Ad production for subscription apps and mobile games, by Growth Therapist.
Creative strategy and production, built from your product, your audience, your competitors and what your account already rewards. We write the concepts, hooks and scripts, decide what each version tests, and direct production through to the finished ad and the next round.
Google
Deutsche Telekom
Zalando
Volkswagen
Bosch
Intel
Oracle
WoogaIf the product is good, the ad usually fails in the first three seconds.
That is where the concept work goes, and it is why games and apps are briefed differently.
Show the product in the first second
A game can put the mechanic on screen straight away and the viewer understands it. The ad opens on the moment that makes the mechanic obvious, and the caption does the rest.
Name the problem before the screen
An app usually has to say what is wrong in the person's own words before the screen means anything. That changes the script more than it changes the format.
UGC, AI and gameplay ads that won their test.
Ten winners from four apps and games. Tap one to play it.
What the ads actually are.
Four formats. Which ones you get is decided by what your account, your product and your competitors show.
Footage taken from inside the product, cut with captions, text and motion graphics.
- gameplay hooks
- app walkthroughs
- screen recording integration
For games gameplay, usually the moment that makes the mechanic obvious in three seconds.
For apps the screen doing the one thing the person downloaded the app to do.
A real person films on their own phone, in the everyday style that reads as a post, not an ad. We handle the direction, the script and the opening line, and the person on camera comes from our creator network. Their fee is quoted on top of the Sprint.
- talking head delivery
- reactions
- voiceover
For games a reaction, or someone playing and talking over it.
For apps someone naming the problem in their own words before anything gets shown.
AI presenters and AI footage, with no shoot behind them. Cheap enough to put an opening line in front of a real audience before anyone books a camera.
- AI presenters
- avatars
- AI voices
- multilingual versions
For games rarely the finished ad, useful for finding which claim is worth filming.
For apps a presenter reading a script, useful for testing which promise lands first.
Short 2D animations that tell the idea without product footage. Useful when the promise is the story, or when the real product is not ready to show.
- animated concepts
- overlays
- UI callouts
- end cards
For games the mechanic explained in motion before the polished gameplay exists.
For apps the problem and the promise drawn out when the screen alone does not carry them.
What is in every ad
- The opening, written for the first three seconds
- The script
- Captions burned in
- Voiceover where the idea needs one, AI or a real voice
- Music and sound design
- An end card
- 9:16 plus one secondary ratio, as 1080p MP4 with named files
- A note of which variable that version tests
Gameplay is captured from your build. 2D overlays and animation, AI presenters and a creator on camera come in with the format that needs them. All of it is inside the price; only the ad spend and a creator's fee sit on top.
3D renders and playable ads are not something we produce. If a test points that way we will say so and point you to a studio that does.
What happens after you book
- 1Brief
The app or game, who it is for, and a look at the ad account if there is one.
- 2Ideas
Written from what the account has already rewarded. For a launch, from the product and the competitors.
- 3You approve
The ideas and the scripts, before anything is produced.
- 4Production
Test versions of every idea, live in your account in about two weeks.
- 5Fourteen days live
Next to the ads you already run, each with a fair share.
- 6The read
What won, what lost, what to make next. The winner gets full production.
The same six steps run in the Sprint and every month of the Loop. How many ideas and versions each round holds is what the two prices decide.
We put an external creative studio on the same account. When Meta chose, it moved the budget to our ads.
Horse Racing SolitaireOne mobile game, one Meta account, December 2025 to April 2026. The studio got the brand kit, the gameplay footage and a written brief. It did not get the account.
| The studio | Our team | |
|---|---|---|
| Best CPI, ads over £10 spend | £1.21 | £0.65 |
| Ads under £1.00 per install | 0 of 16 | 14 of 175 |
| Share of budget once Meta chose | 30% to 4.5% | about 70% to over 95% |
The studio's ads entered in March, so the first two rows set 175 of ours against 16 of theirs, and only four of theirs passed £10 of spend. In March alone, blended CPI was £1.34 on our side and £1.49 on theirs, a gap of 11 percent. The budget row compares like for like over 22 days in the same value optimization campaigns. The account ran in pounds; the prices on this page are in euros.
Every ad started in an install campaign. The best ads from each side then moved into the value optimization campaign, where Meta decides how to split the money. The chart below is that campaign.
CPI only hints at quality. The proof is where Meta puts the money: it tried the studio's ads, gave them 30% of the day's budget on 28 March, and had taken that back to 4.5% by 12 April. In the accounts we run, a new ad takes about a week to find its share and a winner can hold it for weeks, so we judge on the week, not the day.
The studio's craft was fine. Its concepts were guesses, written from a brief and a portfolio, not from what the account had already rewarded. Here the concepts are written by the person who reads the spend allocation, the CPI by concept and the fatigue curve. The edit is the cheap part. Deciding what to edit is where the CPI moves.
