How to check Google Ads attribution, step by step

To check Google Ads attribution, open each purchase conversion action and read three settings: the attribution model, the conversion windows and the count. Then compare models in the attribution reports, and line up click-date and conversion-date columns. If those settings were chosen on purpose, most gaps with GA4 and Shopify usually have a plain explanation.

Google Ads keeps these rules on each conversion action, not on the campaign. Two actions in one account can follow different rules, so check every action that counts a purchase. Set aside time for one coffee, and keep GA4 and your Shopify admin open for steps 8 and 9.

Step by step

1. Open the purchase conversion action. Steps 2 to 4 all happen on this one settings screen. Menu path: Goals icon > Conversions > Summary > the conversion action > Edit settings.
2. Read the attribution model. Data-driven is the default for most conversion actions, and last click is the only other model Google still supports. A change only applies to conversions counted from then on. Menu path: Edit settings > Attribution model drop-down > Save > Done.
3. Read the three windows. Click-through defaults to 30 days, engaged-view to 3 days and view-through to 1 day. Clicks can run from 1 to 90 days depending on the source, and Google recommends at least 7. Menu path: Edit settings > Click-through conversion window, then Engaged-view conversion window, then View-through conversion window.
4. Read the count. Every counts each purchase after an interaction, while One counts one conversion per ad click, which suits leads. Every is the default for website actions. Menu path: Edit settings > Count > One or Every > Save > Done.
5. Look for a second purchase action. The Conversions column only includes primary actions, so two primary purchase actions can count one sale twice. Google's advice is to deduplicate sales with transaction IDs. Menu path: Goals icon > Conversions > Summary, then read down the list of actions.
6. Open the attribution reports. Google keeps four views here: Conversion paths, Path metrics, Assisted conversions and Model comparison. Path metrics shows the days and interactions before a conversion, measured from the first or the last ad interaction. Menu path: Goals > Attribution > Path metrics tab > Measure from drop-down.
7. Compare the two models. Model comparison sets last click and data-driven side by side, with Cost / conv. and Conv. value / cost for each. Menu path: Goals > Attribution > Model comparison > Compare and With drop-downs.
8. Add conversion-time columns before you compare tools. The main columns report on the click date, and the by conv. time columns report on the purchase date. Google suggests the conversion-time view when you compare with other platforms. Menu path: Columns menu > Conversions (by conv. time).
9. Check GA4's side of the link. GA4 lists linked Google Ads accounts under Product links. Its attribution settings decide which channels can earn credit for conversions shared with Google Ads: Google paid channels by default, or paid and organic. Menu paths: GA4 Admin > Product links > Google Ads links, and GA4 Admin > Data display > Events > Key events attribution.

A worked example

Start with the date rule, because it fools the most people. For illustration, say a shopper clicks your Shopping ad on 2 May and buys on 12 May. In the Conversions column, that sale sits on 2 May, the click date. In this worked example, Conversions (by conv. time) puts it on 12 May instead. If you run a report for 1 to 10 May, one column shows the sale and the other does not. Neither column is wrong. They file the same sale on different days.

Now hold the windows against real journey lengths from one store. Store A is one store's anonymised GA4 export, 1 January 2024 to 21 August 2026, holding shares of revenue only. On the Journeys sheet, journeys of two or three touches take 12.5 days to buy, and journeys of ten or more take 16.0 days. Google's default 30-day click window covers both with room to spare.

Google's own example is blunt: with a 7-day window, any conversion more than 7 days after the ad interaction is not recorded. So if this store had cut its window to 7 days, a click that opened one of those longer journeys could fall outside it. The 1-touch journeys would barely notice: on the Journeys sheet they hold 79.5% of revenue, at 0.5 days to buy. In this one store, the window choice can only matter for about a fifth of revenue.

That is the use of step 6 on your own account. Path metrics, measured from the first interaction, tells you whether your window cuts anything off. One store's export cannot tell you that about yours.

Step 7 deserves its own picture. For illustration, take a path with a YouTube ad click, then a Shopping click, then a brand search click. Last click hands the whole sale to the brand search ad. If data-driven finds that paths with the YouTube click convert more often, it moves part of the credit back to YouTube. Google's help page tells the same story with a bike tour: the ad clicked first gains credit.

What to check when the report looks wrong

• Conversions dropped right after a model change. The new model only applies going forward, so the weeks before and after follow different rules. Add the current model columns to restate the past under the new model.
• Data-driven and last click show the same numbers. Google says that can happen, depending on data availability. If most of your paths hold one Google click, there is nothing to move.
• Conversions show decimals. That is data-driven attribution splitting one sale across several clicks. The fractions add back up to whole sales, so nothing is broken.
• Google Ads shows more purchases than GA4. Check dates first: the click date against the purchase date. Then check whether view-through sits in the column you compare, and whether GA4 credits paid and organic channels while Google Ads credits Google only.
• Shopify gives Google less than Google takes. Shopify's marketing reports credit the most recent click within 30 days. An email and a Shopping ad clicked by the same buyer can each record the sale in their own reports.
• Attribution reports disagree with the Campaigns page. Google says the two count networks, time of event and conversion sources differently. Use the Campaigns page for performance and the attribution reports for paths.

What to do this week

1. Screenshot the settings of every purchase action. Pass: one primary purchase action, with a model, windows and count you can explain to your accountant. Fail: two primary purchase actions, or defaults nobody chose on purpose.
2. Run Model comparison for last month. Pass: the campaigns that move under data-driven are the ones you expected to move. Fail: big swings on campaigns you are scaling, so your budget follows a model choice, not a result.
3. Match one week of Conversions (by conv. time) to Shopify. Pass: the two land close for orders from Google-tagged visits. Fail: a gap with no timing reason, so check tags and transaction IDs next.

Check the homework. Your GA4 Attribution paths export already holds the evidence. Causality Engine reads that one file and shows what each channel caused next to what last-click gave it, in 1 to 2 minutes, for €99 once (excluding VAT), refundable within 30 days. Check the homework

Sources, 1 October 2026: About conversion windows (Google Ads Help); About attribution models (Google Ads Help). About conversion counting options (Google Ads Help); Understand your conversion tracking data (Google Ads Help). About attribution reports (Google Ads Help); About data-driven attribution (Google Ads Help). Connect Google Ads to Google Analytics (Google Analytics Help); Select attribution settings (Google Analytics Help). Change the reporting attribution model for key events (Google Analytics Help); Measuring marketing performance (Shopify Help Center).

Related answers

• All Google Ads questions
• How does Google Ads attribution work?
• How does Google Ads attribution work for electronics?
• Attribution window, defined
• The quarterly attribution checkup