Insurance renewals: the calendar changes by line of business and the costliest leak is invisible

At the end of the year, in any broker's book of business, there are two lists of lost policies. You fought for the first one: the customer compared, argued about the increase and left with someone else. You have every name on it and you know who you lost each one to. You never even saw the second one: the debit bounced in March, the rejection arrived in a portal file nobody opened, and the insurer voided the policy for non-payment. That list appears in no retention measurement, because technically they are not lapses: they are logged as voided.

The first one is worked with a calendar, and it does not start on the same day for every line of business: 45 days for motor, 60 for commercial, 30 for home. What sets the lead time is not a communication rule but how long it takes you to reach the price message with a defensible number. If you requote with a single insurer, your motor renewals start comfortably at 30 days; if you requote with four and one of them takes two working days to reply, 45 is the floor.

The second one is not worked with messages: it is worked by opening a file every month, always on the same day. Both hold up by hand for considerably longer than people think, and the point where they stop holding up is not marked by the number of messages.

Renewal is one of the few commercial events with a date known twelve months in advance. The problem is that the broker next door knows it too, and their first question when they hook someone is: when does yours expire? You do not compete against that on price: you compete by having got there first and with data the other one does not have.

The line of business decides when the countdown starts

The column that rules is the third one. The lead time is not an industry convention: it is the time it takes you to reach the second contact with a defensible number. If you requote with a single insurer, your motor renewals start at 30 days at no cost. If you requote with four and one of them takes two working days to come back, 45 is the floor.

The last row is the one that gets looked at least and the one that anticipates half of this text: there are lines of business where the renewal is barely ever discussed and the policy is lost anyway, on the payment side.

The first contact is a data check and does not mention price

It is not an expiry notice: it is the annual data check, and it is written as such.

For motor, four closed questions in a single message:

1. Does the car still sleep in a garage every night?
2. Does anyone else drive it regularly? If anyone is under 25, tell me who.
3. Have you added anything that was not there: CNG kit, alloy wheels, sound system, alarm?
4. Do you use it for work, even occasionally? Deliveries, apps, sales rounds.

For commercial or home multi-peril the questions change, not the logic: what you refurbished, what it would cost today to replace everything inside, whether you added or removed an alarm, whether you rent out part of it.

There are two reasons for this message to exist and the first one is not commercial. If the sum insured has gone stale, the proportional rule applies to a partial claim: a shop insured for 10 million when replacing the contents costs 20 collects half of what it claims. Asking in time is your job, not an excuse to write.

The commercial reason is the other one: requoting with several insurers using real data takes days, and 45 days out nobody is in comparison mode yet. The hard rule is that this message cannot contain the word renewal or a figure. If it does, you put a date in their head a month and a half early and handed that time to the competition.

The price contact carries two coverage options, not two prices

Here comes the number, with four elements:

• What they were paying and what it will be, side by side. Without the comparison, any figure looks arbitrary.
• The reason in one concrete line. It goes up because we updated the sum insured: the car is valued today at X. Never because of the context or because of an insurer adjustment.
• A second coverage option with one lever moved and named. It is not a discount or an alternative price: it is the same policy with a higher partial-damage deductible, or without glass and hail, or moving from comprehensive to own-damage cover with enhanced third-party. The lever is stated by name and with its consequence: with this one, if the crash is your fault you pay X out of pocket instead of Y. A single number turns the conversation into a yes or a no, and the no goes off to ask for a quote elsewhere. Two coverage options turn it into which of the two, and that comparison happens inside your proposal, with your advice, against a competitor who will send a bare number without saying what they took out.
• What happens on the expiry date if they do not reply at all.

The lever is not picked at random: it comes from what they answered in the first contact. If the car sleeps in a garage and is driven by a single 40-year-old, raising the deductible is an honest lever. If they use it to make deliveries four days a week, it is not: there the lever is a different one, or there simply is no second option and you say so in those words. Offering cheaper cover that the customer cannot sustain is the fastest way to win a renewal and lose a claim.

The three-day notice states a consequence, it does not sell

The last message argues nothing. It states the day and time of expiry, states what happens if the customer does nothing, and asks a single binary question.

