The Install Is Just the Beginning: Mick Rigby on Retention, AI Discovery, and the Next Era of App Growth

Full analysis

Mick Rigby, founder and CEO of app-growth agency Yodel Mobile, spent an hour with Signal & Noise hosts Brett House and Rio Longacre making one argument: the install is a garbage metric, and the money you spend chasing it is mostly wasted. His number is that 95% of app users are gone inside 30 days, and 75% quit within three. The interesting new wrinkle is that large language models are now reading your public App Store listing to decide what apps to recommend, which quietly rewrites the rules of App Store Optimization (the practice of tuning a listing so it gets found).

What's actually being surfaced here: not a decision, a shift. Two of them. First, mobile user acquisition is still, five years after Apple's privacy crackdown, a discipline most brands do badly. Second, AI-driven discovery is opening a new front where the winning move is public-facing copy, not hidden metadata. For most programmatic operators this is Type 2, easy to reverse, cheap to test. The forcing function is that AI app discovery is happening now, and the listing text is a free variable you already control.

The Market Analyst. The power map Rigby draws is the whole story. Google, Meta, and TikTok take roughly 67% of mobile ad dollars, and Apple's App Store ad business has grown to an estimated $8 to 9 billion, up from $3 to 4 billion five or six years ago. Treat that last figure as an informed guess, not disclosed revenue. What it tells you: Apple's ATT privacy change (App Tracking Transparency, the pop-up that lets users block cross-app tracking) starved everyone else's targeting while Apple's own ad business tripled. For an informed outsider: the referee started selling tickets. Anyone whose mobile revenue depends on Meta and Google targeting has already lived this. The next front is regulatory, and the EU is the venue.

The Skeptic. The episode is thin on sourcing and thick on round numbers. The 600,000 new apps and 30% jump? Unverified. The $8 to 9 billion? An estimate. The claim that Apple is suing OpenAI over stolen phone plans? Brett House got that wrong, there's no such case on the public record. So discount the specifics. But the retention argument survives even if every number is soft, because it doesn't depend on them. "An install might not even be a human," Rigby said, and that's true whether churn is 95% or 80%. For a non-specialist: he's saying the download counter everyone brags about measures almost nothing worth having.

The Operator. Rigby's real point is organizational, and it's the one operators will fumble. He says product, marketing, and communications have to run as one team, because churn happens when the ad promise doesn't match the app. Tuesday morning, that means the growth lead who bought the install doesn't own the onboarding screen that loses the user on day one. Different team, different budget, different boss. The AI-ASO test is the easy win: rewrite your public listing copy for how an LLM reads it, ship it this sprint, cost is near zero. The retention fix is the hard one, because it's a reporting-line problem, not a media problem.

The Customer / End User. Two customers here. The app marketer, who's being told to model lifetime value before spending a dollar on acquisition, which is correct and which almost nobody does because installs are legible to a CFO and cohort retention curves are not. And the actual app user, who downloads, gets a day-one experience that doesn't match the ad, and leaves inside 72 hours. Nobody is asking for more installs. The user is asking the app to do the thing the ad said it would.

The CFO. The flip-the-funnel pitch is a budget-timing argument dressed as strategy. Model the value first, then spend on acquisition. The catch: LTV modeling on a young app is guesswork, and finance hates funding guesswork. What's genuinely cheap is the AI-listing test, so fund that now. What's expensive and slow is re-orging product and marketing under one P&L, and that only pays back if retention actually moves, which takes two or three quarters to prove. For programmatic operators outside the mobile UA stack, the direct dollar impact of this episode is low. The read-through is the value.

Where the council splits. The Analyst sees a structural power story with a regulatory sequel. The Operator sees a boring org-chart problem that no amount of AI will fix. Those aren't in conflict so much as aimed at different readers: if you sell into mobile budgets, the duopoly framing matters; if you run one, the reporting-line problem is your Tuesday. The Skeptic's job is to keep both honest, because the episode leans on numbers that would not survive a fact-check, and two of them didn't.

What this hinges on. One belief: that LLMs are now materially shaping app discovery off public listing text. If that's real, the AI-ASO test is free money and worth running this month. If it's early hype, you've spent an afternoon rewriting copy and lost nothing. Either way the downside is trivial, which is exactly why the retention and org arguments, the ones with real cost, deserve more scrutiny than the AI angle that's grabbing the headline. Verify the discovery claim on your own listings before you build a product around it.

Honest confidence that this episode moves anything measurable for a broad ad-tech operator in three months is low. It's a strong practitioner conversation about mobile UA with real read-through on concentration and AI discovery, but the specific, checkable claims here are mostly about mobile-app growth, a stack most of the reader base doesn't run. The one grounded, on-topic call available, that Apple's ad business keeps climbing, is directionally obvious and not something a reader could cleanly grade against an anchor, since Apple doesn't break the number out.

No high-conviction prediction this week.