Reconcile Google Ads and GA4 counts with a definition worksheet

Name the event before comparing totals

A report labeled conversions may include a different set of actions from another report labeled key events. Open the definitions and identify the actual outcome: accepted form, contact tap, booking, or imported qualified lead. Do not compare broad all-actions totals with one inquiry action. GA4 distinguishes events, key events, and conversions, and that vocabulary does not establish a customer's business outcome without a verified measurement path.

Source checks: Understand your conversion tracking data · Create conversions from Google Analytics events in Google Ads

Write a definition row for each report

Record platform, report, action names, included channels, date basis, attribution setting, counting method, timezone, and export time. Include whether modeled or imported data is involved when relevant to the configured report. Keep the worksheet with the review. This prevents a later manager from recreating the comparison using different settings and concluding that the unexplained discrepancy has returned.

Align dates before diagnosing missing outcomes

Google Ads' standard conversion columns commonly assign outcomes to the ad-click period; conversion-time columns organize them when the outcome occurred. Choose the basis appropriate to the comparison. A business inbox is usually organized by arrival time, so a click-date report needs a separate cohort view rather than a direct same-day equality check. Also allow for processing and the normal delay between an inquiry and an imported qualification.

Match counting and attribution scope

Check whether repeated events count once or more than once under the relevant settings, and whether the compared report covers only paid-attributed outcomes or all channels. A native Ads tag and a GA4 import may describe the same underlying request; do not add them as separate customers. Confirm linked accounts and tag behavior where a technical defect is suspected. A difference caused by scope should be explained, not repaired by inventing additional leads.

Reconcile to delivered records without exporting personal details

Use the restricted inquiry or booking system to count accepted, unique records in the same operational period. Aggregate the reconciliation for the owner: verified records, platform-attributed outcomes, duplicates or tests, and unresolved difference. Avoid placing names, emails, or customer problem details in analytics parameters or a public report. If matching is incomplete, say so; an unmatchable record is not automatically a lost customer or a broken tag.

Document the answer rather than forcing equality

Separate explained differences from confirmed defects and still-unresolved gaps. Assign a concrete check for each defect, such as duplicate firing or lost identifiers. Keep the comparison settings stable at the next review. The practical outcome is a reproducible accounting of what the numbers measure. Reports can remain legitimately different while the business accurately tracks delivered inquiries and bookings.

Worked example: illustrative figures

For example, an ad is clicked Monday and an accepted form arrives Thursday. The standard Ads click-date view places the conversion with Monday's traffic, while a conversion-time view can place it on Thursday. Separately, GA4 records two events for a repeat submission but the business stores one unique inquiry. A Monday-to-Thursday comparison can therefore disagree for both timing and counting reasons. The reviewer documents those differences before calling any outcome missing. The example does not establish that every real platform gap has the same cause.

Questions about this task