AppLovin posted $1.9 billion in Q2 revenue and roughly $1.3 billion in net income, both up more than 50% year over year — and the stock still fell more than 20% after hours. The numbers were not the problem. The pace was: investors wanted the expansion into ecommerce and consumer advertising to move faster, and it is moving at the speed advertiser budgets actually move, which is slower than any growth narrative survives contact with.
CEO Adam Foroughi’s explanation is the part worth clipping and keeping: “We’re deemed a new bucket, so a testing category. And to graduate up takes time.” Mid-tier ecommerce and consumer brands plan budgets one to four quarters out; search and social still hold most of the money by default; and a new channel — however spectacular its performance numbers — has to survive at least one full planning cycle as a line item labelled “test” before it earns a recurring allocation. Foroughi is playing the long game deliberately, prioritising “client acquisition rather than going for the most widescale possible adoption” over marketing blitz.
Strip out the tickers and this is not really an AppLovin story. It is the clearest public documentation in some time of the physics governing how advertiser money migrates to new channels — physics that apply with full force to every publisher currently launching a commerce media product, a curated audience offering, or newly programmatic inventory.
The numbers in this piece
01The testing-category physics
Every publisher revenue innovation of the last three years — commerce media, curated deals, data products, newsletter programmatic, agent-accessible inventory — is, in the buyer’s planning spreadsheet, exactly what AppLovin is: a new bucket. And new buckets obey rules that have nothing to do with the quality of the product.
The rules are visible in Foroughi’s account. Budgets for mid-tier and non-endemic advertisers are planned one to four quarters ahead, so a product launched in August is pitching into money already allocated through spring. Incumbent channels hold the default: search and social do not have to re-win their allocation each cycle, but every new line has to win its first one. And graduation is sequential — test budget, expanded test, planning-cycle inclusion, recurring line — with each step gated by a planning calendar you do not control. AppLovin has performance numbers most sellers would kill for, a self-serve growth machine, and a decade of ad-platform credibility, and it still cannot compress those steps. Your new product will not either.
That reframes what failure looks like. The most common way publisher innovation dies is not advertiser rejection — it is internal expectation mismatch. A new product gets forecast as if demand were instantaneous, delivers small test budgets for two quarters exactly as the physics predict, and gets killed at month five for “underperforming” — while sitting on a perfectly normal adoption curve. The forecast was wrong, not the product. AppLovin’s quarter is what that same dynamic looks like with a ticker attached: 50% growth, marked down 20% for missing a timeline the planning cycle was never going to allow.
There is a second lesson in Foroughi’s sequencing. Rather than marketing for maximum adoption, he is concentrating on landing clients, documenting their results, and letting reference cases do the scaling. That is the correct sequence for any publisher product aimed at non-endemic money: the second buyer does not believe your pitch deck, but they believe the first buyer’s case study. Reference clients are not early revenue; they are the sales asset that makes revenue possible.
02Why this matters for publishers
| Your new products are someone's testing category | Commerce media, curated audiences, newly programmatic inventory — in the buyer's spreadsheet these are new buckets, subject to the same one-to-four-quarter graduation physics AppLovin just described, regardless of how good the product is. |
|---|---|
| The planning calendar outranks the product | Money committed to search and social through the next several quarters cannot move to you until the cycle turns, no matter what your performance data says. Launch timing against planning season matters as much as the offering. |
| Expectation mismatch kills more products than buyers do | A normal adoption curve read against an instant-demand forecast looks like failure. The forecast is the controllable variable — control it. |
| Reference clients scale; marketing doesn't, yet | For a new bucket, a handful of documented case studies moves the next ten buyers more than any amount of outbound. Foroughi's client-acquisition-first sequencing is the template. |
AppLovin grew everything 50% and got marked down 20% for the crime of moving at the speed advertiser budgets actually move.
03What publishers should do
04The bottom line
AppLovin grew everything 50% and got marked down 20% for the crime of moving at the speed advertiser budgets actually move. Publishers should read that as calibration, not schadenfreude: if the fastest-growing ad platform of its generation needs multiple quarters to graduate out of the test bucket, your commerce media product and your curated audience offering will too — and the difference between the products that make it and the ones that die at month five is mostly whether anyone set the internal clock correctly at launch. Patience, in this market, is not a virtue. It is an operating discipline with a timetable, and the timetable is the buyer’s planning cycle, not yours.