Mission context
The current server-side tracking system is not working properly and is not optimized. As online acquisition is a real challenge for this customer, it is impossible to continue under these conditions, as the data is not flowing properly either in Matomo or in the various advertising networks. The aim of this mission is therefore to correct the server-side tracking that has been put in place.
Tools involved in data collection
Shopify
GTM Web
Axeptio
GTM Server-Side
Addingwell
Connected platforms
Meta Ads
Pinterest Ads
Tiktok Ads
Snapchat Ads
Google Ads
Matomo Analytics
Data flow diagram
flowchart LR
subgraph Shopify store
W[Addingwell \ndataLayer\n app]
Y[GTM Web]
end
subgraph Addingwell
Z[GA4 Client]
A[GTM Server Side]
B[Google Ads]
C[Meta Ads]
D[Pinterest Ads]
E[Tiktok Ads]
F[Snapchat Ads]
G[Matomo]
end
W-->Y
Y-->Z
Z-->A
A-->B
A-->C
A-->D
A-->E
A-->F
A-->G
The Shopify challenge
On a Shopify (or Shopify+) store, the checkout and thank-you pages are in a sandbox. This is a specific environment that doesn’t allow cookies to be sent correctly. If nothing is done to remedy this, there will be many problems with data quality in all our tools.
flowchart LR
subgraph Shopify store
subgraph Sandboxed JavaScript environment
D[Checkout page]
E[Thank you page]
F[Order status page]
D-->E
E-->F
end
end
Addingwell’s Cookie Restore feature restores cookies even in the sandbox.
Results
Meta
Improved number of events
Differences between browser and server requests over the last 28 days.
+9% on purchase events.
+4.7% on payment initiation events (begin_checkout)
+9.8% on page view events (page_view)
Purchase event quality score
=> 7.9/10
This quality score largely represents the user data (email, phone number, etc.) sent to Meta. A good score is between 7 and 10 for the purchase event.
Catalog match rate
=> 100%
Matching products with the Meta catalog ensures optimal distribution of Meta Advantage+ ads.
Google Ads
Enhanced conversions
Merchant Center match rates
Snapchat
Improved number of events
Differences between browser and server requests over the last 28 days.
+4.9% on purchase events.
+5.8% on payment initiation events (begin_checkout)
+7% on Page View events (page_view)
Purchase event quality score
Tiktok
Improved number of events
Differences between browser and server requests over the last 28 days.
- +2.6% on purchase events
- +2% on begin_checkout events
- +4.8% on page view events (page_view)
Purchase event quality score
=> 73
Difference between browser vs. server requests over the last 28 days.
+9.5% on purchase events
+7.3% on page view events (page_view)
Addingwell
44% of returning visitors returned after 8 days on Safari (ITP bypass is active)
On a client-side setup, first-party cookies dropped in Safari are limited to 1 day.
On a classic Server-Side setup, first-party cookies are limited to 7 days.
Matomo Analytics
E-commerce reports correctly populated
Abandoned cart feature enabled
Taking a step back
This case study shows you the positive impacts of server-side tracking both in terms of the volume of events obtained in addition to data quality with user data (enhanced conversions, event quality score) and the extension of cookie lifespan (safari ITP bypass).
In this case study, I’m not talking about the overall impact on sales or cost of acquisition, as I feel that tracking (whether client-side or server-side) is only one piece of the puzzle. What’s more, it’s difficult to establish a precise causality between tracking and business performance. As I often say to my customers:
Server-side tracking is the solution that will provide you with the optimum technical conditions to drive advertising performance.