I've personally managed Meta Ads for hundreds of brands big and small
When the second payment is where the money is, optimising for the first one quietly caps how far you can scale.
Work withMost subscription accounts I look at have the same handful of problems, and none of them are settings problems.
Your reported ROAS looks acceptable but the P&L disagrees, because the platform is counting one payment and your business runs on twelve.
The discount that wins the first order attracts the people most likely to cancel after it, so acquisition and retention quietly work against each other.
Nobody can tell you what a subscriber from a given ad is actually worth six months later, so every scaling decision is a guess.
Creative sells the product but never the habit, so people buy once and never build the routine that keeps them paying.
The fundamentals hold. The economics don't. Here's what I do differently on a subscription account.
Meta will optimise for whatever you feed it. If that's a first order, you'll get people who take a first order. Sending back the events that reflect real value, a retained subscriber rather than a signup, changes who the algorithm goes looking for. That means server-side events, a considered view of what counts as a conversion, and the patience to let a slower signal do its job.
In subscription, the offer isn't a promotion, it's a filter. A heavy first-month discount buys volume and a churn problem. A trial with a genuine reason to continue buys fewer people who stay. That decision affects your CPA more than any bidding change will, so I'd rather argue about the offer than the campaign structure.
Most subscription ads sell the first box. The ones that work show the rhythm of the thing: what it's like in week six, how it fits a normal week, why people keep going. That's a different creative brief, and it needs different people on camera saying different things. It also tends to attract customers your retention team recognises as good ones.
In-platform ROAS flatters everyone and it flatters subscription brands most. What I want to see is cohort behaviour by acquisition source, payback period, and whether the ads produced revenue you wouldn't otherwise have had. Holdouts, MER, and cohort curves tell the truth. Dashboards tell you what you hoped.
I've personally managed Meta Ads for hundreds of brands big and small
A full look at the account, the tracking, the offer and the creative. You get a prioritised list of what's costing you money, ranked by what I'd fix first.
Get the events and values flowing back to Meta so it's optimising toward retained revenue rather than a first transaction.
Angles mapped to who actually stays subscribed, then a testing plan that builds knowledge instead of producing random ads.
Spend increases tied to cohort performance and payback, not to whatever the platform reported this week.
What founders say about working with John on Meta campaigns
John does the work himself, which after three agencies felt almost strange. He found problems in our account nobody had mentioned in two years, fixed them quickly, and explained every decision in plain English. Our creative output has never been stronger.
Giles HumphriesFounder, Mindful Chef
What I value most is the honesty. John tells me when an idea won't work and why, rather than nodding along and billing me for it. He's genuinely part of the business now, not a supplier I chase for updates.
Sandra Roycroft-DavisFounder, Slimpod
It depends on volume. Optimising for the deeper event is almost always better if you have enough of them for Meta to learn from. If you don't, you optimise for the earlier event and manage the quality problem deliberately rather than pretending it isn't there. That's a judgement call I'd make after seeing your numbers, not a rule.
Partly. I can't fix a product or an onboarding experience with creative. What I can do is stop acquiring the people most likely to leave, which is often a discount problem rather than a targeting problem, and put the reason people stay into the ads so expectations are set correctly before they buy.
You'll see creative and cost signals early, but the numbers that matter in subscription take a cohort to mature. That's the honest answer. Anyone telling you they'll prove lifetime value impact in three weeks is describing something that isn't possible.
Probably fewer things, done better, by someone senior. The most common thing I find in agency-run subscription accounts isn't incompetence, it's that nobody has connected the media buying to the retention numbers, because those two things sit with different people who rarely talk.
Tell me what you're spending, what you're seeing, and where it stops making sense. I'll come back with an honest view on whether I can help and what I'd do first.