Blog · Data quality and reconciliation · Insurance
Prepare policy, client and income exports for agency analytics. Match periods, keys and totals before uploading a sample.
An agency may hold all the right files and still be unable to answer a simple growth question. A policy register describes cover, a transaction export records movements, and an income ledger records agency revenue. Treating these files as interchangeable makes the dashboard harder to trust. Begin with one defined question, such as which renewed accounts changed commission income.
One row represents: a single record at the native level of each source file, with that level documented separately.
Useful fields: Source system, extract timestamp, period start and end, client ID, policy term ID, transaction ID, line, producer ID, premium, commission, fee and currency where available.
Write a short file inventory before joining anything. For each export, identify a unique key, the date that controls inclusion and an independently obtained total. Keep original column names in the source copy. Do not request a claims file for a revenue question merely because it is available. Record excluded branches or inactive records so the perimeter is explicit.
The following records and amounts are invented to show the method. They are not customer results, industry benchmarks or a forecast of Covirage performance.
| File | Rows | Control |
|---|---|---|
| Policy terms | 240 | 240 unique term IDs |
| Income transactions | 510 | $82,000 commission and fees |
| Client master | 180 | 180 unique client IDs |
The three row counts should not agree: one client can hold several policies and one policy can generate several transactions. The first acceptance test is that every included transaction maps to a known policy or an explained exception. A match between 510 transactions and 240 policy terms is expected; an unexplained join that produces 900 transactions is not.
An export proves only what the source recorded at that time. It cannot establish complete competitor placements, missing activity or correctly classified revenue. A familiar file extension also does not guarantee that its columns can be mapped without review.
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. A small authorized export can establish column structure and definitions. A production analysis may need more history or related files after the question is agreed.