Architecture
Server-side tracking is not ‘more tags.’ It is checkout truth.
When server-side tracking is mandatory, when the Pixel still matters, and why a webhook with no email is worse than an honest Pixel.
People say “server-side” as if moving a tag to a container fixes attribution. It does not. The question is: does the conversion still exist when Safari lies?
Fire the sale when money changes hands
Checkout complete, paid invoice, trial start, qualified lead. Those events must leave your backend. Thank-you-page JavaScript is optional garnish.
Do not invent a server PageView for every route unless you have a reason. Noise trains Advantage+ on tourism.
Pixel is still how Meta recognizes the visit
| Browser Pixel | Server-side tracking |
|---|---|
| Identifies the visit | Survives blockers |
| Dies on iOS / Safari | Needs email you already collected |
| Easy to install twice | Easy to send without email |
Best path: Pixel on the landing page, server sale on the order, same checkout ID. Caply does that pair. A tagging server can too, if you like operating a container.
The failure mode nobody blogs
A Stripe webhook purchase with no email, no phone, no visit. Delivery is “green.” Match quality is a coin flip. Match Forecast marks that event unmatched and sums purchases Meta may miss from order value.
If you only have an order id, send it as customer ID and still attach the email you showed on the receipt.
Developers: cookie and field names live in /docs.
Caply
Match Forecast on every purchase. Tracking radar on competitor storefronts. Managed server-side tracking without a tagging server.
Start 14-day trial