Why Your Pixel Isn't Tracking (Dedup, iOS, In-App Browsers)

Start with an observable test. Open the same published URL the ad will use, click a DSP button, and inspect the event. A page view is not enough because the campaign action happens on the click-out. If the button sends nothing, fix that wiring before making a claim about iOS or attribution.

In VLVTN, confirm the saved integration in Dashboard > Integrations, use Send Test Events, and select that Pixel in the smartlink editor's Settings tab. The connection test and the published button test answer different questions. Run both before changing the campaign.

Apple's App Tracking Transparency framework limits cross-app tracking when a user does not grant permission. Consent choices, browser privacy controls, and Meta's own measurement rules can therefore make browser-side reporting less complete than a simple count of visits and button actions.

Do not assign a fixed loss percentage to your account. Compare the browser event, server event, and deduplicated result in Meta's diagnostics. The server copy provides another delivery path for the click-out, while the shared event ID prevents the pair from becoming two reported conversions.

EasyList, the most-used filter list behind ad blockers, targets tracking pixels from Meta, Google, TikTok, and others. When a visitor runs a blocker, the fbq() request is blocked and the event simply never fires. There is no error you will catch, no partial signal. The conversion happens and the browser pixel says nothing.

You cannot control a visitor's blocker. Send the matching event through CAPI so the campaign does not depend entirely on the browser request. That improves event delivery, but it does not reveal a stream or any other action inside the DSP.

Most of your paid music traffic comes from a Reel or a TikTok, which means the link opens inside an in-app browser: the webview baked into Instagram, TikTok, or Snapchat. Those webviews suppress third-party cookies and device identifiers, so the signals the browser pixel relies on are not there. On top of that, Safari's Intelligent Tracking Prevention caps JavaScript-set cookies like _fbp at around seven days.

Test the Instagram and Facebook in-app browsers directly because that is where much of the paid traffic opens. Confirm that the page loads, the DSP redirect works, and the browser/server event pair is deduplicated. Do not infer a stream from a successful redirect.

This one is the opposite problem. Once you start sending events from both the browser pixel and the server, you can accidentally count the same conversion twice. Meta only knows two events are the same when it receives a matching event_name plus event_id within a short window. When it sees that match, it dedupes and counts one. Without a shared, identically-formatted ID, it counts both.

Wiring it up means passing event_id as the fourth argument to the pixel call, fbq('track', 'EVENTNAME', {}, { eventID: '...' }), and sending that exact same event_id in the server payload. Get this wrong and your reported numbers wobble: some conversions doubled, some mismatched, none of it trustworthy enough to optimize against.

Every reason above points at the same answer. The Conversions API (CAPI) sends your events from your server directly to Meta, server-to-server, with no browser involvement and no cookie dependency. That means iOS opt-outs, ad blockers, and in-app browsers do not block it. Meta and TikTok both recommend running the browser pixel and the server API together for the most complete coverage, then deduping the pair on event_id so each real conversion counts once.

VLVTN sends the DSP click-out through the browser Pixel and CAPI with one shared UUID. Meta can then match the event name and event ID and count one conversion. This makes the measurement less dependent on the browser path. The amount of additional signal is account-specific, and the event still describes only the click-out.

For the wiring itself, the sibling guide on tracking Spotify conversions walks through the setup, and the smart link pillar covers where this sits in the wider funnel. One honest reminder before you start reading the numbers: a conversion is a tracked click-out to a DSP, not a guaranteed stream. Fixing the tracking makes the count trustworthy. What happens after the click-out is the song's job.

Use the focused zero-conversion diagnostic when Ads Manager shows nothing, then run the pre-spend Pixel and CAPI check. If you need the tracked landing page itself, you can create a VLVTN smartlink.