
Conversion Tracking
Part of Social conversion tracking
Testing browser and server-side conversion events together
When the browser pixel and server integration report the same conversion, test them as a pair.
When the browser pixel and server integration report the same conversion, test them as a pair. The aim is one accurate business event in the platform's final reporting, with both transport paths working under the site's privacy rules.
Prepare a controlled test
Use a permitted test environment or a clearly marked internal transaction. Record its order or lead ID, event name, timestamp, value and currency. Confirm which component creates the shared event ID before you start.
Run one test conversion first, then check each route separately. For the browser route, observe the network call; for the server route, check the send and its response. A successful HTTP response from one route does not prove the final conversion was counted correctly.
For TikTok Pixel and Events API events that overlap, check that the event name and event_id match across both routes; a different ID makes them look like different actions. TikTok’s cross-route deduplication window is after 5 minutes and within 48 hours from the first event.
For Meta Pixel and Conversions API (CAPI), check that the browser and server events use the same event_id where your implementation sends one. A Test Event Code can be used for real-time verification in Meta Events Manager; pass the test only when both routes are received and final reporting represents the one test transaction.
For GA4, pair the browser Google Tag (gtag.js) with GA4 Measurement Protocol. Enable GA4 Debug Mode and use DebugView for live validation; for a purchase, compare transaction_id, value, currency and items with the recorded test. Pass when the event data matches and the final count reconciles to the one transaction.
If your implementation uses UniPixel, its live event-testing console displays events with full payload data, and its optional event logging stores sent events for debugging and audit. Use the payload and log to check that the server route sent the intended event.
Then repeat with the browser route unavailable, followed by a duplicate page load or server retry. Record what the platform shows, including any delay before the reporting view updates.
Platform-Specific Deduplication Windows and Requirements
- TikTok Pixel & Events API
- Deduplication window: 5 minutes to 48 hours after first event
- Meta Pixel & Conversions API
- Requires identical `event_id` for deduplication; use Test Event Code for real-time verification
- GA4 Measurement Protocol
- Validate via DebugView using `transaction_id`, `value`, `currency`, and `items` comparison
Test consent states too
Run an allowed and a declined state according to the site's actual consent policy. Check both the browser and server routes; a server event should not quietly restore information the user declined to share. Keep personal data out of screenshots and test logs wherever possible.
Close the test by reconciling the platform's final event count with the one underlying transaction and documenting any attribution difference. If the two routes are intentionally assigned different events, say so in the event plan; deduplication is needed only where they overlap.



