What operators say about AppLovin.
Named-brand claims pulled from public podcasts and operator interviews. Each carries its source, how it was measured, whether the appearance was sponsored, and the caveat that goes with it. Operator statements are not verified performance.
Ridge
Sean Frank reported spending about $4 million of Ridge's own money on AppLovin in Q4 2024 and said performance matched Facebook ROAS.
Caveat: CEO-reported; no raw data; period was Q4; ROAS definition not published
Listen to the source →Frank said Ridge evaluated AppLovin with Northbeam, incrementality holdouts, post-purchase surveys, and coupon codes, and reported 85% of traffic from iOS devices.
Caveat: Method designs and raw outputs are not published; device mix is one account
Listen to the source →In a later operator episode, Ridge's CMO said AppLovin purchases and revenue had tested as highly incremental relative to Meta, while noting cross-channel budget interaction.
Caveat: Direct operator interpretation; no test report or uncertainty interval
Listen to the source →Sean Frank said he was not paid to discuss AppLovin while disclosing roughly $50,000 of Ridge ad credits and saying the October test was working.
Caveat: Credit-supported self-report; no raw data or causal test
Listen to the source →Frank reported last-click ROAS comparable with Meta at roughly one-fifth the spend and similar $10-$12 CPMs in the early test.
Caveat: Last-click is not causal; no raw data; spend ratio and auction conditions are not portable
Listen to the source →A paid host read said Ridge spent about $2.8 million on AppLovin in 2025 at efficiency comparable with Meta.
Caveat: Paid read; no raw data; not evidence about The Woobles; same Ridge family
Listen to the source →Cuddle Clones
The operator reported scaling AppLovin from zero to about $80,000 per day within seven days during the Black Friday period and about $1 million in the first 30 days.
Caveat: Self-reported; seasonal; repeated appearances are one evidence family; no causal proof
Listen to the source →The operator reported AppLovin at roughly 20-30% of media mix year-round and as the brand's largest spending channel in October 2025.
Caveat: Not a recommended allocation or portable benchmark
Listen to the source →Triple Whale analysis reportedly found less than 5% overlap with Meta buyers across 42,000 Q4 orders.
Caveat: Low overlap and new-to-site status do not prove incremental demand
Listen to the source →The brand reported stronger immediate AppLovin volume for $80-$90 pajamas than for a $200-$250 custom product.
Caveat: One holiday brand and two products; no platform-wide price ceiling
Listen to the source →Purpose-built AppLovin creative reportedly outperformed direct reuse of Meta winners in the account.
Caveat: No controlled asset-level output published
Listen to the source →The operator later described AppLovin as the brand's highest-spending channel entering Q4 with multi-million-dollar monthly spend.
Caveat: Same brand family; self-reported; no causal proof or exact month
Listen to the source →The operator reported AppLovin AOV at least 10-20% higher than Meta and higher LTV.
Caveat: No cohort definition; maturity window; raw output; or causal assignment
Listen to the source →The operator said roughly 15% of post-purchase survey respondents typed a specific game name and described AppLovin-to-Google and one-click or first-click paths.
Caveat: Survey recall and click attribution do not establish causal lift
Listen to the source →A retrospective reported roughly $250,000-$300,000 of AppLovin spend over the Black Friday/Cyber Monday weekend and described checkout and warehouse strain.
Caveat: Retrospective self-report; seasonal; same evidence family
Listen to the source →The operator's broader lesson was that distressed brands often relied only on Meta and lacked retail, organic, or SEO resilience.
Caveat: Anecdotal portfolio observation; not an AppLovin performance result
Listen to the source →True Classic
Episode metadata and the former head of marketing described True Classic as an early brand to scale AppLovin from zero to about $100,000 per day in spend within days.
Caveat: Sponsored; no raw ledger or durable-efficiency evidence
Listen to the source →The guest said the newer Discovery objective had cracked prospecting in his tests and discussed near one-to-one new-visitor and new-customer ROAS reporting.
Caveat: In-platform reporting is not causal evidence; account and test window are not disclosed
Listen to the source →PROOF Wallets
Across the discussed 60-day window the founders reported about $550,000 of AppLovin spend at roughly 1.2 attributed ROAS, compared with Meta's reported 1.2 ROAS on about $1.05 million.
Caveat: Attributed ROAS is not incremental ROAS; sponsored; no raw data
Listen to the source →The founders reported about 10% multi-touch audience overlap and 90-95% new visitors across the major paid channels; they attributed about 35% of new orders to AppLovin.
Caveat: Neither metric proves causal incrementality; wording varies between orders and new orders in episode summaries
Listen to the source →Creative production capacity rather than platform performance was described as the immediate scaling bottleneck.
