Starting position
A direct-to-consumer glass brand with proven demand and Meta as its primary paid channel. The category is visual and crowded, so Meta costs had been rising, and the brand needed a second acquisition channel that reached people Meta was not already serving. Measurement ran through Triple Whale alongside Shopify.
The problem
Growth on Meta was becoming a question of how much more to pay for the same buyers. The open question was whether AppLovin Ads (formerly Axon) would find different customers, or simply re-attribute orders Meta would have won anyway.
What we did
Seeded the AppLovin pixel with the store’s historical Shopify purchase data before launch, so the model started with a buyer baseline. Structured the account around purchase optimization with the Shopify Orders API recovering conversions the browser pixel missed.
Built 9:16 video for full-screen game placements with a product-in-use hook, captions for sound-off play, and end cards carrying the offer and call to action.
Reported from Triple Whale rather than the AppLovin dashboard, tracked new-customer share separately from blended ROAS, and checked order overlap with Meta to see whether the two channels were reaching the same people.
Launched January 2026. Week-by-week spend and ramp markers are being added from the account records.
Results
Cumulative since launch. Every figure is labeled with where it was read.
| Metric | Value | Source |
|---|---|---|
| Attributed revenue since launch | $595K | Triple Whale |
| New-customer revenue | $450K | Triple Whale |
| ROAS | 3.71× | Triple Whale |
| New-customer ROAS | 2.89× | Triple Whale |
| New-customer share of orders | 73% | Triple Whale |
| Share of total store revenue | 18.7% | Triple Whale |
| Click-through rate | 4.04% | AppLovin Ads Manager |
| Cost per click | $0.72 | AppLovin Ads Manager |
Overlap. Triple Whale shows a 61% order overlap between AppLovin and Meta, which means roughly four in ten AppLovin-attributed orders had no Meta touch at all. That is the strongest signal in the account that the channel is adding buyers rather than re-labeling them.
Discovery effect. In multi-touch paths, customers often enter through AppLovin and convert hours or days later through Klaviyo, organic, or Meta. The 18.7% revenue share is a click-attributed figure and likely understates the channel’s influence.
Retention. 19% of new customers acquired through AppLovin repurchased within 18 days (Triple Whale).
AppLovin Ads Manager screenshots of spend, cost per purchaser, and ROAS over time, plus 2–4 creative frames, are being added with the client’s permission.
What we learned
Starting the model with real purchase history shortened the learning phase and kept early cost per purchaser inside a range the brand could tolerate. We would not launch an account without it now.
Blended ROAS looks better than the number that matters. The 2.89× new-customer figure is what justifies the spend as acquisition rather than retargeting.
AppLovin’s click-only attribution and Triple Whale’s model do not match, and they are not supposed to. Pick one source of truth before launch and report from it consistently.