Mobile game ads can put the product on screen in the first second and be understood. So the workhorse is gameplay capture, cut with captions, text and motion graphics, and the whole job is finding the moment that makes the mechanic obvious before the viewer scrolls.
This page is the games side of ad production. It explains why volume comes before polish, what the Horse Racing Solitaire account produced when it was run that way, 175 ads against a studio and a doubled D0 ROAS, and what I do not make. Running the campaigns themselves is a separate service.
How a game brief differs from an app brief
Screen capture leads, but it is not the only format. Each format does a different job for a game than for an app, and the app side has its own page.
| Where it differs | For games | For apps |
|---|---|---|
| The opening | Show the product in the first second; the caption does the rest | Name the problem before the screen, in the person’s own words |
| Gameplay or app capture | Gameplay, usually the moment that makes the mechanic obvious in three seconds | The screen doing the one thing the person downloaded the app to do |
| A creator on camera | A reaction, or someone playing and talking over it | Someone naming the problem in their own words before anything gets shown |
| AI footage and presenters | Cheap enough to test a claim or a hook before anyone books a camera, and rarely the finished ad | A presenter reading a script, to test which promise lands first |
| 2D motion | The mechanic explained in motion before the polished gameplay exists | The problem and the promise drawn out when the screen alone does not carry them |
Volume, then polish
Most game ads lose their test. So I make the first round as cheap test versions and run it in this order. The first four steps are written out in the playbook, with the budget math.
- Setup. One install campaign, one ad set per idea, five test versions in each.
- Launch. A rule pauses every new ad after its first few cents, then all of them restart together, so every ad starts the race at the same moment.
- Read. I leave budgets and ads alone while spend shows which ones pull away, inside the fourteen days of testing. Spend, not IPM, decides which of them won, and CPI is only a sanity check.
- Exit. The ads that pull away, usually one or two per idea, go on to validation.
- Validation. They move into the purchase campaigns in the market you will scale, where Meta decides which ads get budget.
- Polish. Only what wins there gets remade properly and versioned.
What 175 test versions did against a studio
On Horse Racing Solitaire the method held up when an outside studio’s ads ran on the same account. I count the March only row as the fair CPI comparison, because the studio’s ads entered in March. On that row the gap is 11%. Then the studio’s best ads moved into the purchase campaigns, and the chart shows how Meta split the budget there.
| Install test, Dec 2025 to Apr 2026 | Our team | The studio |
|---|---|---|
| Ads tested | 175 | 16 |
| Ads over the £10 spend threshold | 41 | 4 |
| Best ad, CPI | £0.65 | £1.21 |
| Ads under £1.00 per install | 14 of 175 | 0 of 16 |
| Blended CPI, March only, like for like | £1.34 | £1.49 |
What I do not make
3D renders and playable ads. If a test points that way I will say so and point you to a studio that does them.
Who it is for
- Studios in soft launch that need creative volume with a method behind it
- Games whose top ad has fatigued and whose next batch has no test plan
- Publishers comparing an external studio against a testing system on the same account
What you get
- 3 ad ideas written from your ad account, 5 versions of each, 15 ads ready to run
- Gameplay capture cut with captions and motion, a creator or AI footage where the idea calls for it
- Fourteen days of testing in your ad account with a spend rule per ad
- A written summary: what worked, what to make next
Questions people ask
Do you need gameplay footage from us?
Send what you have. The screen capture format is footage taken from inside the product, so the more of the game I can see, the faster the first round is written.
How is the mobile game ad test set up?
As one install campaign where every ad starts the race at the same moment and spend shows which ones pull away; the list above has the steps, and the playbook has the elimination stage with the budget math.
Can game ad production run inside a UA retainer?
Yes, and then the ads are tested inside the campaigns I run, with one owner for the campaigns, the tracking and the ads. That option is quoted on a call, as part of paid UA for mobile games.