Caveat: Single account diagnosis; no production benchmark
Listen to the source →Jones Road Beauty
The operators described scaling an early AppLovin test from roughly $1,000 to $5,000 per day after encouraging Northbeam and Prescient signals while planning a holdout; test credits were disclosed.
Caveat: Early attributed signal; not causal; same evidence family as BPE-S14
Listen to the source →A two-week holdout reportedly found about 3% new-customer lift and 4-5% total lift; total cost per incremental order was $40-something while new-customer CPIA was near $150, making total orders profitable but new-customer acquisition unprofitable.
Caveat: Operator summary; raw study and uncertainty unavailable; tested period and account state are specific
Listen to the source →The operators said the tested setup lacked customer exclusions, new visits were below 50%, and the pattern was consistent in Northbeam and Haus.
Caveat: Historical product state; new-visit rate is not incrementality; no raw data
Listen to the source →HexClad
A sponsored operator read reported mid-seven-figure AppLovin spend, a profitable geo-holdout result, primarily new customers in MTA, and better efficiency than Meta for the cited period.
Caveat: Sponsor copy; no raw test output; comparative period and definitions not published
Listen to the source →AS Beauty
The operator described scaling from about $500 to $50,000 per day around Labor Day and allocating about 30% of BFCM budget to AppLovin.
Caveat: Sponsored; selected success story; not a scale curve or budget-share recommendation
Listen to the source →Kitsch
The operator described AppLovin as the brand's second or third acquisition platform after Meta.
Caveat: Qualitative rank; no spend or causal outcome disclosed
Listen to the source →The team reported using Northbeam as a cross-channel source of truth while watching MER daily and making decisions weekly.
Caveat: Workflow evidence only; does not validate performance
Listen to the source →BRĒZ
The operator said the brand pursued AppLovin partly because a restricted category created account instability and constrained direct-to-product acquisition elsewhere.
Caveat: Sponsored; category-specific; does not establish current eligibility for another advertiser
Listen to the source →The operator reported about $530,000 of spend in the prior 30 days at roughly $137 CAC against a stated $90-$110 target; CPM was about $24 and CPC about $0.83.
Caveat: Self-reported; no raw data; attributed metrics; period and product mix are brand-specific
Listen to the source →The operator described a $12,500-$15,000 daily spend sweet spot and a test that narrowed proven Meta videos to five before varying end cards, calls to action, and landing pages.
Caveat: The spend band is not portable and creative results were not published
Listen to the source →In a later appearance, the founder said access stopped after a cash or payment constraint.
Caveat: One-sided account; episode does not establish billing details or responsibility
Listen to the source →The later discussion warned that adding AppLovin cannot substitute for customer experience and business fundamentals.
Caveat: Commercially aligned qualitative judgment; not a measured platform result
Listen to the source →cbdMD
The operator described AppLovin as a top-of-funnel direct-to-site path for cbdMD's 21+ restricted category.
Caveat: Self-reported; current eligibility and compliance must be verified separately
Listen to the source →The operator said HTML units were performing strongest for cbdMD while noting that internal development resources made the format easier to execute.
Caveat: No controlled output or causal explanation; one brand and team
Listen to the source →The operator recommended measuring by funnel stage with post-purchase survey, last-click, and multiple windows, and starting with a proven product and offer.
Caveat: Workflow evidence; does not validate the account's performance
Listen to the source →Portland Leather Goods
The guest said AppLovin represented roughly 10% of the media mix and that the team wanted to grow it.
Caveat: Sponsored; no spend; return; or causal output; not a recommended allocation
Listen to the source →The guest said AppLovin budget was additive at that time rather than taken from Meta.
Caveat: Additive budget does not prove incremental revenue or profit
Listen to the source →The guest said a new-channel swing should be small enough not to damage the business but large and dedicated enough to matter if it works.
Caveat: Qualitative sponsored-account guidance; no universal budget minimum
Listen to the source →Multi-brand
The original ten appearances represented eight named brands and eight evidence families because both Cuddle Clones and Ridge appeared twice.
Caveat: Superseded for current totals by BPE-C42
Listen to the source →The current 20 appearances represent 12 evidence families because Cuddle Clones appears four times; Ridge four times; Jones Road Beauty twice; and BRĒZ twice.
Caveat: Repeated appearances add chronology and detail but not independent replication
Listen to the source →Sourced from Upscale's AppLovin research knowledge base (145 sources, 12 evidence families). YouTube and Meta facts are not AppLovin facts; repeated appearances are one evidence family, not independent replication.