Conspiracy theories are common among advertisers. When I’m asked about whether Meta can make up conversions, I have a pretty standard response:
The likelihood that there’s a reasonable explanation for inflated results is way higher than the likelihood Meta is lying.
Reasonable Explanations
The list of potential reasonable explanations that would explain your inflated results is a long one…
1. It could be an event setup problem. Is the event firing in places it shouldn’t be?
2. It could be an attribution quirk. Do you have members of your team working on confirmation pages that fire the pixel every time they load? If they see your ad and later fire a confirmation page, conversions can be reported.
3. It could be a deduplication issue. Are you sending both CRM and web events, and are you counting them twice?
4. It could be an interpretation of results issue. The Results column reflects all of your attributed conversions. That could include products that weren’t promoted in your ad.
It could also include multiple conversions by the same person. You can use First Conversion to help isolate unique customers.
And it can include view-through and 7-day click conversions that can conflict with third-party data. Even 1-day click results include clicks that don’t require a click on your link.
Convoluted Results
The results column represents what is actually a convoluted number. Not only is it easy to misinterpret, but its accuracy relies heavily on the quality of data you send Meta.
And that’s why the most likely explanation for inflated results isn’t that Meta is randomly making up numbers to make them look better. Start with understanding why the results are what they are and go from there.