If there is automatic debit: On Friday the 12th, $X will be debited from the card ending in 4417 and cover stays the same. Can you confirm? If there is not: On Friday the 12th at 12:00 the policy is left without cover. Do I go ahead with the cover you have today?

Those four digits are not decoration. They are the only reasonable chance the customer has to notice that the card the insurer has on file is no longer the one they hold, which is exactly the origin of the problem covered further down.

If you reach the expiry date with no reply, what follows is not another message: it is a call, and not out of commercial insistence. In any other industry a customer's silence is a sale that did not happen. Here it is a person who tomorrow may be driving without cover, convinced that they have it. That difference is what forces you to pick up the phone, and it is worth saying out loud as soon as they answer.

Policyholder, insured and payer are rarely the same person

The policy has separate roles and your WhatsApp almost always has only one of them saved. The policyholder signs, the insured uses the asset, and the payer can be a third party: the father who bought the car for his son, the accountant who handles the shop's payments, the company that pays for the fleet, the son who renewed his mother's home policy.

Each contact breaks with a different person. Only the person who uses the asset can answer the data check: the father has no idea whether the car sleeps indoors in La Plata (an Argentine city). Only the person who pays decides on the price. And a rejected payment can only be solved by the cardholder, who is precisely the one who often is not in any conversation of yours.

No system is needed to fix this: what is needed is opening the list you already have and putting three phone numbers where today there is one, labelled with who uses, who decides and who pays. In a book of 300 policies they fail to match on quite a few more than one expects, and those are the ones with the highest premium: fleets, shops and families with several cars on the same policy.

A new quote is the same questionnaire in a different order

When a how much would it cost to insure the car? comes in over WhatsApp, the mistake is to answer with a number. A price that later goes up when you enter the real data costs more than having taken two hours: the customer never believes another figure from you.

What works is two messages. The first one asks for the data in a numbered list, all at once, never one at a time: make, model and year, postal code where it sleeps, whether there is a garage, who drives it and since which year, and by when they need it. That last piece of data is what separates someone buying today from someone browsing, and it costs nothing to ask.

The second message gives a range with the two conditions that move it — deductible and insured value, said in plain language — and a date for the firm number. And it carries something almost nobody writes: how long that quote is valid for. A number from three weeks ago no longer exists, and saying so from the start saves you the conversation of having to take it back later.

The leak that appears in no retention report

Everything above deals with the policy that gets discussed. The one that hurts most is the one that was never discussed.

The full circuit is this. The customer's card was reissued with a new expiry date and they never updated it with the insurer. Or they closed the savings account where the debit was set up. Or the charge bounced against a limit. The month's instalment is not collected. The insurer retries on the following cycle and it bounces again. The rejection travels in a rejected-payments list that appears on the broker's extranet, in a file that gets downloaded when someone remembers.

Meanwhile, the arrears suspend cover. It is written in the general conditions of any policy and it is the only point in this whole text that is worth reading in yours, because the deadline and the reinstatement procedure change from insurer to insurer. The customer does not know. They keep driving, they keep the shop open, they keep thinking they are insured because they never cancelled anything.

Three things make this leak different from all the others:

• It generates no conversation. Nobody tells you they are leaving. There is no price objection, no request for a requote, no I am thinking about it. One day they are simply gone.
• It does not show up as a loss. In the system it is left as voided for non-payment, which is a different status from not renewed. If you measure retention against the month's expiries, that policy never entered the count: it died before reaching the expiry date.
• Winning it back costs more than getting a new customer. When the policy is issued again, the customer starts from zero: they lose the seniority and the no-claims discount they had been accumulating, and on top of that they now compare you against the market price. That conversation is worse than the renewal you lost while arguing.

And there is the scenario that turns this into your problem and not the insurer's: the claim happens in the month the policy was suspended. The call starts with I crashed and continues with I pay you every month. You have the rejection list that said so, with the date it was published.

The check that prevents all of this is not a system, it is a routine with a fixed date: the day the insurer publishes the list, cross-check it against your book and call — call, not write — every rejection that same day. A rejection has a window: between the first bounce and cancellation there are a few weeks and a two-minute task on the customer's side. Once the window has passed, no message fixes it.

