Running a PI Tender: How Larger Firms Should Test the Market

At some point — a new finance director, an audit committee question, a renewal that stung — a larger firm decides to test whether its PI arrangements are as good as they should be. The instinct is to run a tender: invite several brokers, let them compete, pick the winner. Done well, that process is genuinely useful. Done the way it is usually done, it can leave the firm with worse terms than it started with, for reasons that are structural rather than anyone’s bad faith. This page explains the mechanics, because the mechanics are where tenders are won and lost.

A tender is not a review — decide which you actually need

First, a distinction that saves a great deal of wasted effort. A broker tender asks: who should represent us to the market? A programme review asks: is what we buy structured, worded and priced correctly? They are different questions. If your real concern is whether the programme is right, a review — which can be done without disturbing the market at all — answers it faster and more cheaply than a tender, and may conclude that the incumbent broker is doing a perfectly good job. If your concern is the broker relationship itself — service, expertise, conflicts, fees — then a tender is the right instrument. Firms that run a tender to answer a review question tend to get noise: three glossy documents, each optimised to win the appointment rather than to diagnose the programme.

How many brokers — and why more is worse

The intuition from every other procurement exercise — more bidders, better price — fails in insurance broking, and it fails because of how the underwriting market works. A broker does not manufacture the product; every broker is approaching the same finite panel of insurers on your behalf. When an underwriter receives the same risk from two brokers, the standard response is to quote through neither until the firm designates one, or to release identical terms to both — at which point the competition between your brokers has produced nothing except confusion about who controls the quote. This is market blocking: each additional broker does not add capacity, it freezes it. An underwriter who has seen a risk shopped by four houses may decline to spend serious time on it at all, reasoning that the submission is being sprayed rather than presented.

The practical optimum for a substantial PI risk is small: the incumbent plus one or two challengers is usually enough to create genuine tension without degrading the firm’s standing in the market. Beyond that, each addition subtracts.

Exclusive market allocations: the mechanism that makes a tender work

The professional solution to blocking is to allocate markets exclusively before anyone approaches anyone. Each participating broker is assigned specific insurers — in writing, agreed by the firm — and approaches only those. The incumbent typically retains the holding markets; challengers are allocated insurers where they can credibly claim access or relationships the incumbent lacks. Every insurer sees the risk once, from one broker, presented properly. The firm then compares real, uncontaminated alternatives rather than three versions of the same quote.

Allocation has a second effect worth understanding: it forces each broker to reveal, before the process starts, where they actually believe they can add value. A challenger who asks for the same markets the incumbent already holds is telling you something. So is one who requests markets nobody else considered.

The tender pack: what to give them and what to ask

The quality of responses is capped by the quality of the pack. It should contain the current programme structure and schedule, the claims summary (insurer-produced confirmed claims experience, not the firm’s recollection), a description of the firm’s work and its split by discipline and client type, the contractual limit requirements the programme must satisfy, and the renewal date and decision timetable. It should ask each broker for: proposed markets and why; proposed structure if different from current, with reasoning; the team who would actually service the account, by name; claims-handling approach and who advocates when a claim is disputed; remuneration, disclosed fully — fee, commission, and any income from insurers or facilities; and references from comparable clients.

What the pack should not do is invite speculative premium indications on incomplete information. An indication produced without full underwriting information is a marketing number, and comparing marketing numbers rewards the most optimistic guesser rather than the best broker.

Comparing responses beyond the premium column

Premium is the most visible line and the least durable. It reflects one year in a cyclical market; the appointment will likely run for many. The comparisons that repay attention: wording quality — not the brochure summary but the actual policy language on the definition of claim, notification, defence costs and aggregation; claims handling — who inside the broking house fights a reserved or disputed claim, and what their record is; insurer security — the ratings and the longevity of the proposed panel, because cheap capacity that exits in eighteen months converts a saving into a re-placement problem; and service structure — whether the people in the presentation are the people on the account. A response that wins on premium and loses on all four of these is not the winning response.

Conflicts deserve explicit questioning. Ask each broker what they earn from insurers as well as from you: profit-sharing arrangements, placement into facilities the broker operates, work placed with a broker-owned managing general agent. None of these is improper when disclosed; all of them shape which market a recommendation favours. A broker’s answer to “who pays you, and for what?” tells you a great deal about the advice that will follow. An independent broker remunerated transparently has the fewest reasons to steer the outcome.

Frequently asked questions

How often should a larger firm tender its PI broking?

Rarely, and deliberately. A tender every year destroys the relationship value that makes a broker effective, and the market notices serial tenderers. Every three to five years, or on a genuine trigger — a service failure, a merger, a material unexplained cost movement — is a defensible rhythm, with a lighter-touch programme review in between.

Can we just ask two brokers to quote and see who comes back cheaper?

That is precisely the process that triggers market blocking: both will approach the strongest insurers, those insurers will freeze or duplicate, and the comparison you receive will be distorted. If you want competing quotes, allocate markets exclusively first — it is one page of correspondence and it changes the entire quality of the exercise.

Does the incumbent broker have an unfair advantage in a tender?

The incumbent holds real advantages — the claims history, the relationships, the renewal data — and a well-run tender neutralises what it can by putting complete information in the pack for everyone. The remaining advantage is legitimate: knowing the account is part of what you are evaluating. The purpose of a tender is not to handicap the incumbent but to test whether familiarity has become complacency.

Apex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority (FRN 724952). Every risk is different: nothing on this page is advice on your own programme, and outcomes depend on your firm’s circumstances and the market at the time.