Read the full comparison, caveat included →We've worked closely with Growth Therapist, led by Samet, for the past two years as we scaled our first product. He led creative and UA strategy through our soft launch and beyond, and consistently drove fast, high-quality creative experimentation. He's proactive, communicative, and very transparent.
The ads and the campaigns, in one house.
Most creative is briefed across a gap. The people who see the numbers write a document, and the people who make the ads work from the document. Here the same people do both, so the work is faster and it starts from data.
The concepts are written by the person who reads the spend allocation, the CPI by concept and the fatigue curve. The brief is the data, not a document about it.
The person who sees an ad tire is the person who writes its next version. No brief to write, no account manager in between, no waiting for a reply.
Each test ends in a written read, and the next ideas are written from it. Winners become versions, losers are not made again.
Inside a UA retainer this is the default. On the Sprint and the Loop we read your account before writing, and hand the read back with the ads.
Two ways to buy the ads.
- We go through your ad account, your product and the ads your competitors are running
- 3 ad ideas written from what that shows
- 5 versions of each idea, so 15 ads ready to run
- Test versions first. Only the winner gets full production
- Live in your account in about two weeks, then fourteen days of testing set up so the ads get a fair read
- A written summary at the end: what worked, and what to make next
- Credited against the monthly plan if you continue within 60 days
- 2 new ideas a month, 5 versions of each, so 10 new ads
- 5 new versions of the ad that won, so 15 ads a month
- The winner itself finished properly, on top of the 15
- Bigger loops: more ideas and more versions a month, quoted to the account
- Creators on camera as often as the ideas need them, their fees on top
- 14 days notice, no long contract
On top of either price: the ad spend, which runs in your own account, and a creator's fee when an idea needs a person on camera, typically EUR 100 to 300 per creator per Sprint. Other languages, raw footage and project files are quoted on request. On a UA retainer, the ads are made and tested inside the campaigns Samet runs, quoted on a call.
Questions people ask.
Who is this a good fit for?
Apps and games with an ad account already running, or a launch with a product ready to show, and a team that wants the ideas written from its numbers instead of a brief. We start on Meta, with the ads running in your own account next to the ones you already trust. It is not the right fit if you want a single finished ad without versions, because the versions are how the test finds the winner.
What do you need from me to start?
The app or game itself, your audience, and read access to your ad account if you have one. The ideas come out of what is already running. For a new launch with no account yet, they come from the product, the audience and your competitors, and the first test sets the baseline. Alongside that: brand assets (logo, icons, characters, item animations, sound effects, music and an end card if you have one), two or three reference ads, anything the ads must not claim, and for gameplay capture a build that reaches the moments worth filming. If you already have creator footage or clean screen recordings, they go into the Sprint where they fit an idea, at the same price, and the same goes for concepts or hooks of your own that you want tested.
How long until the ads are live?
About two weeks from booking to the fifteen ads running in your account. The fourteen days of testing start after that.
What does the price include, and what is extra?
Both prices cover the ideas, the scripts, the editing and the test. The ad spend is yours and runs in your own account. When an idea needs a person on camera, the creator's fee is quoted on top, typically EUR 100 to 300 per creator for a Sprint, depending on the creator and the market.
What differs between the five versions of an idea?
It depends on the idea. Usually the opening, sometimes the length, the format or the person on camera. We decide per idea from what the account shows, and the brief says which, so the test reads one variable at a time.
Which languages can the ads be in?
English by default. Any other language is quoted per language, with a native speaker reviewing the script and the captions before anything is produced.
What length and format do the ads come in?
Every delivery is 9:16 plus one secondary ratio decided by your placements, as 1080p MP4 with captions burned in. Length and revisions are agreed in the brief for each project, and the brief can override the defaults.
Where do the ads run?
In your own ad account, alongside everything else you have live. That is what makes the result comparable to the ads you already trust.
Who owns the finished ads?
You do. The final files are yours once the work is paid for, and you keep running them whether or not we carry on. The one exception is a creator's face. Their paid ad usage runs for a term that is agreed per creator and written into the quote before production, and it can be extended.
Do I get the raw footage and the project files?
On request, quoted on top. Both the Sprint and the Loop deliver the finished ads.
Do you guarantee a CPI or a ROAS?
No. The test tells you which idea earned the spend, and the comparison on this page shows what that looked like on one account. The result also depends on the product, the offer and the budget behind it, which is why the ads run in your own account, where you read them yourself.
What happens if none of the ideas win?
The summary says so plainly and sets out what the test ruled out, which narrows what is worth making next. A round that kills three ideas is still a result you can act on.
Do you make ads for games or for apps?
Both, and they are briefed differently. A game can put the product on screen in the first second. An app usually has to name the problem before the screen means anything.