Start with the business outcome
Check the enquiries or orders in your own records, then compare how each marketing report credits them. A platform result and a business total can answer different questions.
Know the event and reporting window
A visit, button click, saved enquiry and purchase are different events. Confirm which one your report counts and how far back it looks for an advertising interaction. Keep the account’s current settings beside the report.
Compare the current attribution labels
Meta announced changes in March 2026 separating link-click credit for website and in-store conversions from other social interactions under engage-through reporting. The announcement described a phased rollout, so check the labels and settings available in your account when comparing periods. Source: Meta’s measurement update (Spanish).
Why Meta and website analytics differ
Tools can observe different interactions, apply different credit rules and use different reporting scopes. Date ranges, time zones, consent choices, devices and event duplication can also affect the totals. Compare equivalent measures before investigating the difference. Source: Google Analytics traffic-source scopes.
Count each order once
Illustrative example: a customer sees an ad, later clicks an email and places one £100 order. Meta and an email report may both credit an interaction, but the business has one £100 order. Reconcile channel reports rather than adding their revenue columns together.
Read ROAS alongside business costs
ROAS divides attributed purchase value by advertising spend. Use it alongside margins, fulfilment, fees and returns when assessing the return to the business. Incrementality asks a further question: which outcomes happened because of the advertising? A suitable lift study can help answer that.
Build a useful report
Include spend, dates, the event counted, attribution settings and the matching business outcomes. Explain the campaign changes and the next action in plain language.
Reconcile a reporting difference
Put the business total beside each platform total for the same completed period. Record timezone, event definition, credit window, reporting scope, consent coverage and refunds or cancellations. In the £100 example, keep one order in the business total even when two channels claim credit. Label any unexplained difference and the next check rather than changing one report until the numbers match.