Incrementally on Meta – The Metric That Actually Matters

Most advertisers still obsess about the wrong things. They argue over attribution windows. One day click or seven day click. Should we include view-through or not. Should we match Ads Manager to Shopify or GA4. They build entire reporting structures around which window makes them look the best. The truth is none of those debates matter if the sale was going to happen anyway. Counting it under a one-day or a seven-day window doesn’t change the fact that you just paid for a conversion that wasn’t caused by your ads. That’s the entire problem with how most people look at performance, and that’s where incrementality comes in. Incrementality asks the only question that really matters: did the ad actually cause the conversion. Not who should get credit, not which click counted, not how many platforms claimed it, but whether the impression actually moved the customer to buy. And this is not just a measurement exercise anymore. For the first time, Meta now lets you attribute and optimize directly for incrementality inside Ads Manager. That means you can stop optimizing campaigns to win attribution credit and start optimizing them to actually create sales. It is a complete shift in how the platform delivers, how results are reported, and how budgets should be scaled.

The way Meta’s reporting works today is designed to make Meta look good. Somebody sees an ad, clicks, buys in the next week, and Meta takes credit. If that person was going to buy anyway, it doesn’t matter. Meta still reports it as a conversion. This is why retargeting always looks like the hero in an account. You show ads to people already in the funnel and Meta gives itself the win. But when you take away the ads and look at a holdout group, a big chunk of those buyers come back and purchase anyway. The retargeting ROAS that looked so amazing was just inflated credit. Prospecting always looks worse by comparison because it’s actually trying to create demand. Platform reporting punishes the thing that drives growth and rewards the thing that collects inevitable sales. We’ve seen this play out across every platform. Google does it. TikTok does it. They all want to show they drive results, so they grab as much credit as possible. That’s why you can have Meta reporting 100 conversions, Google reporting 100 conversions, TikTok reporting 50 conversions, and your store only showing 120. All three are fighting to claim the same sales. Add them up and you’ve bought more conversions than you actually had. Incrementality cuts through that noise. If the sale would have happened without the ad, it doesn’t count. If it only happened because of the ad, it’s incremental. That’s the distinction.

Before this update, the only way to really measure incrementality on Meta was through Conversion Lift studies or geo holdouts. Valuable, but they were one-off tests. You ran them, you got a report, maybe you adjusted some budgets, and then you went back to the same old platform ROAS reporting. Now, with the incremental attribution option inside Ads Manager, you can tell Meta to not just measure but actually optimize for causality. Instead of the system chasing people who are most likely to convert anyway, it chases people where an impression is predicted to make a difference. Reporting shifts with it. That toggle is the difference between paying for credit and paying for growth. What happens when you flip it is uncomfortable at first. CPMs usually rise. Retargeting starts to look weaker. Prospecting often gets more delivery. Spend starts to shift across markets. Why? Because the model stops rewarding cheap credit grabs and starts rewarding actual lift. If a market was inflated by attribution lag, like we’ve seen before with Europe stacking conversions on 7-day click, that market shrinks. If a market like the U.S. or AU is actually responsive to ads, it gets more spend. You don’t have to reallocate budgets manually. The system reallocates because it sees where ads really work.

This change also impacts your entire funnel. When Meta stops flooding you with inevitable buyers and starts sending you traffic that actually needs convincing, your downstream performance improves. Your email capture looks better. Your retargeting on Google and TikTok looks better. Even your overall blended ROAS improves, because the top of the funnel is now filled with higher-intent traffic. Optimizing Meta for incrementality doesn’t just fix Meta’s numbers. It makes other platforms stronger, because the audiences they retarget are more incremental. Of course, when you flip to incrementality, your dashboard looks worse. Your conversions drop. Your ROAS shrinks. But ask yourself, what changed? Did the ads suddenly stop working, or did the illusion disappear. Take a simple example. You spend $100,000. Ads Manager reports 300 conversions and $300,000 in revenue. That’s 3x ROAS. Looks incredible. Flip to incrementality and you see 180 conversions, $180,000 revenue. That’s 1.8x iROAS. The number doesn’t feel as good, but it’s the truth. If your breakeven is 1.4x, you still scale. But you scale with a realistic ceiling instead of scaling a fake 3x and wondering why your bank account doesn’t match Ads Manager.

