Ask a commercial account manager when a renewal "starts" and most will say 60 days out, because that is when the carrier's renewal terms show up. Ask the same person when the account was actually lost and they will trace it back further: the loss runs that were requested on day 45, the producer who last spoke to the client in February, the indication that came in 18% up with no time left to remarket. The renewal did not fail at 60 days. It failed at 120, quietly, and nobody was looking.
This post is about what "looking" means in practice, and why it is a data problem before it is a people problem.
The 120-day window
For a commercial account of any complexity, remarketing takes time you cannot compress:
- Loss runs. Carriers take 10 to 15 business days to produce them, longer around quarter end, and you usually need five years from more than one carrier.
- Updated applications. ACORD 125/126/140 with current payroll, receipts, and schedules. That means a conversation with the client, and the client is busy.
- Market feedback. Submitting to three or four carriers and getting real indications takes two to three weeks.
- Client decision. A week, if you are lucky and the difference is obvious.
Add it up and a serious remarket needs roughly 75 to 90 days from decision to bind. If the decision is made at day 60, you are already choosing between the incumbent's terms and a rushed submission. At day 120 you have a real choice. The revenue does not leak at renewal. It leaks in the 60 days when a choice was still possible and nobody made one.
Predicting problem renewals
Agencies that do this well have a person, usually a senior account manager, who "just knows" which accounts are going to be trouble. Ask that person how they know and you will get a list. It is a surprisingly consistent list across agencies:
- A carrier appetite shift on that class of business, usually visible in a bulletin or in what happened to the last three similar accounts.
- Loss activity since the last renewal: an open claim, a reserve increase, a frequency pattern.
- A rate indication or a filed increase for that carrier in that state.
- Something outstanding in the AMS activity log: a loss-run request not returned, a supplemental not received, an inspection recommendation not answered.
- No producer touch in a long time. The number varies, but 60 days without a logged call, meeting, or substantive email on a mid-sized commercial account is a warning sign in every agency we have talked to.
- A change on the client's side mentioned in an email thread but never recorded: new location, new vehicles, a big contract, a sale of the business.
None of those signals is hard to detect. All of them live in different places: the AMS policy record, the AMS activity log, the carrier portal, the carrier's email notices, and the account's inbox. The senior account manager holds them together in her head for the 40 accounts she knows best. Nobody holds them together for the other 400.
A data problem
The AMS knows the expiration date, the carrier, the premium, and every activity that was logged. It does not know what the carrier said in a portal notice last Tuesday, and it does not know that the client mentioned a new warehouse in an email that never became an activity. The carrier portal knows its own renewal status and nothing about the other four carriers on the account. The inbox knows everything and is searchable by nobody.
A renewal list is only as good as the worst of these sources. Most agencies build theirs from the AMS expiration report alone, which is why it is a list of dates and not a list of risks. It tells you when, and the whole problem is which.
What a radar watches
If you were to build the senior account manager's judgment into a system, here is what it would do, per account, starting 120 days out:
- Read the policy and account from the AMS: lines, carriers, premium, last renewal outcome.
- Read the activity log and list anything open older than 14 days.
- Check carrier status wherever it is visible: portal, download, or notice email.
- Read the account's recent email and pull out anything that looks like a change, a complaint, or a question that was never answered.
- Compute a touch gap: days since the last logged producer contact.
- Rank, with the reason on the line: "Carrier X 14% indication; two open loss-run requests; no producer touch since 3 March."
Then it drafts the remarketing packet for the accounts you decide to move, because the day you decide is the day the clock starts, and the applications and loss-run summary are the slow part.
The important design choice is that this does not send anything. It produces a list and drafts. The account manager still makes the call and the producer still makes the phone call. What changes is that at day 120 they are looking at 20 accounts with reasons instead of 400 with dates.
A manual test
If you want to test the idea before spending a dollar, run this manually for one month:
- Pull the AMS expiration report for 120 to 150 days out.
- For each account over a premium threshold you choose, have someone spend five minutes checking the four other sources: activity log, carrier portal, notice emails, inbox.
- Write one sentence of risk per account.
- Count how many you would remarket that you would not otherwise have looked at until day 60.
Most agencies that do this exercise find the number is not zero, and that the five minutes per account is precisely the work nobody has time for. That is the case for making a machine do the reading.
If you want the machine built on your book rather than described, that is what Renewal Radar is.