Meta Doesn't Invent Conversions That Never Happened

Meta doesn’t invent conversions that never happened.

One of the most popular myths among advertisers is that your conversion results can’t be trusted. Meta wants so badly to make you think that you’re getting good results that they’ll simply make them up.

This is nonsense. And if you’ve fallen for this claim before, I want you to understand why it doesn’t make any sense.

I know that when I’ve tried to explain this in the past, I’m prone to be called a Meta apologist, but I’m not. The problem isn’t that Meta makes up conversions. The problem is that most advertisers don’t understand what their conversions mean.

So let’s start from a very basic level.

Assuming you’re optimizing for purchases, the Results column reflects the number of attributed purchases that have resulted from your ads. First, Meta knows that these conversions happened because the purchase event fired.

So whether it was a website pixel event or Conversions API event from another source, Meta can connect that conversion to someone who saw or clicked your ad.

Before we move any further, it’s really important to understand that this is entirely how Meta understands a purchase. Without conversion events, they didn’t happen.

So before you blame Meta for making up conversions, first confirm that your events are set up properly in the first place. Look at the overview in Events Manager to see when and where those events are happening.

Events Manager, of course, will report all purchase event fires, regardless of whether it came from your ad. It could be from you or a team member simply refreshing a confirmation page.

It’s the raw number before deduplication, so if you’re sending events from multiple sources, it could be double what you’d expect. Don’t panic when you see that.

But go into the details of the event in Events Manager to see if Meta has detected any deduplication issues. Because this is one of the most common causes for Meta reporting results that didn’t happen.

If you send events from multiple sources, you may have failed to properly deduplicate them.

But also make sure that the event is firing when it should, and not prematurely. Far too often, I’ll see purchase events fire when someone clicks the wrong button or when they’re at the wrong stage.

No matter how you do it, it should only fire when the purchase has completed.

Test this as a user and go from the ad’s landing page all the way through the completed purchase. Monitor what events are firing and when to make sure that it’s working as it should.

Next, don’t forget your ad set’s attribution setting.

If you’re using the default, Meta can report on conversions that happened within 7-day click, 1-day view, and 1-day engage-through.

Advertisers love to look at referral numbers from third-party reporting and say that Meta is lying, but there are so many holes in that analysis.

Not everyone will convert immediately upon clicking to your website, even those who fall within 1-day click. Some will come back days later, and they may even do it from a different device.

View-through and engage-through conversion results don’t require a click on your link at all, so these people won’t be confirmed with UTM parameters and other third-party reporting.

More often than not, these are conversions that happened as a result of remarketing. They were shown your ad and then converted from another marketing channel within a day.

So already, even if attribution was set up properly, your conversion results can be misleading if you don’t understand how Meta attributes conversions in the first place.

But that’s less Meta’s fault than your inability to know this.

Use the breakdown by attribution setting to get a sense of how many conversions fall under each setting. The more that are 1-day click, the better.

But we’re not done.

Advertisers often fail to realize that the Results column almost always reflects results for a standard event. For example, Meta will report all purchases that happened after someone saw or clicked your ad within the attribution setting windows.

But that doesn’t mean they bought the product you were selling in the ad.

Once again, it’s critical you understand how Meta defines conversions. Meta doesn’t care whether someone bought the product that was featured in your ad.

Meta only cares that someone performed the action that you want, in this case a purchase, within your defined attribution settings.

Know that.

And I have one more.

By default, Meta is reporting ALL CONVERSIONS that happened after someone engaged with your ad. For example, assume the following situation.

Someone saw your ad and clicked it, driving them to your landing page. They performed a purchase.

Then, three days later, they came back. It doesn’t matter what they bought, but they made another purchase.

Assuming you were using an attribution setting that included 7-day click, both purchases would be counted as separate conversions.

When you don’t understand this, you might be confused when reporting is higher than what you’d expect for total customers. It’s because you have a number of customers who may have performed multiple purchases.

You can sort this out by using the breakdown by conversion count to separate first conversions from all other conversions.

So here’s the bottom of the glass.

Meta doesn’t make up conversions, and claiming it does just makes you look like an advertiser who doesn’t know what they’re doing.

Understand what the results mean.

But also know that if Meta truly is overreporting your conversions, it’s a problem on your side to fix.

Mastering the meaning behind your conversion results is a key point in my Master Brief.

Read it at jonloomer.com/master.