Conversion Tracking: The Setup Most Ad Accounts Get Wrong
Roughly a third of the ad accounts I audit have broken conversion tracking. In those accounts every downstream number is wrong: the algorithm is optimising toward noise, ROAS is fiction, and the reporting the owner is making decisions from does not describe their business.
This post is the setup I would do. It takes about two hours and it is the foundation everything else sits on.

Decide your conversion event before you install anything
The most common mistake is installing a pixel with every possible event enabled and letting the algorithm pick. Then it optimises for PageView because it happens most often, and you wonder why you get cheap clicks and no customers.
Decide first: what is the single event that means money? For most advertisers it is Purchase (ecommerce), Lead or Submit Form (services), or a qualified booked call.
Everything else is secondary. You track it for reporting; you do not optimise toward it.
Step 1: install the base code properly
Facebook base code goes in the <head> of every page, on your main domain, loaded synchronously. Not in a footer, not conditionally on non-landing pages, not only on the thank-you page.
Verify it: Meta Pixel Helper extension, or Events Manager showing live traffic. If Events Manager shows no activity while your site has visitors, the pixel is not firing and nothing else matters yet.
Step 2: set up the Conversions API
Browser-only tracking loses 20–40% of conversions to ad blockers, ITP, and platform restrictions. The Conversions API sends the same events server-side, so you recover most of that.
- Server events — Meta's own integration for Shopify, WooCommerce, BigCommerce. Usually the right choice for ecommerce.
- Conversions API Gateway — the hosted option, handles deduplication for you.
- Custom implementation — for anything bespoke. Most work, most control.
Step 3: deduplicate properly
This is where most implementations go wrong, and it is the hardest thing to get right.
When the Pixel and the API both send the same purchase, Meta will count it twice — which makes your ROAS look twice as good as it is and teaches the algorithm nothing useful. Deduplication requires an event ID: a single identifier for that specific transaction, sent identically by both the browser and the server.
The test: in Events Manager, a single purchase should show as roughly 1 event with both Browser and Server sources, not 2 separate events. If you see two, deduplication is not working and your numbers are inflated.
Step 4: test the whole flow before you trust it
Do this end to end, once, with a real low-value transaction if you can spare it:
- Events Manager → Test Events, open in one window
- In another, complete a real form or test checkout on your live site
- The event should appear within 30 seconds, with the correct value and currency
- Check the source shows both Browser and Server
Do this monthly afterwards. Tracking breaks silently — a site update, a plugin conflict, a consent change — and you will not notice until the numbers look wrong.
Step 5: send values and currencies correctly
Value-based optimisation needs real values. Sending the same static number, or sending a value without a currency code, makes value optimisation useless and can actively mislead the algorithm.
- Send the actual transaction value, not a guess
- Always send
currencyin ISO 4217 format (USD,GBP) - For leads, send an estimated value or drop value-based optimisation entirely
If you take one thing from this: an account optimising for value with no values is strictly worse than one optimising for conversions with no values.
Common failure signatures
| Symptom | Almost always |
|---|---|
| Events Manager shows no activity | Base code not firing, or consent blocking it |
| Clicks high, purchases zero | Purchase event not firing, or firing on the wrong URL |
| ROAS far higher than reality | Deduplication failing — double counting |
| Delivery erratic, CPA swings weekly | Low event volume, or two conflicting optimisation events |
| Works in Ads Manager, nothing in Events Manager | Debug mode filter still enabled, or wrong pixel ID |
Not sure whether your tracking is actually working?
Points 17 and 18 of the 27-point audit are exactly this, verified against Events Manager rather than assumed. Free version scores it in five minutes.
Run the Free Leak CheckThe bottom line
Decide your money event, install properly, add the API, deduplicate with event IDs, and verify end to end before you spend another dollar. Everything else in your account depends on this being right, and it is the one thing that fails silently.
Running Facebook or Instagram ads? The Meta Conversions API guide covers server-side tracking and deduplication step by step.
Frequently asked questions
Do I need the Conversions API if I already have the Pixel?
How do I know if my tracking is double counting?
Should I optimise for leads or purchases?
Free Leak Check
What’s your ad account’s leak score?
27 quick yes/no questions, the same checks I run in every paid audit. About 3 minutes, and you see your score and biggest leak right away. No email needed.
Prefer email? Get the checklist instead
Want us to find the leaks for you?
Professional audit, 48-hour delivery, money-back guarantee — from $97.
Get My Ads Audit →