Client Retention Reviews for Broking Firms: A Quarterly Process That Protects the Book

A broking firm learns it has lost a commercial client at renewal, when the client renews elsewhere or a broker-of-record letter arrives naming someone else. By then nothing can be done, because the decision was made weeks or months earlier and the renewal is only its paperwork.

That timing is the whole problem with how most firms manage retention. They manage it as a renewal event, which means they manage it at the one moment they have no room to change the outcome. The account was winnable in the quarter before, when the client was quietly dissatisfied and had not yet talked to a competitor. It was defensible when the claim went badly, when the third service request went unanswered, when the new CFO arrived with a broker she already trusts. Each of those was a signal, and each passed unnoticed because the firm was not looking until the renewal calendar told it to.

Retention as a process means looking earlier and on purpose. It is a standing quarterly review that reads the leading indicators of defection across the book, names the accounts at risk while there is still time to act, and treats every loss as data about the next one. This is deliberately the low-technology version. There is a place for the modelled, insurer-side churn prediction that scores a whole portfolio, but a mid-size broking firm does not need a model to protect its book. It needs a meeting, a short list of signals, and the discipline to act on an at-risk account before the client has stopped taking its calls.

A commercial client sends signals before it leaves. They are not subtle once you decide to watch for them, and the value of naming them is that they can be logged and reviewed rather than felt in hindsight.

The reliable indicators, roughly in order of how strongly they predict a loss:

1. A claim that went badly. A disputed, delayed or rejected claim the client blames the broker for is the single strongest predictor of defection. A client forgives a hard renewal; it does not forgive feeling abandoned in a loss. Every friction-laden claim is a retention event, which is why claims service and retention are the same problem, worked in the claims-servicing SLA piece.
2. Unreturned or slow service. Endorsement requests that sit, certificate-of-insurance requests that arrive late, queries answered in days rather than hours. Service latency is felt as neglect, and neglect compounds.
3. A new decision-maker. A new CFO, procurement head or risk manager arrives with existing broker relationships and a mandate to review costs. The incumbent broker is the default thing a new finance chief re-tenders.
4. Competitor activity. An RFP floated, a rival broker seen at the client, or a request for policy copies and claims history that hints at a broker-of-record move in progress.
5. Going quiet. A client that used to call and now does not, that declines the review meeting, that stops sharing information. Disengagement precedes departure.
6. A single-line relationship. An account where the firm holds one policy is structurally easy to displace, because the client has no switching cost and no breadth of relationship to lose.

None of these requires software to spot. They require that someone writes them down against the account when they happen, so the quarterly review has something to read.

The engine of the process is a standing meeting, once a quarter, whose only job is to identify accounts at risk and assign action. It is short, structured, and separate from the renewal-planning meeting, because renewal planning is about execution and this is about early warning.

A workable structure:

• Attendees. The servicing team leads and a member of management who can authorise a service intervention or a fee concession within policy. Retention decisions that need sign-off should not wait for a second meeting.
• Input. A pre-circulated list of accounts flagged on any leading indicator during the quarter, each with the signal noted. The flags come from the servicing team's own logging, not from a special exercise done the night before.
• The rating. Each flagged account is rated on likelihood of loss and value at stake. A simple red-amber-green on likelihood, crossed with the account's revenue band, is enough to sort where attention goes. A red-likelihood, high-value account is the meeting's priority; a green, low-value one is noted and left.
• Output. For every red and amber account, a named owner, a specific action, and a date. Not "stay close to the client." A service audit by a named person within two weeks, a senior-level meeting requested, a claims escalation opened, a coverage review offered.

How a firm measures retention determines what it protects, and the common measure protects the wrong thing. A firm that reports it retained 95 percent of its clients is counting accounts, and counting accounts flatters the book, because the accounts a firm loses are rarely its smallest.

The number that matters is retention by revenue: the proportion of last year's commission that renewed, weighting each lost account by what it actually paid the firm. The gap between the two measures is routinely wide. A firm that lost five accounts out of two hundred looks like it retained 97.5 percent by count. If two of those five were large accounts, it may have retained only 85 percent by revenue, and the 12-point difference is the entire story the count measure hid.

