A number of Littledata customers have noticed a deterioration in the percent of orders being attributed to marketing campaigns in Google Analytics 4 (GA4) recently.
Without good marketing attribution it’s hard to be confident in marketing or product decisions. And it’s true that the rate of attribution – or visibility of marketing campaigns – has gotten worse.
But it’s not the fault of GA4 directly.
Switching tracking from UA to GA4 has caused marketers to look more closely at attribution rates, and spot issues which have been growing for some time. GA4’s attribution model is actually very similar to Universal Analytics previously, but there are some additional settings and other things to pay attention to. We focus on the best of both first-party cookies and server-side tracking at Littledata, to support the best possible data capture – whether you want to look at that data in GA4 or on top of BigQuery.
Here are the real reasons attribution is getting worse. The good news is that as a complete data solution that goes beyond simple attribution, Littledata‘s app fixes many of them out of the box for Shopify and Shopify Plus.
1. You’re relying on Shopify Analytics
Shopify’s built-in analytics is useful for a quick overview of purchases and products sold, but it is very simplistic when it comes to marketing attribution.
Shopify announced improved attribution reports in summer 2023, but the root problem seems to be that the data capture of UTMs and so marketing campaigns is missing.
I don’t expect Shopify to catch up with Google Analytics any time soon, so please give GA4 another try! GA4 is also the best way to share audiences and retargeting data with Google Ads. That’s why brands that implement our server-side tracking for Shopify and GA4 see an average 20% decrease in CAC in Google Ads, while at the same time improving both reach and LTV.
2. You’re looking at the Traffic acquisition report in GA4
GA4 has two different standard attribution reports: traffic acquisition (reporting on session channel) and user acquisition (reporting on first user channel).
The traffic acquisition report will only show you the channel that brought that session to purchase. And if the user has a period of activity of more than 30 minutes between buying and completing the checkout, the purchase will be in a new session and the channel of that session will be Direct.
By contrast, the User acquisition report looks at what was the first known source that brought the user that purchased. This should always show less Direct attribution.
Many internet users don’t store cookies for more than a week (see #6 below) so your reporting is unlikely to be polluted by last month’s campaigns.
If you do want to report strictly on session attribution, then you could extend the duration of a session from 30 minutes to 4 hours to give a user more time to complete a purchase.
If you’re interested to see what’s possible with GA4 reporting, book a demo of Littledata Plus.
3. Landing page views are not tracked
Marketing attribution in GA4 is based on tracking UTM parameters (UTM source, medium and campaign) from the landing page URL.
If the landing page is not tracked, then these UTM parameters are never captured, and the session is Direct.
The landing page might get missed because the user is quickly redirected to another page, or other scripts on the landing page delay the Google Analytics tracking.
Littledata fixes this by picking up UTMs as soon as possible from the landing page and tracking them on a subsequent page if necessary. This includes the ability to track both paid and organic channels (like Klaviyo and Attentive) via UTMs, linking them with browsing behavior and checkout activity.
4. Your Thank You pages are not being tracked
Similar to landing pages in reason #3, the thank you page might not get tracked. This is a growing problem as users familiar with ecommerce know they can rely on an email or other notification to confirm an order and have no reason to wait for the thank you page to load.
Users may be browsing with an ad-blocker enabled (generally used by at least 40% of the global population). This will block both landing pages and purchases – a double hit on attribution.
This is one of the core problems Littledata solves. By tracking purchases server-side we can ensure 100% of conversions are tracked in Google Analytics, and as many as possible are attributed to campaigns.
5. Users have opted out of tracking
GDPR regulations in Europe, and increasingly in states across the US with regulations like CCPA, give users a legal right to opt of being tracked by first-party cookies.
This inevitably reduces the proportion of revenue that can be attributed, but if this cookie consent is implemented incorrectly, it can further erode tracking – for example, by not tracking the landing page.
Compliance with cookie consent and accurate tracking can go hand-in-hand, and it’s essential to get this right now instead of waiting for a future project to pay attention to Shopify’s Customer Privacy API.
6. Apple is wiping cookies and blocking tracking
Apple has a long-stated intent to limit tracking of users on Apple’s products. This includes wiping cookies on iOS web browsers (mainly Safari) after 7 days and, with the latest iOS 17 roll out, blocking tracking in private browsing mode.
Safari’s share of the global browser market is only 30%, so the detrimental impact on overall attribution is probably less than 10% – but it has grown every year.
7. Your checkout flow includes a redirect
Most credit cards now need a confirmation from the bank for larger purchases. Usually this ‘3D secure’ process is handled within the frame of the Shopify checkout, but sometimes it involves a redirect to another domain.
When the user then returns from a different domain to the thank you page, the attribution of the session is to a referral from ‘paypal.com’ or similar.
The solution is to add these payment domains to the referral exclusion list in GA4, and ensure they do not start a new session.
8. Reporting identity is set to Blended
Hidden in the settings of GA4 is how the different ways of identifying a user are used together.
Along with cookie-based device identifiers, the default Blended setting uses Google Signals – or the Google account system across Gmail, Android, YouTube etc – to link users across devices. Unfortunately, Google Signals doesn’t seem to work well in GA4 yet, and most brands see better attribution when switching off the use of Google Signals.
You can change this (along with a general review of GA4 settings) by switching to Device-ID Only identity. This setting works instantly to recalculate attribution based on just the device ID.
9. Rise of purchases on Shop Pay
Shopify has been heavily promoting Shop Pay for convenience for customers and pushing accelerated checkouts via Shop Pay.
Many stores are seeing an increase in orders from the Shop App, where customers can browse products on hundreds of thousands of brands – and may in time compete with Amazon for product range.
Luckily Shopify just changed the behavior, so Shop Pay now redirects to the thank you page on your domain. This will improve attribution of the orders, but only if the thank you page is tracked.
10. The user completed the purchase on another device
The average number of devices per customer has increased steadily over the last decade, so the chances that a user browses your site on one device and buys on another is higher than ever.
This is not a problem that cookie-based tracking can solve. It should be something that Google Signals helps with, but as discussed in point #8 Google doesn’t have this working yet.
However, it is a problem that can be solved in Facebook Ads using Meta Conversions API. By sharing the customer profile with Meta (securely, directly from Shopify to Meta’s servers), the purchase can often be linked to a Meta account.
This means, if the user is logged in on Instagram on their phone and also on Facebook Messenger in their laptop browser, the purchase on the laptop can be linked to the ad viewed on the phone.
Setting up Conversions API (CAPI) via Littledata takes just a few minutes and is essential to getting accurate targeting in Facebook Ads.
You don’t have to accept poor marketing attribution on your Shopify store.
There’s no one thing to fix it, but with less than an hour of your precious time many of the problems above can be addressed.
The barriers to great marketing attribution are increasing all the time, but this is what I obsess about at Littledata – and you can be sure we have your back when it comes to accurate marketing data.