Why do two reports disagree about the same product campaign?

Write down what each number means

For an illustrative launch review, one report might show clicks, another sessions and another attributed purchases. Those are different units. Record the metric name, date range, time zone and selected campaign before exporting the totals. Keep the raw report definitions with the review.

Google Analytics documents attribution settings and eligible lookback periods. Check the actual configured rules rather than assuming that an advertising platform and website report assign credit in the same way. Do not change the rules simply to make the totals match.

Create a reconciliation sheet

Resolve a definitional difference separately from a missing event or broken link. A combined total can conceal which problem remains.

Compare equivalent slices

Use the same product campaign and a clearly defined interval. Check whether refunds, duplicate events or later actions are included according to the relevant system. Ask the analytics owner to investigate implementation defects; creative generation cannot repair reporting configuration.

State the practical limitation

The final review should name the chosen reporting basis and any unresolved mismatch. Do not merge incompatible dashboards into one unsupported return claim. Use creative comparison for production variables and accepted-asset cost for a separately defined production metric.

A mismatch that needs definitions

A fictional campaign review compares an advertising dashboard’s attributed purchases with a website report’s sessions. The totals disagree because the metrics are not equivalent before any implementation issue is investigated. Write the counted action for each, align dates and inspect the configured attribution rules. If a mismatch remains after equivalent slices are compared, ask the measurement owner to investigate. Keep both original report definitions. The purpose is to produce a defensible comparison basis, not to adjust settings until a preferred creative appears profitable.