here's a popular video (by my standards) from my youtube going over Cost per conversion

Their server-side tracking worked perfectly. It was just sending Meta the wrong data. A client was running Facebook ads to a Typeform funnel. Every completed form was counted as a conversion, including unqualified leads. They had already hired multiple technical people and installed an advanced server-side tracking setup. Technically, everything was firing correctly. But Meta was being told that every form submission was equally valuable. In a 45-minute call, I found the underlying logic problem. We changed the conversion event so it fired only when someone met the actual qualification criteria, while preserving the rest of the tracking infrastructure. Now Meta receives a qualified-lead signal instead of being trained on every form submission. The lesson: More advanced tracking does not automatically produce better data. If your conversion definition is wrong, server-side tracking simply sends the wrong signal more reliably. Before changing campaigns, ask: • What event is Meta optimizing toward? • Exactly when does it fire? • Does it distinguish qualified leads from junk? • Does it match what appears in the CRM? If you’re unsure what your account is actually optimizing toward, describe your funnel below and I’ll tell you what I would check first.

The client helps freelancers and small business owners with access to things like health insurance, payroll, and employee benefits. The service was only a good fit for people earning more than $50,000 per year. The Problem The client was getting leads from Meta ads, but most of those leads were not a good fit. Before we made changes: The ads produced 500 leads. Only 200 leads were qualified. 300 leads were not qualified. Just 40% of the leads met the client’s basic requirements. The ads were producing names and email addresses, but the sales team was wasting time on people who could not buy the service. The client did not simply need more leads. They needed Meta to find more of the right people. What We Changed 1. We Added One Important Question We added an income question to the lead form. This allowed us to separate people who earned more than $50,000 per year from people who did not meet the client’s requirements. Before this change, every form submission looked like a good lead inside the ad account. After the change, we could tell the difference. 2. We Changed What Counted as a Conversion Qualified and unqualified leads were sent to different pages after filling out the form. Qualified leads reached a page that counted as a conversion. Unqualified leads reached a different page that did not send the same conversion signal back to Meta. This gave Meta better information. Instead of teaching the system to find anyone who would fill out a form, we started teaching it to find people who were more likely to qualify. 3. We Tested the Ads More Carefully We tested new images, headlines, and messages. Instead of changing everything at once, we changed one main part at a time. That made it easier to see what was helping and what was not. 4. We Kept the Brand Easy to Recognize We used the same colors, style, and main message across the ads. The ads still had different hooks and images, but they looked like they came from the same company. 5. We Used Lead Quality to Make Decisions