How to trace ad attribution for one order, step by step

Pick one recent order and follow it through each tool. Check that the ad's mark reached your site and read the order's journey in Shopify. Then note the model and window in GA4, Google Ads and Meta. If each tool credits its own ad for the same order, you have watched attribution work, twice.

Step by step

1. Confirm that Google Ads tags your clicks. Auto-tagging adds a GCLID to the landing URL, so GA4 and Google Ads can match the visit to a later sale. The box "Tag the URL that people click through from my ad" should be ticked. Path: Google Ads > Admin > Account settings > Auto-tagging.
2. Check that the mark survives your redirects. Google's help suggests a test click on your ad, and says a site with redirects must pass the GCLID to the final landing page. If the parameter is gone when the page settles, sales from those clicks can lose their match. Path: your ad's final URL, opened in a browser.
3. Open the order's journey in Shopify. The conversion summary shows the customer's visits in the last 30 days and where their first visit came from. View conversion details adds the timeline between the first and the latest visit. Path: Orders > the order > Conversion summary > View conversion details.
4. Note GA4's rules for key events. Write down the reporting attribution model, the channels that can receive credit and the key event lookback window. These decide which touches GA4 may credit at all. Path: Admin > Data display > Events > Attribution settings.
5. Find a path like your order's in GA4. The paths report lists the channel sequences before key events, up to 20 touchpoints long. Filter by path length to find one that looks like your order, then export the table. Path: Advertising > Key events > Key event attribution paths > Share this report.
6. Compare GA4's models on the same data. Pick two models in the Attribution model (non-direct) columns and read how each channel's revenue shifts. A channel that gains under one model and loses under another is one the rules disagree on. Path: Advertising > Attribution > Attribution models.
7. Read Google Ads' own settings. Open your purchase action and note its attribution model and its click-through conversion window. Then compare last click with data-driven for your campaigns. Path: Goals > Conversions > Summary > your action > Edit settings; Goals > Attribution > Model comparison.
8. Read Meta's counting rules. Set the same ads side by side under 1-day click, 7-day click and 1-day view. In March 2026, Meta announced that click-through for website conversions would count link clicks only. Some accounts may still use the older rule while it rolls out. Path: Ads Manager > Columns: Performance > Compare attribution settings.
9. Set the claims against Shopify's orders. In a Shopify marketing report, switch the Attribution menu between Last non-direct click and Any click. Any click gives every clicked channel the full order, so the gap between the two totals shows how much the tools overlap. Path: Analytics > Reports > Category > Marketing > Performance by referring channel.

A worked example

For illustration, take one invented order with four touches. On day 1 the shopper clicks a Meta ad. On day 3 they click a Google Shopping ad. On day 8 they scroll past another Meta ad without clicking. On day 9 they click a tagged email and buy.

Suppose Meta is set to 7-day click and 1-day view. It cannot count the first click, which came more than a week before the purchase. It counts the day 8 impression instead, because the purchase followed within a day. Google Ads counts its day 3 click, well inside its default 30-day click window. Its model makes no difference here: with one Google click on the path, last click and data-driven both give it the whole conversion. That holds when the purchase action uses the Google tag. If the action is imported from GA4, Google's help says only key events where Google Ads is the last non-direct click become Google Ads conversions. The email came last, so this order would not count there.

GA4, on paid and organic last click, gives the whole sale to the email, the last channel clicked. On data-driven, it may split the credit among the clicks it recorded. Shopify's default for marketing data, last non-direct click, also picks the email.

So the tally for one order: Meta claims it, Google Ads claims it, and GA4 and Shopify both give it to email. Add Meta's purchases, Google's conversions and your email sales, and this order counts three times. Shopify's Any click model shows the same overlap inside one report: Meta, Google and email each get the full order.

In one store's export, most revenue came from journeys far simpler than this one. Journeys with 1 touch hold 79.5% of revenue on the Journeys sheet, with 0.5 days to buy. On the Channels sheet, Direct holds 57.7% of revenue in all three views. For those Direct sales, GA4 found no ad mark at all. If a platform still claims some of them, it used a view or a click that GA4 never saw.

What should you check when the trail goes cold?

• No gclid in the landing URL. Auto-tagging is off, or a redirect drops the parameter. Fix that first, because nothing further down can match a visit without its mark.
• No conversion summary on the order. Shopify says apps and private browsers can block cookie tracking, and orders that began as drafts can have no summary. Details can also take up to 48 hours to appear.
• Google Ads and GA4 disagree about a day. Google Ads books conversions on the click date, while GA4's attribution reports use the time of the key event by default. Add the by conv. time columns in Google Ads before you compare.
• Meta claims a purchase that GA4 gives to Direct. Meta can credit an ad the person saw but never clicked, and it reports across devices. Meta's help on reporting differences lists both reasons.
• Shopify names a different first visit than GA4. Shopify resets a visitor's first interaction after 30 days without a purchase, and it counts sessions with its own cookies. Compare each tool with itself before you compare the two.
• The numbers jumped after a settings change. A new window in Google Ads only applies to conversions from that day on. Do not compare across the change without noting it.

What to do this week

1. Trace three recent orders end to end. Run steps 3 to 9 for each one. Pass: for every order you can name the mark, the window and each tool that claimed it. Fail: an order has no mark or no summary, so note which and fix the tagging behind it.
2. Run Meta's comparison on 1-day click alone. In Compare attribution settings, set 1-day click beside your current setting for last month. Pass: most purchases remain, so link clicks carry Meta's count. Fail: most vanish, so the count leans on views and engagement that GA4 cannot see.
3. Put Any click next to Last non-direct click in Shopify. Use last month in Performance by referring channel. Pass: the two totals sit close, so few orders mix channels. Fail: Any click runs far above, so each platform's own number includes orders the others also claim.

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 auto-tagging (Google); Viewing order conversion summary (Shopify); Select attribution settings (Google); Key events attribution paths report (Google); Key event attribution models report (Google); About conversion windows (Google); About attribution models (Google); Understand your conversion tracking data (Google); Compare attribution settings in Meta Ads Manager (Meta); About attribution models and attribution settings (Meta); Simplifying Ad Measurement for a Social-First World (Meta); About conversion count differences between Meta Ads Reporting and third-party reporting tools (Meta); Marketing reports (Shopify)

Related answers

• All Attribution models questions
• How does ad attribution work?
• What is the default Meta attribution setting?
• Causal attribution, explained