I've personally managed Meta Ads for hundreds of brands big and small
Getting people through the door is the cheap part. Everything worth optimising happens after the download.
Work withApp accounts fail in a distinctive way. The dashboard looks healthy right up until someone asks what the users are worth.
Cost per install looks excellent and revenue has not moved, because you are buying the cheapest possible download rather than a user.
Your best performing creative attracts people who open the app once, which means the algorithm is being taught to find more of them.
Attribution is fragmented across platforms and nobody trusts any of the numbers enough to make a spending decision.
Creative is a store screenshot with a caption, so people arrive on the listing with no idea what they are meant to do next.
The principles hold. The signal chain is longer and more fragile, so more of the work sits in what you send back to Meta.
An install is a step, not an outcome. Whether that outcome is a subscription, an activation, a purchase or a return visit, that is the event Meta should be learning from. It needs enough volume to learn, which sometimes means optimising slightly earlier than you would like and managing quality on purpose rather than by accident.
You are asking someone to leave the feed, read a listing and commit to a download, and most drop off happens in that gap. The ad and the listing need to tell the same story. I treat them as one unit, the same way I treat an ad and a landing page.
Screen recordings of the actual experience consistently earn more attention than polished brand films, because people want to know what they are getting. The hook is usually the moment the app does the useful thing. Get to it fast.
Between SDK setup, aggregated measurement, delayed reporting and platform limits, app measurement is genuinely harder than web. That is not a reason to guess. It is a reason to get the events, values and configuration right before spending more, because everything downstream depends on it.
I've personally managed Meta Ads for hundreds of brands big and small
Events, SDK configuration, campaign structure and creative, plus an honest read on which of your numbers can be trusted.
Get the events that reflect real value flowing back cleanly, with sensible values attached, so the algorithm optimises toward users worth having.
In app footage, hooks built on the moment of usefulness, and a testing plan that treats the store listing as part of the journey.
Spend follows cohort behaviour rather than install volume, so growth does not quietly fill the app with people who never come back.
What founders say about working with John on Meta campaigns
Having one experienced person accountable for everything changed how fast we move. Questions get answered the same day, tests go live the same week, and I always know exactly what's being worked on and why it matters.
CyrusFounder, re:sleep
We'd been told our category was too niche for Meta to work properly. John rebuilt our creative approach around who actually buys from us, and the difference was obvious within weeks. No jargon, no endless meetings, just work.
Jason GrimaFounder, Swallow Co.
In app events, whenever you have enough of them for Meta to learn from. If you do not, the honest answer is that you optimise earlier and accept a quality trade off while you build volume, rather than pretending a cheap install is a result.
Your developers will need to implement, but I will specify what needs to exist and check it is firing correctly. Most of the app accounts I see have a measurement problem rather than a media buying problem, so this part is not optional.
It limits the granularity of what you can see and slows reporting, so decisions have to be made on fewer, more considered signals. It does not prevent the platform working. It does punish anyone who was relying on granular attribution instead of a strategy.
It changes measurement more than it changes the Meta work. When several channels claim the same install, you need a view of incrementality rather than a comparison of self reported numbers, otherwise you scale whichever platform is most generous to itself.
Tell me what you're spending, what you're optimising toward and what happens after the install. I'll come back with an honest view on whether I can help.