Insurance brokers: renewals without a spreadsheet

A renewals spreadsheet never warns anyone

Short answer: an insurance broker stops losing renewals by turning every expiry date into a sequence of dated reminders, prepared in advance and approved by a person, rather than rereading a spreadsheet the day it comes to mind. The spreadsheet stays a good database. It is a poor alert system.

In a brokerage, the expiry date is when the relationship is played again. The policy reaches its term, the client compares, a competitor calls. If your first contact comes three weeks ahead, you are discussing terms. If it comes the day before, you are discussing deadlines.

The problem is almost never missing data: the date is in the spreadsheet, in the management software or on the policy. The problem is that nobody is warned at the right moment, with the right content, and that it all rests on a habit. A habit holds while the week is calm. It breaks during holidays, claim peaks or when a colleague leaves. The studio’s approach for brokers is described on the studio page.

What breaks in spreadsheet tracking

Four failures typically come back with spreadsheet tracking. None is dramatic, which is why they last.

• The date is entered once, then ages. An endorsement, an insurer change or a partial cancellation moves the real expiry date. The spreadsheet keeps the old one, and the reminder goes out at the wrong time, or not at all.
• Sorting is done from memory. Everyone looks at this week’s rows, but nobody looks at the ones two months out, which need the most preparation: offer comparison, information request, meeting.
• Every reminder looks the same. The same message goes out for household cover and for a vehicle fleet. The first is routine, the second deserves a call.
• Nothing records what was done. A reminder sent by email does not appear in the spreadsheet. A colleague chases a second time, or assumes it left when it stayed in drafts.

None of these is fixed by a tidier spreadsheet. They are fixed by moving the work: the machine spots and prepares, you decide.

The reminder calendar, step by step

Definition: a renewal reminder calendar is the sequence of contacts, dated relative to the expiry, that a firm plans for each policy: a detection, a preparation, a first message, a follow-up, then a call if the client does not answer. The exact dates depend on your portfolio. The public sources we read recommend very different lead times, from 45 to 120 days before expiry depending on the provider: keep the principle rather than the values.

1. Detect. Every day, the system reads the policy list and isolates those whose expiry falls inside a window you set, for example within the quarter. It also flags missing or inconsistent dates, which a spreadsheet never does.
2. Prepare the file. For each policy detected, it gathers the useful history on one sheet: previous years’ premiums, declared claims, latest exchanges, known documents. This is the longest part by hand, and the most mechanical.
3. Draft the message. It proposes an email suited to the client and the type of policy, in the firm’s usual tone. You see the text, you edit it or you reject it.
4. Approve and send. Nothing goes out without a person’s agreement. This rule is not cosmetic caution: it is what lets the machine work without you watching every line.
5. Follow up and escalate. With no answer after the planned delay, a second reminder is prepared. After the second, the line moves to a list of calls to make. The phone remains your tool, the machine only says when to use it.
6. Record. Every step is logged on the file: message proposed, approved by whom, sent when. The question “did we chase this client?” gets an answer in ten seconds.

The principle is the same as for a quote follow-up: a short sequence, dates set relative to an event, a human who approves. The tone advice in that article carries over almost unchanged to a renewal.

Segment: not every policy deserves the same treatment

The sequence above gains value as soon as it stops treating every policy alike. Three profiles are enough to start, and you can define them without any tool.

• The routine policy. Modest premium, no claim, stable client. The prepared message goes out after a quick approval, one follow-up comes next, and a call is only the last resort.
• The policy to watch. A claim declared during the year, a premium increase announced by the insurer, a client who asked questions. Here the prepared sheet matters more than the message: you want the facts in front of you before writing.
• The key account. A fleet, a business activity, several policies with the same client. The system warns early and prepares the file, but the first contact is a call or a meeting, never an automatic email.

This segmentation avoids the most common trap: a uniform reminder, read as a circular. It is set by simple rules you write yourself (amount, presence of a claim, number of policies for the client), not by an opaque model. If you cannot explain in one sentence why a policy is classed “to watch”, the rule is too complicated.

One last point on message content: messages reuse facts present on the file (date, insurer, type of policy) and nothing else. A prepared message that invents a cover or an amount is a serious defect, which is why human approval cannot be removed.

What the machine prepares, what stays with you

The last row matters most. This article is only about how work is organised: detecting, preparing, chasing, recording. Advice and the obligations specific to your profession stay in your hands, and nothing here replaces them.

A calculation to run before deciding

Take a fictional firm, to show the method and not to promise a result. Assume 800 policies renewing yearly, about 67 expiries a month. Assume ten minutes of preparation and follow-up per policy when all goes well, and twenty-five when the history has to be dug out. If a third of files fall into the second case, monthly time comes to roughly 67 × (2/3 × 10 + 1/3 × 25), about 1,000 minutes, a little over 16 hours.

These figures are assumptions. Yours have to be measured: take ten recent renewals and time what they took, from the first search to the message sent. Also note how many policies were chased late, and how many were lost while the client was happy. That last line often weighs more than the hours: a lost recurring commission is not recovered.

When this is not the right project

• Your portfolio is small. With a few dozen policies, a shared calendar and discipline are enough. Automation would cost more than it returns.
• Your dates are wrong. If the spreadsheet is unreliable, the machine will chase at the wrong time, faster than a human would. Clean the expiry dates first, a thankless but necessary job.
• Your key clients are handled by phone. For high-relationship accounts, a prepared message does not replace your call. Keep automation for volume, not for accounts that need your voice.
• You have not defined your process. Automating a practice each person does their own way freezes the confusion. Agree on the sequence first, on paper.

Confidentiality: where your clients’ data goes

A policy portfolio holds personal and sometimes sensitive data: claims, family situation, vehicles, business activity. Before connecting anything, ask three questions. Where is the data processed, in Switzerland or locally, or with a foreign provider? Who can read the sheets produced? Are access and sends logged? We cover this for small businesses in our article on Swiss data protection law and the data you send to an AI model. This article is not legal advice: validate the framework with the person responsible for compliance in your firm.

Where to start this week

Pick ten policies renewed in recent months. For each, find the real expiry date, the moment you first made contact, and what took time. Then write down, in a few lines, the sequence you would want running: how many contacts, on which dates, in what tone. These notes form the specification. Keep them short: one page is enough, and if it takes longer to write than the sequence takes to run, the process is still too vague. Share it with a colleague and see whether they would run it the same way. To place your case before deciding, see how the studio works on the processes it automates.