Here’s a good way to confirm your results…
If you feel that Ads Manager is overcounting or undercounting conversions, or you just want a more precise measurement, this is what I often do.
The specific examples when I’ve done this are when I’ve run tests of targeting. I created multiple ad sets and wanted to see which approach led to the most results. But, I didn’t want to simply take Meta’s word for it.
While I did this to test targeting, you could technically do it for any situation when you want to increase your confidence in the reported results. It’s a good secondary backup.
What to Do
Create a unique landing page for your advertising. The main thing is that the landing page isn’t otherwise publicly accessible (it could be a duplicate of your normal landing page). Use a form on that landing page (for purchase or leads) that is only used for these ads. Then make sure that those who convert get a special tag. How this is done will depend upon your CRM.
Most recently, I did this when testing three different targeting methods when driving people to a lead magnet (I also used this approach here). Each ad set used different targeting. The ads within them sent people to a different landing page (one of three that looked the same but had a different form).
I was then able to run a report in my CRM to confirm the number of conversions that came through each form and page. Even though there’s the possibility of including the random organic conversion, this is rare, and I’d argue that these are worth counting anyway.
Regardless, I’ve found it to be more precise than Ads Manager. It’s just good practice to have a backup method for confirming results.
This is my preference. It may be a little extra work, but it gives me results that I trust.