Since iOS 14.5, browser-based tracking alone leaks data badly. If you're still running on the Pixel by itself, Meta is making decisions on a fraction of your real conversions — and spending your budget accordingly.
Key takeaways
- Browser-only Pixel tracking loses events to ad blockers, privacy restrictions and cookie limits.
- The Conversions API sends the same events server-side, restoring signal the browser drops.
- Run both together with a shared event ID so conversions are deduplicated rather than double-counted.
- Verify in Events Manager that events are received, matched and deduplicated — do not assume the install worked.
Why the Pixel stopped being enough
The Pixel fires from the browser, which means it's blocked by tracking prevention, ad blockers, and opt-outs. Every conversion it misses is a conversion the algorithm never learns from. Less signal means worse optimisation, inflated cost-per-result, and reporting you can't trust.
The Conversions API (CAPI) fixes this by sending events server-to-server, independent of the browser. Run both together and deduplicate, and you recover a large share of the visibility iOS 14 took away.
Get the foundations right first
- Verify your domain in Business Manager — non-negotiable for conversion configuration.
- Set up Aggregated Event Measurement and prioritise your eight conversion events, with your primary purchase event at the top.
- Confirm the Pixel base code is on every page and your key events fire where they should.
Add CAPI without double-counting
The goal is one logical event reaching Meta through two paths, recognised as the same event. The two things that make that work:
- event_id — send a matching identifier from both the Pixel and the server so Meta can deduplicate. Get this wrong and you'll inflate your numbers.
- Rich customer-information parameters — hashed email, phone, name, and the fbp/fbc cookies. The more match keys you send, the higher your Event Match Quality, and the better Meta attributes conversions.
Better signal in means better delivery out. CAPI isn't a reporting nicety — it's optimisation fuel.
Verify, don't assume
Use the Test Events tool to watch events arrive in real time and confirm both the browser and server versions show up deduplicated. Then check Event Match Quality in Events Manager — anything in the "Good" to "Great" range means Meta has enough to attribute well. A "Poor" rating usually means you're not passing enough customer parameters.
What good tracking is worth
Recovering even 20–30% of lost conversion signal changes everything downstream: the algorithm optimises against reality, your reported ROAS lines up closer to your bank account, and you stop turning off campaigns that were actually working. It's unglamorous plumbing — and it's often the highest-leverage fix in the whole account.