GA4 Key Events vs Google Ads Conversions: Compare the Right Numbers

When GA4 shows 16 purchases and Google Ads shows 20 conversions, which number should guide the next campaign decision? Choosing whichever looks more credible will not resolve the difference. You first need to know what each report counts—and whether two conversion actions are giving credit for the same purchases.
We separate three questions: what happened across the store, what Google Ads credited, and which outcome the campaign uses for bidding. The aim is an explained comparison and a deliberate bidding choice. Two matching totals would not, on their own, establish either.

Give each report a clear comparison basis
A GA4 key event marks an important recorded behavior. A Google Ads conversion created from an Analytics event is a related advertising measurement, not a replacement name for every key-event total.
Google's explanation of conversions and key events distinguishes these views. GA4's Advertising section can show conversion performance, while standard Reports and Explore use key-event reporting rather than Google Ads conversion metrics. Where available, start with the Advertising conversion report when checking the corresponding advertising conversions.
Label each export with its product, report, metric, filters and date basis. “Purchases across all channels” and “conversions credited to this campaign” describe different sets of activity. Selecting June 1–7 in both reports does not remove that distinction.
Check the property you actually use. Controls and availability differ, and Google's streamlined management feature excludes subproperties and roll-up properties. The useful first step is identifying the available report and its meaning, not hunting for a control shown in someone else's screenshot.
Map the actions to the purchase they represent
A store may have a native Google Ads purchase action and a GA4-sourced purchase action in Google Ads. Google's Analytics-to-Ads conversion guidance supports both routes. Their coexistence can be useful for comparison; it does not make the two totals additive.
Put the actions side by side. For each, record its ID, name, source, business event, first import date where relevant, and campaigns using it. Then ask what makes an order enter that action. Similar names are a clue; the event mapping and implementation establish whether the actions describe the same purchase.
Add the current optimization role to this record. Actions created through the Analytics interface start as secondary, but later edits can change that role. Check the present setup instead of assuming that an imported action is automatically bidding or permanently excluded. Pre-import historical data is absent too, so compare periods both routes actually cover.
Separate 16 store purchases from 20 action credits
This is a fictional teaching example, not client performance. A store's June 1–7 order register contains 16 completed purchases. Its GA4 all-channel purchase key-event count is also 16. We assume one valid event per purchase, fully processed data and no modeled or fractional credit.
To compare with those order dates, the analyst opens the Google Ads action breakdown using conversion-time reporting for June 1–7. The comparison record reads:
Purchase — Ads: ten conversions through the native action.
Purchase — GA4: ten conversions through the Analytics-sourced action.
Observed overlap: the implementation and matched order records establish that both actions represent the same ten purchases.
The actions total 20 credits, but their observed purchase set contains ten orders: 10 + 10 − 10 overlapping orders = 10. The other six completed store purchases sit outside that Google Ads-credited set. They remain valid orders; this exercise does not assign them to another advertising channel.
The action map also reveals that both actions were explicitly selected as primary purchase outcomes for the same campaign after an earlier setup change. There are two measurements of one commercial outcome in the bidding setup. The analyst records that duplication for correction instead of treating 20 as evidence of extra customer demand.
This conclusion depends on the evidenced overlap. Real reports can include modeled or fractional credit that cannot be matched one-for-one to visible orders. Without the simplified conditions and matching evidence used here, a difference between totals would be a reason to investigate, not proof of duplicate purchases.
Explain eight versus ten without changing the orders
There is still a date difference to resolve. Two of the ten customers clicked their Google ad on May 31 and purchased on June 2. The other eight clicked and purchased during June 1–7. In this simplified single-click example, each action therefore shows eight against June's click dates and ten against June's conversion dates.
Google's conversion-reporting guidance distinguishes standard click-time reporting from columns labeled “by conv. time.” Use conversion time for the order-date reconciliation. When assessing advertising spend, return to the appropriate interaction group so the spend and outcomes refer to the same set of clicks.
The analyst can now explain both differences: 20 credits versus ten observed purchases comes from overlapping actions; eight versus ten per action comes from the date basis. Neither explanation requires moving a sale or changing settings to force agreement.
For any remaining gap, compare event definitions, counting, credited channels, attribution model, window, time zone and processing status. Our guide to comparing attribution windows covers those broader boundaries. Correct a setting when it represents the business outcome incorrectly, not merely because it produces a different total.
Make one purchase outcome responsible for bidding
In this store, the native action has passed the order-level checks, so the owner approves it as the campaign's purchase outcome and retains the GA4-sourced action for comparison. That is an evidence-based choice for this setup, not a claim that native measurement is always better.
The advertising specialist checks every affected campaign, changes the duplicate action to secondary and verifies that no selected custom goal still uses it for bidding. Our guide to primary and secondary Google Ads conversions explains that campaign-level check. A changed label alone is not the acceptance test; the selected goals must reflect the approved choice.
The GA4 purchase key event stays in place. Removing the duplicate bidding objective does not require deleting useful site measurement. An All conversions total can still contain both actions, so it remains a view of action credits rather than a count to relabel “unique purchases.”
Close the review with a decision the owner can use
Save the action IDs, previous roles, approved choice and effective time with the original exports. The record should explain 16 all-site purchases, 20 action credits, ten observed Google Ads-credited orders and the eight-versus-ten date comparison. A later settings change should not be promised to rewrite that history.
At the next complete reporting review, check the chosen bidding outcome, the separate diagnostic action and the order register again. The owner keeps the spending limit unchanged until the corrected measurement supports a commercial decision. A lower conversion total after removing duplication does not establish that sales fell.
Our Google Ads management service can help turn conflicting exports into this kind of usable explanation. Bring the exact reports and action names: we can identify which differences are legitimate, which setup needs correction and which purchase outcome should guide the campaign.
Comments