How to Set Up the Snapchat Conversions API
Snapchat traffic is heavily mobile and heavily blocked. Here is how to recover the events the pixel never sees.
Snapchat traffic is almost entirely mobile, almost entirely in app, and lands on your site through a webview with aggressive privacy defaults.
If any channel in your mix is losing browser events, it is this one. The Conversions API is not an optimisation here, it is closer to a requirement.
The Snap Pixel first
Install the Snap Pixel through GTM using the community template, with your pixel ID.
Then fire Snapchat’s standard events: PAGE_VIEW, VIEW_CONTENT, ADD_CART, START_CHECKOUT, PURCHASE, SIGN_UP, SUBSCRIBE.
Snapchat’s event names are uppercase with underscores, which differs from every other platform and is a common source of events that silently do not register.
Why the browser loses so much here
The Snapchat in app browser is a webview. Storage behaves differently, cookies are more restricted, and users frequently bounce back to the app mid journey.
Add to that an audience skewed young and mobile, which is the demographic most likely to be on a device with tracking prevention enabled by default.
The result is that Snap Pixel only setups routinely report a fraction of the conversions the channel actually produced, and Snapchat gets cut from budgets on the strength of that.
Setting up the Conversions API
In Snapchat Ads Manager, go to Events Manager and select your pixel.
Open the Conversions API setup and generate a long lived access token.
Send events to the Snapchat conversions endpoint with the pixel ID, event name, timestamp and hashed user data.
Verify in Events Manager that server events are being received.
Snapchat also has a server side GTM template, which is the path of least resistance if you already run a container: feeding every ad platform from one server container.
Matching data
sc_click_id. Snapchat’s click identifier, on your landing page URL. Capture it. This is the strongest signal by a wide margin.
Hashed email. SHA256, lowercase, trimmed.
Hashed phone. E.164 format before hashing.
Mobile advertiser ID. If you have an app and can legitimately collect it.
IP address and user agent. The visitor’s, passed through.
Capturing the click ID matters more on Snapchat than almost anywhere else, precisely because the cookie based identifiers are so unreliable in the in app browser.
Deduplication
Send a client_dedup_id with both the pixel event and the server event. Same value on both sides.
Snapchat deduplicates on that identifier within a window. Your order ID works well.
Verify after launch by placing a test order and confirming a single purchase in Events Manager. If you see two, the IDs are not matching.
Event timing
Send events as close to real time as you can. Snapchat’s attribution windows are shorter than some platforms, and a batched nightly upload can push events outside the window entirely.
If you can only batch, batch hourly rather than daily. Daily batching on Snapchat loses real attribution.
Reading the diagnostics
Events Manager shows you received events, deduplication rate and match quality.
Three numbers to watch:
Deduplication rate. If very few events are being deduplicated, your IDs do not match and you are double counting.
Match quality. Low means your identifiers are not landing. Check click ID capture first.
Server versus browser event counts. Server should be higher. If server is much lower, your server side implementation is missing events.
Landing page considerations
Since nearly all your traffic arrives in a webview on mobile data, the landing page itself matters more than usual.
A page that takes six seconds to load will lose people before any tag fires. A consent banner that blocks the entire page on a small screen will lose more. These are tracking problems as much as conversion problems, because an event that never fires is an event you cannot optimise toward.
Test your landing page on an actual phone on actual mobile data, not on desktop with the responsive view. It is a different experience.
Setting expectations
After enabling the API, expect reported conversions to increase noticeably, often more than on other platforms, because the baseline was so incomplete.
Then give the campaigns two weeks to relearn before judging performance, and compare against your actual order count rather than against the previous Snapchat numbers.
Is Snapchat worth this effort?
Honest answer: only if it is a real line in your budget. If you spend a small amount there as an experiment, the Snap Pixel alone is fine and the numbers will be pessimistic.
If it is a meaningful channel, the API is the difference between measuring it and guessing at it, and the effort is a day at most.
Need this built properly?
The channels with the worst browser conditions gain the most from server side events. I set up and repair tracking for ecommerce stores and lead generation businesses, so the numbers in your ad accounts match the money in your bank account.
Have a look at what I do, or book a call and we can go through your setup together.
