Practical guide

Reconcile counts before returning results

Last materially reviewed 2026-09-27

Quick answerEvery expected row needs an accounted-for outcome—even when that outcome is held or unknown.
What to know

Count the right denominator

Begin with the defined input batch and its documented duplicate policy. Compare expected result rows with observed result rows, then account for valid exclusions or transformations separately. A matching total is necessary in many workflows but does not prove that individual results map correctly.

What to know

Investigate before patching

If totals differ, look for duplicate output rows, missing records, filters or a partial download. Do not add blank rows to make totals match. Use the original snapshot and mapping rule to identify the discrepancy.

What to know

Use the local worksheet

Our worksheet accepts only aggregate counts. Enter expected rows and mutually exclusive result buckets; it calculates the difference and unclassified count. A zero difference means arithmetic reconciliation only. It is not a verification service, a permission check or evidence that each row was correctly matched.

What to know

Put it into practice

In a fictional batch, the input contains 1,000 distinct expected result rows. The four reviewed buckets contain 900 favourable, 50 unfavourable, 30 risky and 20 unknown results. The total is 1,000, and the uncertain portion is 50. That establishes arithmetic completeness only. Now imagine one row appears twice and another is missing: the same total can still be produced. This is why the next control is a key-level exception report, not an automatic return to the live list. If input deduplication intentionally changed the expected output count, record that transformation before using the worksheet. Do not subtract unknowns from the denominator to make a completion percentage look better. Keep the counts, the mapping rule and the unresolved action queue in the same receipt. Each answers a different question and none alone proves a safe production update.

Continue when useful

Next: Join results to records with an exception report

A successful join should expose unmatched and multiply matched rows instead of hiding them.

Open Join results to records with an exception report →

Sources used for this page

These records support the facts and comparisons above. Merchant-controlled records are labelled so you can separate product claims from independent evidence.

  1. Bouncer API introduction — Merchant documentation · docs.usebouncer.com · Merchant-controlled · checked 2026-09-27