Measurement
Your reports may disagree for a reason.
Advertising, website analytics and CRM systems can count different events under the same familiar label. Reconciliation starts with definitions.
Ask what each number actually represents.
An advertising report can credit an interaction within a defined attribution window. Website analytics records observable activity under its collection and consent conditions. A CRM records the stages entered by the business. “Conversion” is not a universal unit across them.
Before investigating a discrepancy, write the event, period, attribution rule and system behind each number.
Separate collection from attribution.
Collection asks whether the event was recorded correctly. Attribution asks how credit was assigned. A duplicated event is a collection problem. Two platforms crediting the same sale is an attribution issue. They need different investigations.
Reconcile in a useful order.
- Align dates, time zones and reporting periods.
- Confirm that the events mean the same action.
- Check duplicate firing and missing collection.
- Review attribution windows and crediting rules.
- Compare with CRM records where a reliable match is possible.
Do not force agreement by hiding the difference.
Some differences should remain visible because the systems answer different questions. A useful report explains the relationship and the remaining gap rather than changing definitions until the totals match.
Similarly, a matched customer record does not prove that a particular exposure caused the purchase. Attribution evidence and causal evidence are different.