Most Advertisers Should Only Use Incremental Attribution

Most advertisers should only use incremental attribution.

So my opinion on this has evolved during the past year, but there’s a good reason for that. There are also some important caveats behind this recommendation that I’ll make sure to address.

I truly believe that making incremental attribution the default would solve all sorts of problems.

Let me explain.

Meta first announced incremental attribution nearly two years ago in August of 2024.

Now standard attribution is the default attribution model. And when standard attribution is used, credit is given to ads when people convert after clicking or viewing your ad.

For example, the default standard attribution setting is 7-day click, 1-day engaged-view, and 1-day view. So Meta will report on any conversion that happens within any of those conversion windows.

But that also means that conversions can happen incidentally where your ad may not have actually caused the result.

For example, an impression of your ad appeared and the person it was shown to then converted within a day. They may not have even seen your ad, but they responded to another marketing channel later on.

This is an example of remarketing, and it’s exactly why I often warn that remarketing isn’t incremental.

So when using standard attribution, your default results can be a bit, if not grossly, inflated and misleading, especially if you don’t care enough to dig. And if you isolate your remarketing audiences to only target these people, that is even more likely to be the case.

If your results are made up of a high percentage of view-through and engage-through conversions, a strong argument can be made that many of these results would have happened without your ad.

This is where the incremental attribution model comes into play.

It does not rely on click-through, engage-through, and view-through attribution to give ads credit. Instead, incremental attribution uses advanced models to predict whether a conversion was actually caused by your ad.

And that’s an important distinction.

Cause and effect aren’t required for standard attribution, and that’s how results can become misleading.

The applications of incremental attribution evolved from a purely reporting option to also include an optimization option during the past year or so.

That means you can use incremental attribution to impact both delivery of your ads as well as the reporting of your results.

When using standard attribution, Meta can load up on view-through conversions if it will improve the number of results reported. But that’s not an option with the incremental attribution model.

Now, all of this said, I don’t always use incremental attribution as my attribution model when creating an ad set.

The reason behind this is simple: I obsess over my results, and I know what they mean.

So if there are 100 reported conversions, I know whether those results are propped up by view-through, engage-through, 1-day click, or 7-day click results.

I also know how many of those results Meta believes to be incremental, since I’ll add a column when using the Compare Attribution Settings feature.

But most advertisers don’t do this.

Most advertisers use the standard attribution model, and they don’t obsess over what the results mean. They take their results at face value.

Even advertisers who specifically target remarketing audiences do this, which I find absolutely insane.

One of my biggest pet peeves lately is that most advertisers completely misunderstand their results. They don’t dig beyond the surface, and this leads to bad interpretations of what is working and what’s not.

It leads to really bad decisions regarding scaling or turning something off or starting over.

When you don’t understand your results, you’re not in a good place to make good decisions.

So I’ve been fighting this problem for the longest time, and it occurs to me that Meta could put an immediate end to this fight with a change.

The problem isn’t necessarily that everyone should use incremental attribution instead of standard, since I still often use standard.

The problem is that the vast majority of advertisers make mistakes because they don’t know what to do with results that come from standard attribution.

They don’t break down their results by attribution setting. They don’t use the compare attribution settings feature to understand how many of the conversions Meta believes are incremental.

So they’ll try to scale a campaign that is largely view-through and engage-through conversions, and they’ll be confused when that scaling fails.

Or they’ll turn off an ad that doesn’t appear to be performing as well as another, when the ad it’s compared to is propped up by view-through conversions.

Meta could save a lot of advertisers from themselves by simply making incremental attribution the default attribution model.

Now, the obvious conflict for Meta is that making this change will mean fewer conversions reported by default.

A very large percentage of advertisers are blissfully ignorant that their ads are doing more than they actually are, and that leads to more ad spend.

On one hand, it’s not entirely Meta’s fault that advertisers don’t use these tools to better understand their results. But you also can’t ignore the bad decisions based on bad data that could be avoided.

There’s also the matter of so many advertisers claiming Meta makes up conversions, and that’s probably a reputation that Meta doesn’t want.

And to be clear, Meta doesn’t make up conversions, it’s just that most advertisers don’t understand the results.

So I’ve found that results from incremental attribution tend to line up more or less with your click-through attribution results from standard attribution.

If most of your conversions come from click-through attribution, which mine do, you’re unlikely to see much of a change when using incremental.

That said, if this change were made, your results absolutely would drop on the surface. But if you add columns for view-through and engage-through attribution, you’re bound to find out that those are still happening.

Meta just isn’t reporting them by default.

I truly believe that the vast majority of intermediate to advanced advertisers who have tried and abandoned incremental attribution only did so because of this drop.

They think it isn’t working as well, but the reality is that it’s probably only a perceived drop more than anything. Instead, these are the core results that actually mean something.

I get that there’s a risk to Meta implementing a global change like this due to that perceived drop in results.

But they could also mitigate this by adding a default column for conversion results from standard attribution that could exist alongside incremental.

Or, I don’t know, do a better job of educating their advertisers about what to use, when, and what the results mean.

So here’s the bottom of the glass.

While I don’t always use incremental attribution, I believe strongly that the vast majority of advertisers would benefit by switching to it full time.

If I’m to be honest, it’s because the vast majority of advertisers don’t put in the work to understand what their results actually mean.

Now, my preference would be that advertisers simply make this decision themselves, but I doubt that’s going to happen.

Meta could push or force incremental attribution, which could lead to much smarter decisions based on meaningful results. The risk is that it could lead ill-informed advertisers to believe their results are worse.

And I get it, that’s a potential problem.

If you don’t switch to incremental attribution full time, I’m begging you. Understand what your conversion results actually mean.

Know how many of the default results Meta considers incremental before making any decisions about scaling or pausing campaigns, ad sets, and ads.

When you do, your clients will thank you for it.