The way to prove it is simple. Duplicate a campaign. Keep everything the same — creative, budget, audience — but change the attribution model on one version to incremental. Run both side by side for a few weeks, long enough to cover a purchase cycle. Don’t interfere. At the end, compare not by platform ROAS but by iROAS and iCPA. The incremental version will almost always show fewer conversions but a higher percentage of them are real. Delivery will have shifted. Prospecting will look stronger. Retargeting will look weaker. And your finance team will actually be able to reconcile the ad numbers with real revenue.

This is also where attribution windows finally make sense. In Ads Manager you can now line up 1-day click, 7-day click, 28-day click, and incremental attribution side by side. That view shows you the exact gap between what you would have reported under old models and what you actually created. We’ve done this with market reports before. Some regions leaned more toward 1-day click, others stacked conversions in 7-day. Incremental attribution exposes which markets were living off attribution lag and which markets actually responded to ads. And because the optimization is tied to it, Meta naturally shifts delivery toward the markets with real lift and away from the ones inflated by windows.

The practical impact is big. Under the old model, you’d see U.S. retargeting eating half your spend because it reported a 6x ROAS. Under incremental optimization, that budget redistributes. Retargeting still runs, but it only gets spend if it proves incremental. Prospecting gets healthier. Creative that actually changes minds performs better. And your account mix shifts away from the vanity buckets that looked great but weren’t driving growth. This isn’t magic. Bad ads will still fail. Weak offers will still struggle. Incrementality is not a band-aid. What it does is expose reality. It forces the system to prove its value. And when you combine it with strong creative and offers, it gives you a more honest, more scalable account.

It’s worth pointing out that a lot of advertisers rely on third-party attribution tools, MMPs, or a patchwork of incrementality tests run separately across platforms to try to piece together the truth. The problem is that most of those tools are building their models on top of what the platforms already report. They try to find the right balance of credit to assign, but the results are still skewed based on how good each platform is at attributing closer to the conversion. And they almost never measure impression value. They only measure once a user lands on the website, which is just a fraction of the journey. That’s why Meta’s ability to optimize for incrementality is such a big deal. It gives advertisers at least some confidence that the ads are being targeted toward users who wouldn’t convert without it. And that doesn’t mean these are always new customers. They could be repeat buyers or warm audiences. Incrementality doesn’t equal “new.” It simply means the conversion would not have happened without the ad. And in theory, that’s exactly the metric advertisers should want to pay for.

And this is why the endless debates about attribution windows don’t matter anymore. One-day click, seven-day click, view-through on or off — those are just reporting views. They were a crutch for a world where every platform fought for credit and nobody could agree on whose number was “right.” Incrementality ends that debate. It tells you whether the ad caused the sale. Everything else is noise.If you care about profitable scale, you have to run on incrementality. It won’t always make your dashboard prettier, but it will make your business healthier. You’ll stop wasting money on campaigns that collect conversions and start scaling the ones that create them. You’ll stop fighting over attribution overlap and start managing budgets by real outcomes. And once you make the switch, you’ll never go back, because you’ll realize how much time and money you wasted chasing credit instead of growth.

At the end of the day, ads shouldn’t be judged by how many sales they can claim inside a window. They should be judged by how many sales they actually create. That’s what incrementality tells you. And that’s why it’s the only metric that matters. And don’t be surprised when you see other platforms start adopting the same approach. This solves the problem of platforms not working together, each one grabbing credit for the same sale. If they want to keep proving value, they’ll have to work for individual incrementality instead of inflated attribution. Meta was the first to roll it out, but you should fully expect Google, TikTok, and others to follow. Because once advertisers see the difference, credit doesn’t cut it anymore. Only causality does.