Blog · Wallet share and penetration · Insurance
Define comparable client peer groups and disclose small samples. Avoid turning a thin sector median into a confident client target.
An agency can calculate a sector median from three accounts, but the result may shift sharply if one account is unusual. Larger groups can also be misleading when they combine different client sizes or arrangements. The quality of a peer reference depends on comparability and evidence, not merely the ability to compute it.
One row represents: one client-period observation in a defined peer group, with the measurement and eligibility basis stated.
Useful fields: Client ID, broad sector, size band, relevant exposure band, observed measure, evidence status, observation date, group definition and inclusion reason.
Select attributes relevant to the question before inspecting the results. Report median, count and a spread measure where the data supports it. Avoid using the same client in its own benchmark for sensitive comparisons, or clearly state the convention. Keep missing external placements and unverified product tags outside confirmed-measure comparisons.
The following records and amounts are invented to show the method. They are not customer results, industry benchmarks or a forecast of Covirage performance.
| Peer group | Recorded lines per client | Median |
|---|---|---|
| Three comparable clients | 1, 2, 7 | 2 |
| After one additional client | 1, 2, 7, 8 | 4.5 |
| Mixed size group | Not consistently comparable | Do not use one target |
Adding a single observation changes the small group's median from two to 4.5 recorded lines. Neither number proves that a reviewed client needs that many policies. The sensitivity illustrates why the report should display count and spread and avoid a precise cross-sell target from a thin sample.
No universal sample-size threshold makes a peer group valid. Concentration, missing evidence and heterogeneous client needs can undermine a large group too. The agency should decide when to suppress a reference or show it only with a caution.
Bring a small, authorized sample to Covirage for insurance agencies and brokers. Use the sample to discuss the fields and views your business needs. A dashboard or AI analyst can help explore this question when the required data and definitions are available; missing records still need to be resolved.
Upload sample data to check its structure. Keep unnecessary personal, claims and policyholder details out of an initial sample. The sample check does not establish that every analysis in this guide is available automatically.
These references provide terminology or governance background. The worked example and proposed review method above are original illustrations, not prescribed industry standards.
No. It is a descriptive reference from the observed group. A client's appropriate arrangement depends on its own circumstances and qualified review.