AppLovin's Click-Only Era Is Ending, and Your Old Numbers Were Probably Wrong

AppLovin just quietly started measuring your ads in a whole new way, and it’s about to change what you think that channel is actually worth.

Up until now, AppLovin only counted clicks. No view-through credit at all, which actually made it one of the more honest attribution setups out there.

That is changing, and the direction it is changing in deserves a closer look than most brands are giving it right now, especially with AppLovin’s ecommerce ad product only recently opening to self-serve advertisers at scale.

Where Click-Only Actually Stood

AppLovin’s in-platform reporting has run on 0-day and 7-day click windows, with no view-through reporting at all, for as long as ecommerce brands have been running ads on the network.

That made the Axon dashboard intentionally conservative compared to platforms like Meta that include view-through credit by default, according to Causality Engine’s guide to AppLovin’s attribution model.

Click-only did not claim conversions from people who merely saw an ad, which is a real advantage over platforms that do.

But conservative is not the same as accurate.

Independent incrementality research has been finding a real gap between what AppLovin’s click-only dashboard reports and what actually happened.

AppLovin’s own blog acknowledged this directly, stating that the incremental factor observed, a brand’s geo-holdout-measured iROAS relative to its AppLovin in-platform reported ROAS, has more than doubled over the past year, and that a meaningful share of what the channel drives simply is not visible in its own click-only reporting, according to AppLovin’s own measurement analysis.

What’s Actually Changing

The shift toward view-through and halo measurement for AppLovin is not coming from a single flipped switch inside the platform.

It is happening across the measurement ecosystem the channel sits inside.

Some attribution platforms added AppLovin as a launch partner to its Clicks and Deterministic Views attribution model in early 2026, which combines high-intent click actions with qualified view-based exposures rather than counting clicks alone. Purchases from people who saw the ad without clicking, or who bought a different product from the same brand shortly after exposure, are now eligible to be counted in a way AppLovin’s native click-only dashboard was never built to capture.

Some Incrementality Data Backs This Up

This is not just a modeling story.

A 15-month analysis of AppLovin incrementality tests across DTC and omnichannel brands found the channel’s halo effect climbing toward levels seen on other core paid social channels, though the same research is careful to note it has not fully closed that gap.

AppLovin drove roughly a 25 percent halo effect beyond direct DTC sales for omnichannel brands in the sample, compared to a 40 percent median halo effect across other channels tested for the same brands, according to Haus’s incrementality research.

Separate third-party analysis found AppLovin’s halo rate climbing to roughly 22 percent, putting it close to Google’s, and that during peak retail periods the gap to Meta narrowed further still, according to Prescient AI’s halo effect data.

So if you wrote this channel off as small or flat based on click-only reporting alone, the data suggests you were probably measuring it with a stick that was too short.

That part of the story checks out.

Now the Part Worth Slowing Down For

Here is where the caution comes in, and it is not a small caveat.

View-through and halo attribution are modeled credit, not observed fact.

They are built on assumptions about what a person would have done without seeing the ad, and those assumptions can be wrong in either direction, sometimes dramatically so.

One case study makes this concrete.

A geo holdout test on a single window of AppLovin spend produced a 1.25x ROAS on AppLovin’s own dashboard, a 4.77x ROAS according to Triple Whale, and zero incremental lift according to the actual controlled experiment, according to Jetfuel Agency’s tactical guide to AppLovin.

Same spend, same window, three completely different answers, and the answer with the most credit attached was the furthest from what actually happened.

That is the exact risk that comes with adding view-through and halo credit on top of a click-only baseline: the new number can be more right, or it can be more wrong, and a dashboard has no way to tell you which.

This Is the Same Discipline AdBeacon Applies Everywhere Else

AdBeacon has walked through this exact tension before in what brands are getting wrong about view attribution and the final verdict on click vs view attribution.

View-through credit is not inherently dishonest. It is a real attempt to capture influence that a click-only model genuinely misses.

But it is also exactly the kind of self-reported, platform-published number that deserves the same scrutiny AdBeacon applies to any platform-reported ROAS versus actual ROAS.

AppLovin moving away from click-only does not exempt it from that scrutiny. If anything, it puts the channel on the same footing Meta and TikTok have always been on, where the honest read has never been the in-platform number alone.

What To Actually Do

Go pull your AppLovin numbers again once view-through and halo credit shows up in your reporting stack, whether that is inside AppLovin’s own dashboard, Triple Whale, or another measurement layer you run.

Do not treat the new, higher number as automatically more accurate than the old click-only one. Both are measurements, not ground truth.

Run or re-run a geo holdout or incrementality test on your own AppLovin spend before reallocating meaningful budget based on a halo or view-through figure alone.

Compare the direction and magnitude of the new number against what independent incrementality research has found for similar brands, roughly a 22 to 25 percent halo effect in the current data, and treat anything wildly outside that range as a reason to investigate, not a reason to celebrate.

If your Amazon revenue is meaningful, pay particular attention. Multiple studies have found AppLovin’s halo effect is disproportionately concentrated in Amazon sales for brands with a strong Amazon presence.

Keep a record of what each measurement layer reports side by side, AppLovin’s own dashboard, any third-party tool, and any incrementality test you run, so a single quarter of divergence does not quietly become your new baseline assumption.

The channel you thought was doing nothing might genuinely be working harder than your dashboard ever gave it credit for.

It might also be working exactly as hard as the old click-only number suggested, with the new figure simply modeling credit that never converts into real incremental revenue.

The only way to know which story is true for your brand is to test it, not to read the new number and assume it settled the question.

If you want to see your AppLovin performance measured independently, with click-based first-party data rather than platform-modeled credit, book a live AdBeacon demo and look at the real numbers.

—-

FAQ

Has AppLovin always used click-only attribution?

Yes, AppLovin’s native platform reporting has historically used 0-day and 7-day click windows with no view-through credit, making it more conservative than platforms like Meta that include view-through attribution by default.

Is AppLovin’s channel value actually undercounted?

Independent incrementality research suggests yes, to a meaningful degree. Studies have found AppLovin’s halo effect climbing toward 22 to 25 percent, and AppLovin’s own published data shows the gap between click-only reporting and geo-holdout-measured incrementality has widened over the past year.

What is halo attribution?

Halo attribution credits an ad with driving sales beyond the exact product or channel it directly promoted, such as a purchase of a different product from the same brand, or a sale on a different sales channel like Amazon, following exposure to the ad.

Should I trust a new view-through or halo number more than the old click-only figure?

Not automatically. Both are measurements with real limitations. A documented case study found the same AppLovin spend producing three very different ROAS figures across the in-platform dashboard, a third-party attribution tool, and an actual geo holdout test, with the highest-credit figure furthest from the true incremental result.

How can I verify whether AppLovin’s new attribution numbers reflect real value?

Run a geo holdout or incrementality test on your own AppLovin spend rather than relying on any single dashboard figure, and compare your results against the current range found in independent research, roughly 22 to 25 percent halo effect for most brands.