Insurance renewal retention is usually treated as a commercial problem — the client left because somebody else was cheaper. Occasionally true. Mostly not. Most lapses are administrative: the client did not open the notice, did not return the call, or the phone number on file stopped working in 2023.
That distinction matters, because a commercial problem needs a better price and an administrative problem needs a calendar. One of those is considerably easier to fix.
What follows is the ninety-day calendar our assistants work, planned backwards from expiry, plus the weekly at-risk list that catches the accounts about to go quiet while there is still time to do something about it.
Plan backwards from expiry
The default is to work forwards from whenever the system fires a reminder, which for most agencies is somewhere between thirty and forty-five days out. By then the useful conversations are already behind you.
Plan backwards instead. Expiry is day zero. Everything else is a subtraction, and each contact has a distinct job:
| Days out | Contact | Job | Owner |
|---|---|---|---|
| 90 | Data check | Confirm exposures, payroll, vehicles, contact details. Not a sales call. | Assistant |
| 75 | Market prep | Submission assembled, loss runs ordered, gaps flagged internally | Assistant |
| 60 | The conversation | Review meeting booked and held. Changes, concerns, anything new. | Producer |
| 45 | Options presented | Terms delivered with a recommendation, not a menu | Producer |
| 30 | The decision | Confirm intent in writing. Anything unresolved escalates today. | Producer |
| 21 | Paperwork | Applications signed, payment method confirmed | Assistant |
| 14 | Rescue | No response by now means phone, text and email on the same day | Assistant + producer |
| 7 | Final | Personal call from the producer. Every time. No exceptions. | Producer |
The ninety-day data check is the one most agencies skip and the one that makes everything after it work. It is a five-minute call that is not asking for anything, which is precisely why people take it. It also surfaces the thing you needed to know — they bought two more vans, they moved premises, they hired eleven people — early enough that the sixty-day conversation is about their actual business rather than a generic renewal.
The weekly at-risk list
Rebuilt every Monday. Capped at twenty-five accounts. Both of those constraints are doing work.
An account lands on the list when any of these are true:
- Two contact attempts with no response
- Renewal inside thirty days with no confirmed intent
- A claim in the last twelve months with no conversation since
- Any premium increase above your threshold, flagged the day the terms arrive
- A returned email or dead phone number — which is a lapse in progress, not an admin note
- A payment failure of any kind
The cap matters more than the criteria. Twenty-five accounts is a list somebody clears in a week. Ninety accounts is a report, and reports get read, nodded at, and filed. Capping it forces the prioritisation that everyone agrees with in principle and skips in practice.
Silence
The leading indicator of a lapse — not price, not a competitor, not dissatisfaction. Two unanswered contacts is the single most predictive signal on the at-risk list, and the cheapest one to act on.
The reason this is an admin problem, not a commercial one
Who makes which call
The split follows the licensing line, which is not negotiable and not a matter of style.
An unlicensed assistant can confirm details, schedule reviews, chase documents, send approved reminders, and rebuild the at-risk list. That covers the ninety, seventy-five, twenty-one and fourteen-day contacts — more than half the calendar.
Anything that quotes, advises, compares coverage or solicits a policy requires a license. In every state. Without exception. Your state insurance department is the authority on where exactly that line sits, and it is worth reading rather than guessing, because the penalty for getting it wrong is not a stern email.
Draw the boundary in writing before the first call, and put the escalation route directly next to it. When a client asks the assistant a coverage question — and they will, on roughly the second call — the assistant needs a scripted holding line, not a judgement call: acknowledge, book the producer, do not answer.
Two scripts worth having
The ninety-day data check
“Hi [name], Priya from [agency] — nothing needed from you today, I’m just updating the file ahead of your renewal in the spring. Has anything changed on your side this year? New vehicles, new premises, headcount up or down?”
No ask. No pitch. It takes four minutes and it routinely surfaces something that changes the renewal materially, which is a good return on four minutes.
The fourteen-day rescue
“[Name] — your cover expires on the [date] and I haven’t been able to reach you. I don’t want you to end up uninsured by accident. Can you reply with a yes or a no either way? If you’ve moved to another broker that’s completely fine, I just need to know so I stop chasing.”
Direct, and it works because it names the actual risk. “Uninsured by accident” gets replies that “following up on your renewal” never will — and the permission to say no is what makes it feel like a courtesy rather than a chase.
What to measure
Three numbers, monthly. Not more.
- Retention rate by month of expiry. Aggregate annual retention hides the month that went wrong. Broken out by expiry month it tells you which season your calendar failed in.
- Contactability. The share of the book you could actually reach on the first attempt. This is a data-quality number wearing a retention costume, and it is usually worse than anyone expects.
- Lapses with no recorded conversation. The pure administrative loss. This is the number the calendar exists to reduce, and the only one that tells you honestly whether it is working.
That second number is where most renewal calendars quietly break. A perfect ninety-day sequence run against a database where a quarter of the phone numbers are wrong is a perfect sequence delivered to nobody — which is why the two-week data cleanup is usually the right thing to do before the renewal calendar rather than after.
The same backwards-planning logic applies anywhere appointments and deadlines drive revenue, which is why it looks almost identical in a clinic: filling cancellations from a waitlist is the same calendar with different stakes.
If your renewal season currently starts at forty-five days because that is when the system reminds you, the calendar above is the whole fix. See how we run it, or book twenty minutes and we will scope it against your actual book.