The Loss-History Data Room: Why Well-Kept Claims Records Buy Better Terms

When an underwriter prices a commercial renewal, the single most important input is the account's own loss history. It drives the experience rating, the burning cost, the view of the risk and, in the end, the terms. What many buyers do not register is that this history exists in two versions, and it matters a great deal which one the underwriter reads.

The first version is the insurer-held loss run: the record the current insurer produces from its own claims system. It is often incomplete, coded for the insurer's purposes rather than the buyer's, silent on the causes and the remediation, and produced late and grudgingly when the account is going to market. If that is the only record on the table, the underwriter prices off a document the buyer neither controls nor can fully explain, and every ambiguity in it is resolved against the buyer.

The second version is the buyer's own standing loss-history data room: a record the company maintains itself, complete, coded to its own risks, current, and able to tell the story behind every loss. A buyer who brings this to a renewal frames its own risk. It presents the losses in context, shows what was done about each one, and gives the underwriter a clean, credible basis to price on. The difference in terms between the two situations is real, and it is available to any buyer willing to keep its own records rather than rely on the insurer to keep them. This is not about tooling to extract loss runs; it is about the buyer-side discipline of owning its claims data and the pricing payoff that follows.

A loss-history data room is not a folder of past claim letters. It is a structured, maintained record designed to answer the questions an underwriter will ask, before they are asked.

At its core it holds a multi-year claims record by line of business, conventionally five years, because that is the window most underwriters weight and the period over which a credible experience emerges. For each line the record shows the claims arranged so that development is visible: the year of loss, the amounts paid, the amounts still outstanding or reserved, and how each has moved over time. Arranged as claims triangles, by year of loss against development period, this shows not just what has been paid but how the account's losses develop, which is what an experienced underwriter actually reads.

Around that spine sit the fields that turn numbers into a risk story:

• Cause-of-loss coding, so every claim is classified by what caused it (fire, water damage, theft, machinery breakdown, third-party injury), letting both the buyer and the underwriter see concentrations and trends rather than a flat list.
• Closed-versus-open status, so it is clear which claims are settled and final and which are still developing, because an open claim carries uncertainty that a closed one does not.
• Risk-improvement actions taken after each loss, recording what the company changed in response, which is the field that converts a loss from a liability into evidence of a managed risk.

Maintained this way, the data room is a live asset, not a renewal-time scramble. It is assembled once and kept current, so that at any renewal or remarketing the company can present a complete, coded, up-to-date picture of its own risk without reconstructing it from scratch under deadline pressure.

An underwriter presented with a clean, complete, well-coded loss history reads more than the numbers. The presentation itself is a signal, and it works in the buyer's favour before the loss figures are even assessed.

A company that can produce a five-year record by line, coded by cause, with open and closed status clear and remediation documented, is telling the underwriter that it understands its own risk, tracks its own losses, and manages the account deliberately. That impression matters, because underwriting a large commercial account is partly a judgement about the quality of the risk management behind it, and a buyer that keeps this kind of record is demonstrating exactly the discipline an underwriter wants to see. The data reduces the underwriter's uncertainty, and lower uncertainty supports better terms.

The reverse impression is just as powerful. A buyer that arrives with a partial insurer loss run, no cause coding, unclear claim status and no account of what was done about past losses signals the opposite: an account that is not closely managed, whose true experience is uncertain, and whose future losses are therefore harder to predict. The underwriter prices that uncertainty. So the presentation of the loss history is not a formality that precedes the real pricing; it is part of the pricing, because it shapes the underwriter's confidence in every number that follows. Clean data does not just inform the terms, it improves them.

The cost of a poorly kept loss history is not neutral. Gaps and disputes in the record do not leave the underwriter with a blank they ignore; they leave the underwriter with an uncertainty it charges for.

An underwriter faced with missing years, unexplained large claims, claims of unclear status, or figures that do not reconcile has to make an assumption, and the prudent assumption is the conservative one. A gap in the record is read as a possible hidden loss. An open claim with a vague reserve is read at or above its stated outstanding. An unexplained spike is read as a risk that could recur. Each of these is an uncertainty loading: a margin the underwriter adds precisely because it cannot see clearly, and the buyer pays for the poor data quality through the terms.

Disputed entries are worse than gaps, because they introduce doubt about the buyer's credibility as well as its risk. A loss history that the buyer contests in part, or that contradicts the insurer's own record, forces the underwriter to decide which version to believe, and the resolution is rarely in the buyer's favour under time pressure at renewal. The remedy is to find and fix these problems before the account goes to market, not to argue them at the underwriting table.

The field that most distinguishes a loss-history data room from a bare claims list is the record of what the company did after each loss, and it is the field with the largest effect on terms.

Every significant loss is, to an underwriter, a data point about future risk. But a loss followed by a documented remediation is a very different data point from a loss followed by nothing. A fire followed by a sprinkler upgrade, better hot-work controls and a revised storage layout tells the underwriter that the specific cause has been addressed and is less likely to recur. A theft followed by improved security and access controls tells the same story. The loss happened, but the risk that produced it has been reduced, and the underwriter can price the improved risk rather than the raw loss.

This is why the remediation record has to be kept as the losses happen, not reconstructed at renewal. A company that logs, at the time, what it changed after each loss builds a running account of a managed risk that improves with every event. A company that tries to assemble the same story under deadline pressure produces something thinner and less credible, and often finds the details are no longer available. The remediation story is only convincing if it was recorded when it was true.

A loss-history data room is only as useful as it is usable, which comes down to the formats it holds the data in, the reconciliation that keeps it honest, and the governance that keeps it alive.

On format, the record should hold the raw insurer loss runs by line and period, and also a standardised summary the company maintains itself: the multi-year record by line, the claims triangles showing development, and the cause, status and remediation fields. Insurers accept loss data in the formats they are used to, so the data room should be able to produce the standard claims-experience presentation an underwriter expects, while also holding the fuller record the company keeps for its own management. The point is to be able to hand a market a clean, complete, credible pack without reconstructing it.

On reconciliation, the buyer's record and the insurers' records must be kept consistent, because a discrepancy discovered at renewal is a problem and one discovered a quarter earlier is an email. Periodic reconciliation against the insurer loss runs, resolving differences as they arise, keeps the data room trustworthy and removes the credibility risk before it matters.

On governance, the record needs an owner and a cadence. The risk or insurance function should own the loss-history data room, with a defined update cycle, quarterly for a large account, so that new claims, movements on open claims and remediation actions are captured close to the event rather than at renewal. A record with no owner decays, and a decayed record is worse than none because it looks complete while missing exactly the recent movements that matter. Owned and updated on a cadence, the data room is a standing asset that makes every renewal and every remarketing start from the company's own evidence.

The payoff is straightforward and repeatable. A buyer that maintains a clean, coded, reconciled, remediation-rich loss history walks into every renewal framing its own risk on its own evidence, reduces the underwriter's uncertainty, and earns terms that reflect a well-managed account rather than an unclear one. The record costs a modest, standing effort to keep. The alternative, letting the insurer's incomplete loss run be the only version of the company's own history, costs a margin on the terms at every renewal, which over the years is far more than the record would ever have cost to maintain.