lead — not their son, not the receptionist, not a wrong number who happened to answer. Agents inflate their contact rate constantly by counting any human voice. Then they wonder why their close-per-contact looks terrible compared to the benchmarks other producers report. The denominator was wrong the whole time.

Disposition at hang-up, not at the end of the session

The most common failure mode is not a bad code set. It is batching. An agent runs a two-hour block telling themselves they will clean up the notes afterward, and afterward they are tired, and the cleanup becomes forty rows of “NA” that erase every nuance from the block.

Dispositioning has to happen in the same motion as the call — one click, before the next number connects. That is a workflow property, not a discipline property. If tagging a call means alt-tabbing to a spreadsheet and typing into a cell, you will stop doing it by Wednesday, which is one of the five places a spreadsheet quietly breaks. When the disposition buttons live on the call screen and the next lead loads automatically, the habit maintains itself.

Notes and codes do different jobs

A code is machine-readable: it drives the queue and rolls up into your reporting. A note is human-readable: it is what you read for four seconds before the next dial so you open with “you mentioned your daughter handles the bills” instead of a cold intro.

Keep them separate and keep the note short. The useful ones are always the same three things: what they said about money, who else is in the decision, and when they said to call. Everything else is you journaling.

The strongest note habit I know is writing the next openerinstead of the last summary. “Ask if Medicare supplement premium went up in July” beats “seemed hesitant,” because in three weeks the first one still tells you what to say and the second one tells you nothing. That pairs naturally with how you handle the first fifteen seconds on any call.

Attempt caps: when to stop dialing a lead

Dispositions only keep the pipeline honest if something eventually retires a lead. Without a cap, your No Answer bucket becomes a treadmill — the same 300 unreachable numbers cycling through your list forever, eating dials that should go to fresh contacts.

A cap that works in practice: six to eight attempts across two to three weeks, varying the time of day, then the lead drops to a low-frequency cycle rather than the daily list. Morning, midday, and early evening within legal hours cover most people's schedules; six calls all at 10am on Tuesdays is really one call repeated six times.

“Retired” should never mean deleted. Those contacts are exactly the raw material for a structured aged-lead revival pass a few months out, when circumstances have changed and the number is no longer being hammered.

The three codes that carry legal weight

Most of your codes are operational. Three of them are evidence, and they deserve stricter handling than the rest.

• Do Not Call — must suppress the number permanently across every list, including lists you import later. The failure case is real: an agent honors the request, then loads a new batch six months on that contains the same number, and dials it. Suppression has to live at the account level, not the list level.
• Bad Number / Wrong Person — when the person answering is not the person on the lead, that is a reassigned or mis-keyed number. Continuing to dial it is how agents end up calling strangers who never consented to anything.
• Not Qualified – State — if you are not licensed or appointed somewhere, the record should say so, because it explains why the lead was never worked if anyone asks.

What makes these defensible is not the tag itself but the timestamp attached to it. A demand letter is answered with a record showing the request came in on a date and the number never rang again after it. That is the same evidentiary standard that makes a power dialer defensible under the TCPA in the first place — the log, not the intention.

The ten-minute monthly audit

Once a month, pull your disposition counts for the last 30 days and look at four things. This is not analytics theater — each one has an obvious fix.

• Blank dispositions. Any real number here means calls are going untagged and your queue is running on stale information. Fix the workflow, not the agent.
• Bad Number rate by source. One batch running dramatically higher than your others is a data-quality problem worth raising with whoever provided it.
• Callback Set kept vs. missed. Missed callbacks are the most painful leak in the whole pipeline, because those were prospects who asked you to call.
• Not Interested vs. Do Not Call ratio. If Do Not Call is climbing relative to Not Interested, something upstream has changed — a harder opener, an older list, or a source whose consent trail is weaker than you were told.