In the previous two posts on purchase events, we focussed on building data models and gaining insights from the GA4 events. But purchase is the only event with a clear equivalent in shopify (orders) and a join key (transaction_id.)

This means we can join the data from the two sources! This is how i have answered clients who wonder about data accuracy, trusting GA4 data, or trusting data models!

But there’s more to this than shutting up doubters. Let’s explore.

Getting data from shopify orders into GA4

Shopify orders table has more details about the purchase than the GA4 event. In fact, it has every detail about an order.

GA4 purchase event covers 80-95% of shopify orders depending on the users privacy settings and GTM setup. All of these can be mapped to shopify orders. Once we map the events, we can take all the details from shopify (like customers email id, first time customer or returning customer flag, payment method, tax, shipping, coupon, customer geography, and so much more) and put them in our data models.

Returning customer flag.

GA4 is not good at tracking returning users. A user who made first purchase through marketing can come back directly and GA4 will most likely tag it as a new user.
The same person can come from desktop and get tagged as new user.
The same person can come from from an ad on different marketing platform, and get tagged as new user.

But if a user makes a purchase, we get a purchase event for them in shopify. And here, we can find out if it’s a first time purchaser.

We can then pass this info back to GA4 and improve our returning users data. This can seem like futile effort – because the number of visitors is so much higher than buyers. BUT we are much more interested in buyers than other visitors.

Correcting buyer data is a huge advantage to further analysis.

Payment method and other details

Shopify orders table also adds more details to the payment information. The extent of new information depends on what you are getting in GA4 by default

Most important of these is refunds and cancellations. These are not captured in GA4. By adding this extra information, you can get an exact idea of revenue made. You can also account for losses through items shipped but returned, COD orders not completed and so on.

For some stores, this can significantly bring down the ROAS on marketing – as well as challenge existing ideas about best performing campaigns.

Complete picture with shopify data

While GA4 data is good for analysis, it is incomplete. Shopify data solves this problem. It gives you all the orders and revenue, with refunds included.

Shopify also gives you the referral site. This can be different from GA4 for many orders. This is expected. Shopify is better at tracking where a user came from than GA4. So take shopify’s data as the better of the two.

With this, you can assign the orders against your marketing campaigns on meta and google ads – and get ROAS at a campaign level, everyday.

Line items

Some of the analysis mentioned in part 2 can also be achieved through shopify orders table. This is true for analysis related to items in the order.

Data models

shopify_orders

Matches against GA4, adds more details by cleaning up shopify referral , landing page data. Also takes refunds and cancellations into account.

shopify_line_items

Breaks the above into line items.