The two leaks in a book of business, with numbers

Take a book of 300 annual policies: 25 expiries a month. With an average annual premium of USD 600 and a commission of 18%, each policy is worth USD 108 a year and USD 2,700 of annual commission passes across your desk every month. It is a template for you to run your own figures through, not an industry benchmark.

The leak you see. Of those 25 expiries, the ones that do not renew after receiving the price. If they are 4 a month, that is USD 432 of annual commission that does not come back until you find four new customers. You know them one by one, you know why they left and you have the conversation saved. And each one costs more than it looks: a lost policy does not cost one instalment, it costs twelve, plus the other policies that customer had with you.

The leak you do not see. Take the part of the book on automatic debit — say 200 policies — and apply your monthly rejection rate to it. Most bounces recover on their own at the retry on the following cycle; the problem is the ones that bounce twice in a row. If one or two a month reach cancellation, that is between 12 and 24 policies a year lost without anyone having compared anything: on the order of half of what you lose arguing about price, without a single conversation in between.

None of those numbers is yours. The only one that matters takes twenty minutes to get: ask the system for the policies voided for non-payment over the last twelve months and count them. If that number surprises you, you already know which of the two leaks to deal with first, and it is not the one you were looking at.

That same book of 300 has a second reading, and it is the one that decides whether the calendar above holds up with a spreadsheet or not. With 25 monthly expiries, writing is never the problem: the 75 messages get dispatched while you do something else. What eats the month are the three tasks that surround those messages, and none of them is fixed by writing better: requoting on each portal, cross-checking the rejections file and reading what comes back in free text.

The number that decides is not how many policies you have: it is how many times you type the same data. A broker with 400 policies with a single insurer and payment by transfer lives comfortably with a spreadsheet. One with 150 policies across four insurers, three lines of business and automatic debit already has a time problem, and it is not one of volume: it is that the same model, year, postal code and garage get keyed in four times per renewal into four portals that do not talk to each other.

On the payment side the same thing happens. With one insurer, reviewing the rejections is a short monthly routine. With four it is four files, four formats and four publication dates, and it is the first task that gets skipped when the week is heavy. Precisely the one that has no second chance.

What does not get automated in insurance

The claim, never. When an I crashed or a they broke in comes through, a person answers from the very first message, and that first message does not ask for paperwork. There are notification deadlines that run from the event and a version of what happened that is worth hearing before the customer writes it into a form: what they declare in the first few hours conditions the whole file.

Coverage advice does not get automated either, and here the limit is not commercial judgement but licensing. Recommending a sum insured, removing a cover or accepting a higher deductible is an act of advice given in your name and for which you answer to the regulator. An automatic message suggesting that the customer lower their cover so the price works out is not an automation with a bad tone: it is your professional responsibility delegated to a template.

Underwriting, even less so. Telling someone whether their risk is acceptable or not — a car more than twenty years old, a shop with an activity other than the declared one, a driver with a recent claims history — is an answer the insurer gives case by case. If you give it automatically, you then have to stand behind it.

The number almost no broker has at hand

Before asking anyone for a quote, put together a figure that is on no dashboard: how many policies of yours were voided for non-payment over the last twelve months. Not voluntary cancellations, which are a different conversation and can even be healthy for the book. The other ones: the ones lost because a debit bounced, the insurer published it in its rejections list and nobody read it before the reinstatement deadline expired. It comes from adding up twelve files you have already received.

Divide it by your book and you have the number that orders everything else. If it comes out below 1% and you work with a single insurer, you can sustain the three-contact calendar with your own expiry list and a diary: there is no project here, and your time is going somewhere else. If it comes out above 3% and you requote on four portals, you already know that your problem is not writing renewal messages: it is the repeated typing and a monthly file that a heavy week swallows.

And then one more question, which is the one that hurts: of those cancellations, how many were lost while the reinstatement deadline was still running? Those you did not lose on price or because the insured went to another broker. You lost them by not opening a file in time, and they are the only ones you win back without discounting commission.

That percentage, and nothing more than that percentage: info@striqtech.com.