How much of your account book has usable wallet evidence?

Share-of-wallet data coverage tells a reader how much of the account book the reported analysis can describe. A 52% book share based on customers representing half of revenue is a partial result. Showing only the share invites the reader to assume that every account has comparable evidence.

This is a coverage measure for the evidence behind wallet analysis. It does not replace the source-grade policy or the customer-level wallet calculation.

Start with an independently measured book

Use the customer ledger to establish included accounts and net category revenue. Record the reporting period, entity level, currency, active-account rule and any exclusions. Retain all included customers when you assess evidence, even if their wallets are missing.

A zero-sales prospect and an existing customer are different reporting populations. Combining them makes count coverage difficult to interpret. Show a separate prospect population if the review includes one, and keep credits or negative balances visible rather than turning them into artificial positive weights.

For revenue-weighted coverage, agree how negative-net accounts are handled before calculation. The simple example below contains positive revenue only; a book with significant returns needs a separately documented weight basis and reconciliation to signed net revenue.

Assign one evidence state per account

Keep current declaration, inferred estimate and unknown as mutually exclusive reporting states. Preserve more detail underneath: declaration date, customer-provided category, estimation method, freshness, missing size inputs and scope mismatch.

Stale evidence is not automatically unknown. A declared policy may reclassify it as inferred while a refresh is pending. An invalid denominator, such as a different category or missing period, should remain unusable until resolved. A numeric value alone does not establish fitness for the analysis.

Avoid a single opaque quality score that hides these distinctions. A reader should be able to identify the population with direct evidence and the population whose result depends on an inference.

Compute counts and revenue coverage separately

Count coverage = accounts with usable wallets / all included accounts

Revenue coverage = revenue of accounts with usable wallets / all included revenue

Declared-evidence revenue coverage = revenue with current declarations / all included revenue

Use the same population and revenue basis for every state. Count an account once at the declared decision level, even if it has several site records.

A synthetic evidence review

The following ten-account book has $1 million of positive annual category revenue. These amounts are invented for the method.

Usable count coverage is 7 / 10 = 70%. Usable revenue coverage is $800,000 / $1,000,000 = 80%. Direct declared evidence represents 45% of revenue, a different and more demanding disclosure.

Suppose wallets for the seven usable accounts total $1.6 million. Their measured share is $800,000 / $1.6 million = 50%. The numerator must be the same seven accounts' revenue. Dividing all $1 million of revenue by their wallets would incorrectly produce 62.5%.

Choose refresh work from materiality

Rank unknown accounts by measured revenue, approaching commercial decisions and the reason the evidence failed. A missing category scope may be resolved quickly; an unknown total spend may require a customer discussion. Record an owner, next evidence request and due date.

Use the customer-master checks to distinguish missing evidence from missing identity mapping. For inferred coverage, backtest the estimator before extending it to a much larger unknown population.

Publish the limitation with the result

Show the declared population, measurable population, excluded revenue and evidence-state table beside the wallet-share total. When coverage changes between periods, use a consistent population or disclose the effect; newly measured accounts can change the share without any commercial movement.

Microsoft's SUMIFS reference supports state-specific revenue totals, while its SUMPRODUCT reference documents weighted calculations. The coverage definitions above are the proposed review policy, not prescribed industry benchmarks.

Bring your ledger and wallet-evidence fields to Covirage to discuss a review that makes gaps in the evidence visible. Available results depend on the supplied records and the definitions agreed for your book.