croissant
Home / Blog / Reporting

Why Google Ads and GA4 conversion counts differ: an agency reconciliation guide

Croissant editorial team · · 5 min read

Compare Google Ads and GA4 conversion reports with a practical scope checklist, synthetic examples and clear review boundaries.

Different Google Ads and GA4 conversion counts do not automatically mean tracking is broken or leads were lost. Compare the action, report scope, date basis and settings first. Keep both values labeled until evidence supports a more specific explanation.

Start with a comparison contract

Before explaining a gap, record the Ads account and GA4 property, conversion action, report and column, date range, time zone, filters and extraction time for each number. Include whether each view uses interaction time or conversion time, its channel coverage and attribution settings. Similar labels do not establish that two tiles measure the same population.

Google says conversions created from Analytics events appear in the GA4 Advertising section when an Ads account is linked; standard Reports and Explore serve different analytical purposes. For campaign comparison, start with the relevant conversion performance view and selected action rather than assuming a website key-event count is an equivalent Ads conversion count. [1] Our guide to key events and conversions explains that initial choice.

Check dates and counting rules

A click near the end of a reporting period can lead to an action in the next. Google documents that its usual Ads conversion reporting attributes the conversion to the click date, while Analytics attributes it to the action date. Google Ads also offers conversion-time columns. Compare an appropriate conversion-time view when that is the question, but do not assume changing the column will resolve every difference. [2]

Inspect the selected action and counting method. Google documents One and Every counting options and says the Ads Conversions and All conversions columns have different coverage. A report containing secondary actions or different conversion sources should not be presented as if it had the same scope as a primary-action-only view. [2] Record actual settings; do not change them merely to make the numbers agree.

Google also notes that Analytics property and Ads account time zones can differ, and that eligible attribution channels can differ between reports. Check actual settings and date boundaries before assigning a cause. [3] Matching dates and action names does not establish that underlying populations are identical.

Worked example: a boundary and a count

This example is synthetic. A fictional agency reviews one imported enquiry action over Monday through Sunday in matching time zones. A person clicks an ad on Sunday and completes the form on Monday. Assume the interaction and action are eligible and all other relevant settings are aligned. A click-date Ads view can place the conversion in the first week while an action-date Analytics view places it in the next. This illustrates a date-basis difference, not evidence about any real account. [2]

Now consider a separate hypothetical: one eligible ad interaction is followed by two form completions. An Analytics key-event count records two, while an imported Ads action configured to count One may report one. Google gives this type of example; other eligibility rules still matter. [2] Test the scenarios separately. Do not add their differences or claim either explains a real gap without account evidence.

A client note might say: ‘These figures use different date bases and are not directly comparable. We are checking the same action in a conversion-time view and will report any remaining difference separately.’ The reporting owner should verify this statement against actual exports before the account lead approves it.

Build a repeatable investigation

Preserve both source snapshots and extraction times. Confirm the intended account and property, then list each selected conversion action and whether the Ads column includes it. Match the reporting interval and time zone; label any unavoidable difference. Compare the action's conversion-time view with the relevant Analytics conversion view where available, while keeping the original view for the campaign question it serves.

Inspect counting method, attribution channel eligibility, report filters and conversion windows. Google notes that settings for Analytics-based conversions generally match by default, but eligible channels and account time zones can differ. Treat defaults as a starting point for inspection, not a substitute for checking this account. [3] Allow for source processing and record when each extract was taken; a newer export does not automatically prove the earlier report was wrong.

Log findings as confirmed, possible or unresolved, with evidence and an owner. If values remain different, avoid assigning the whole gap to one setting. Do not subtract the numbers and call the result lost leads, or sum them as separate outcomes: they may overlap and follow different rules. If a report was issued, use the client report corrections guide to decide whether a reviewed revision is needed.

Automate checks and keep human judgment

A small team can start with a shared comparison sheet. A recurring workflow can collect approved report metadata, check for an unexpected action or date basis, and route mismatches to the reporting owner. Test it with fictional cases: matching inputs, a Sunday click with a Monday action, a One-versus-Every mismatch, an unavailable export and an unexplained gap. For the last two, create a review task rather than a confident explanation.

AI can summarize verified findings and label uncertainty, but a specialist should review source evidence and any proposed configuration change. The account lead approves client wording; an authorized person separately decides tracking, bidding or budget changes. A reconciliation workflow should not alter settings simply because two numbers differ.

Explore the reporting demo for a scoped workflow discussion. Bring a fictional comparison and the decisions your team needs to make through the contact form, without client exports or credentials.

Implementation checklist

  • Name each account, property, action, report and column.
  • Record dates, time zones, filters, extraction times and date basis.
  • Check action coverage, counting method, attribution channels and windows.
  • Compare suitable conversion-time views without assuming they must match.
  • Separate confirmed causes from possibilities and unresolved differences.
  • Test boundary, counting and missing-source cases with synthetic data.
  • Require human approval for explanations, configuration changes and delivery.

The sources behind the sauce.

Primary research and official documentation behind the cited points. The workflow examples and checklists are Croissant’s own guidance.

  1. Google Analytics Help — Conversions vs. key events in Google Analytics
  2. Google Ads Help — Create conversions from Google Analytics events in Google Ads
  3. Google Analytics Help — Select attribution settings