Track both, and manage by revenue:

• Retention by count tells you about the health of the relationship base and about churn in the long tail of small accounts.
• Retention by revenue tells you what actually happened to the firm's income, and it is the number that should drive where the at-risk review spends its attention.

The practical consequence is that the at-risk meeting weights its effort by value at stake, not by account count. Losing ten small single-line accounts and losing one large multi-line account are the same on a count measure and nothing alike on a revenue measure. A firm managing to revenue puts its scarce retention effort on the account whose loss would actually hurt, which is also usually the account with the most relationship to save. This is the same weighting logic that governs concentration risk in the commission book: the accounts that dominate revenue are the ones whose defection the firm can least afford not to see coming.

Identifying an at-risk account is worth nothing without a defined response, and the response has to match the signal that raised the flag. A generic reach-out to a dissatisfied client can make things worse by signalling that the firm noticed the problem and did nothing specific about it.

Match the recovery to the cause:

• Where the flag is a bad claim, the recovery is a claims intervention: a senior person takes the claim personally, gets the client a clear status and a realistic timeline, and, where the client has a genuine grievance, advocates it hard with the insurer. The client needs to see the broker fighting for them, which is the thing they felt was missing.
• Where the flag is service latency, the recovery is a service audit: pull every open item on the account, close the backlog, and put a named service contact and a response standard in place, then tell the client what changed.
• Where the flag is a new decision-maker, the recovery is a relationship reset: a senior meeting to re-establish the firm's value to a person who did not choose it, before the person defaults to a broker they did.
• Where the flag is competitor activity, the recovery is a value case: a review that shows what the firm has delivered, where the programme could improve, and why continuity is worth more than a first-year price cut, made before the client is committed rather than after.

The common thread is specificity and speed. An at-risk account has a short window in which intervention still reads as care rather than damage control. A firm that acts within weeks of the flag holds accounts a firm that acts at renewal has already lost.

Some accounts leave despite everything, and the firm that learns nothing from a loss is condemned to repeat its cause. Loss-reason coding turns a departure from a disappointment into an input.

When an account leaves, code the reason from a fixed, short list rather than recording it in free text that never aggregates:

• Price. The client moved for a cheaper placement or a broker who promised one.
• Service or claims. The client left over how it was handled, in a claim or in day-to-day servicing.
• Relationship. A new decision-maker brought their own broker, or the relationship owner left the firm and took the account.
• Competitor displacement. A rival won a formal tender or a broker-of-record move on breadth or specialism the firm lacked.
• Client event. The client was acquired, closed, relocated, or restructured in a way outside the firm's control.
• Coverage or capability. The firm could not place a line the client needed, or lacked expertise in a class the client valued.

The value is in the aggregate. One loss coded price is a lost account. A quarter where six of ten losses code service or claims is a systemic failure the firm can fix, and one it would never see from six separate post-mortems. Review the coded losses at the same quarterly meeting, and let the pattern drive change: a run of claims-coded losses funds a claims-service investment, a run of relationship-coded losses funds a succession plan for key relationship owners. A loss you coded is a lesson. A loss you sighed about is not.

A lost account is not a closed file, and treating it as one forfeits some of the most convertible new business a firm can write. A client who left knows the firm, and the firm knows the client, which strips out most of the cost of winning a stranger. What win-back needs is timing, because the same approach lands very differently at different moments.

There are two windows worth working:

1. Immediately, where the loss was a mistake. If the account left over a specific, fixable failure, or over a competitor's promise that will not hold, a prompt, honest approach that owns the failure and shows what changed can recover it before the new arrangement beds in. This is rare and time-limited, and it depends on the firm having coded the loss honestly enough to know the failure was real.
2. Around the next renewal, where the loss was to a competitor. A client who moved to a rival broker is most winnable roughly ten to fourteen months later, as the new broker's first-year attention fades and the client discovers whether the promises held. A firm that stays lightly in touch through that year, without pestering, is positioned to be the alternative when the client's second renewal under the new broker comes into view.

Run a win-back pipeline the way the firm runs its new-business pipeline: lost accounts with the loss reason, the departure date, and the next renewal date, reviewed so the approach lands in the window rather than at random. The firms that do this find that a meaningful share of their new business each year is old business returning, won at a fraction of the cost of a cold pursuit, because the relationship and the risk knowledge were never fully lost, only paused.