"We tried Facebook ads and they didn't work." I hear this often, and when I look at the account, the ads were usually running on missing or broken tracking. The campaigns never had a fair chance. This post explains what actually happens inside Meta when tracking is missing or partial, so you can see why I fix tracking before touching a campaign.
How does Meta use your conversion data?
Meta's delivery system decides who sees your ads by predicting who is likely to take the action you optimise for. It makes those predictions from the conversions it has seen: who bought, what they looked like, where they came from. Your conversion events are the training data.
So the quality of your ads' targeting can never be better than the quality of the events you send.
What breaks when tracking is missing or wrong?
1. Meta optimises for the wrong thing
Without a Purchase or Lead event, you can only optimise for clicks, landing page views or reach. Meta then finds people who like clicking, not people who like buying. Costs per click may look good while sales stay flat.
2. Meta learns from a distorted sample
Partial tracking is the hidden version of the same problem. If the browser Pixel misses a share of purchases (ad blockers, Safari, the Instagram in-app browser), Meta learns from only part of your buyers. If it sees duplicates, it believes some ads are better than they are.
3. Retargeting and lookalikes get weaker
Audiences like "added to cart but did not buy" and "lookalike of purchasers" are built from the same events. Missing events mean smaller, less accurate audiences, and duplicate or wrong events mean the wrong people end up in them.
4. You pause good ads and scale bad ones
When Ads Manager under-reports, profitable campaigns look unprofitable. When it over-reports, weak campaigns look like winners. Either way, budget moves in the wrong direction.
5. You cannot compare channels
If Meta's numbers are unreliable, you cannot fairly compare them with Google Ads or GA4. Budget decisions become opinions instead of data.
What does "proper" tracking mean for Meta?
| Element | Why it matters |
|---|---|
| Meta Pixel on every page | Browser events and audiences |
| Conversions API | Server events that ad blockers and browsers cannot stop |
Shared event_id |
Meta counts browser and server copies once |
| Value, currency, content IDs | Meta can optimise for revenue and match catalogue products |
| Hashed email and phone | Better matching, shown as Event Match Quality |
| Tested against real orders | Proof the numbers are right |
I explain each part in Meta Pixel vs Conversion API, Meta CAPI event deduplication and improving Meta Event Match Quality.
What does it look like after tracking is fixed?
Fixing tracking does not guarantee better results on its own, but it gives every other improvement a chance to work. In my projects, the visible changes are measurable ones. For example, on ProCloudify match quality went from around 3–4 to 9–10 out of 10 after advanced matching, and on Tea of Sylhet synced event IDs stopped purchases being counted twice.
How to check your own setup in 10 minutes
- Open Events Manager and note last week's Purchase or Lead count.
- Compare it with real orders or leads from your store or CRM for the same week.
- Open one event and check it is received from both Browser and Server.
- Check that purchases carry a value and currency.
- Look at Event Match Quality for your main event.
- Run your website through the free Tracking Checker.
If any step fails, fix tracking before changing campaigns.
Final thoughts
Meta ads without tracking are not really "Meta ads not working"; they are ads that were never allowed to learn. Get the events right first. My Meta Ads management service always starts there, or you can book a free audit.