AppLovin Incrementality Testing for Ecommerce Brands

AppLovin incrementality testing answers the question platform ROAS cannot: is the revenue AppLovin claims actually new revenue for your business? We set a blended MER baseline, design a geo holdout or on/off test, and compare incremental ROAS with your breakeven rather than with the platform number. The result then maps to a decision rule: scale, hold and fix, or cut.

Have Us Run an Incrementality Read →See the AppLovin Ads Service →

Our method, shown on an illustrative example

5,000+ creative sets deployed

$5M+ Axon ad spend managed

50+ Axon accounts

The example below is illustrative, not a client result.

LineIllustrative value
Axon spend, test period$60,000
Platform-reported ROAS3.2x
Blended MER before Axon2.40x
Blended MER with Axon2.51x
Revenue lift over baseline~$108,000
Incremental ROAS~1.8x
Breakeven ROAS at 55% gross margin1.82x

Incremental ROAS is revenue lift over the pre-Axon baseline divided by Axon spend, not the platform figure. Breakeven ROAS is 1 divided by gross margin. In this example the platform reports 3.2x, while incremental ROAS of about 1.8x sits just under the 1.82x breakeven.

How we run an AppLovin incrementality test

Platform ROAS and blended MER split for three reasons:

  • Overlap. Axon is claiming conversions another channel also claims, so the same dollar is counted twice.
  • Substitution. Axon is reaching buyers who would have converted anyway through search or email, which looks like customer acquisition without adding new customers.
  • Scale. Axon spend is too small a share of total spend for MER to register it.

As a rule of thumb, a channel under roughly 10% of total spend will not move blended MER detectably in a month.

  1. Set the baseline. Four weeks of blended MER with Axon off or held flat, with no other channel changes.
  2. Pick the method. Use a geo holdout if Axon is 10% or more of spend: split matched regions, run Axon in half and pause it in the other half. If it's smaller, run an on/off test by pausing the channel for two weeks and reading the MER curve.
  3. Run it long enough. Two to four weeks.
  4. Hold everything else still. No new creative on Meta, no email calendar changes, no price moves during the window.
  5. Compare to breakeven. Compute incremental ROAS, then compare it to breakeven, not to the platform number.

Platform ROAS vs incremental ROAS vs MER

MeasureWhat it dividesUse it for
Platform ROASRevenue attributed to AppLovin ÷ AppLovin spend, in the reporting mode and cohort window you selectOptimization inside the platform
Blended MERTotal revenue ÷ total ad spend across every channelBusiness-level judgment
Incremental ROASRevenue lift over the pre-Axon baseline ÷ Axon spendComparing with breakeven, 1 ÷ gross margin

Decision rules after the read

Incremental ROAS vs breakevenReadingNext step
Above breakeven by 30%+Clear contributorScale in budget steps, retest at 2x spend
Within 30% of breakevenMarginal, margin-dependentHold spend, fix creative or feed, retest
Below breakevenSubstitution or overlapCut or restructure before spending more

Incrementality test FAQs

Is AppLovin incremental?

AppLovin traffic can be incremental, but attributed conversions alone do not prove it. The cleanest answer comes from a controlled geo holdout or another well-powered incrementality test that compares exposed markets with similar control markets. Treat incrementality as a testable business question, not a platform promise.

What is the difference between AppLovin ROAS and MER?

Blended ROAS, often monitored as MER, compares total business revenue with total marketing spend. Because channels can overlap and reporting rules differ, use platform ROAS for optimization and blended performance for business-level judgment.

How long should an AppLovin geo holdout test run?

Two to four weeks, after a four-week baseline. Shorter reads are dominated by weekly seasonality and promo calendars.

What if AppLovin is a small share of our ad spend?

Below 10% of spend, we use an on/off test instead of a geo holdout: pause the channel for two weeks and read the MER curve. It's noisier, but usually enough to settle a budget argument.

Can a post-purchase survey replace a holdout?

No. Post-purchase surveys and marketing mix modeling corroborate a holdout; neither replaces one.

Who wins on AppLovin Axon Ads (and who doesn't)

AppLovin Axon Ads isn't for everyone. Here's how to know if it's for you.

Brands that win

  • Proven product-market fit with an existing customer base
  • Prior paid scale of $10k+/day with positive unit economics
  • Strong creative volume and production capacity
  • Willingness to move fast, test aggressively, and iterate
  • Clear attribution model and realistic ROAS expectations

Not a fit (right now)

  • Still validating product-market fit or finding your audience
  • Limited creative resources or slow production cycles
  • Smaller budgets that require immediate ROI to survive
  • Brands with low product quality or negative reviews
  • Looking for a "set it and forget it" advertising solution

Platform ROAS vs blended MER

Blended MER is total revenue divided by total ad spend across every channel. AppLovin reports that 80% of web conversions happen within 24 hours of the click, which makes the platform number fast.

Treat platform ROAS as the engine's scorecard for its own optimization, and judge the channel on blended and incremental results. Our position is that no channel earns scaled budget on its own reporting, Axon included, and the brands that have scaled on Axon are the ones that tested first.

Watch: how we read incrementality on AppLovin

Tommy Patterson walks through what AppLovin attribution counts, why platform ROAS runs ahead of blended MER, and how the holdout test works. The full written guide: AppLovin attribution explained.

Related AppLovin services

What our clients are saying

"Your team's attention to detail on the daily optimizations has been incredible" — Beau, The Bomb Co.

"You guys are the best agency we have ever worked with" — Nick, RP Diet

"Revenue jumped by 320% while cutting CAC by 43% in 60 Days." — Denise, Alala

Ready to compound?

Ready to partner with a team of performance marketing veterans who execute at the highest level? We've put our own money on the line, scaled and exited our own brands, and we know the playbook because we've run it.

Whether you are launching on Axon Ads Manager or trying to scale past a plateau, we will tell you what we would do before you commit to anything.

Have Us Run an Incrementality Read →See the AppLovin Ads Service →