How many ad creatives does a mobile game need?
Enough to replace every winner before it wears out and, because most new ads don't win, many times more. A planning model, a calculator, and which benchmarks actually exist.
The short answer: enough new ads to replace every winner before it wears out and, because most new ads don't win, many times more than that. How many depends on three numbers from your own account. They are how many ads you keep live, how long a winner lasts, and how often a new ad beats the one it's tested against. This post turns those three into a production plan, with a calculator, and it's honest about which benchmarks exist and which don't.
How much creative the market actually runs#
Before the model, the market. Game advertisers refresh constantly, and the biggest ones produce at a scale no hand-built pipeline reaches.
The last number is the one to hold on to. Spend concentrates on a tiny share of creatives, so most creatives in the market barely run. Volume only helps if it's aimed at finding winners and extending them.
The gap between an ordinary advertiser and the top of the market is large:
| Median app advertiser | 13 |
|---|---|
| Top 25% of app advertisers (53+) | 53 |
| Top gaming advertisers | 192 |
What those creatives are has shifted fast too. In one year, static images fell from 48.9% to 33.0% of gaming ad impressions while video rose to 53.7% and playables doubled to 13.3% (Sensor Tower, 2026).
How fast do game ads wear out?#
Here's the uncomfortable part: nobody has published a game-ad lifespan with a method behind it. The day-counts that circulate for video and static ads on Meta trace back to one page. That page has since withdrawn them, saying they "have no origin". Other headlines, such as SocialPeta's "average ad creative lifespan is now under 5 days" in mature markets, don't define lifespan or show a method. What we can say with sources:
- Sensor Tower's State of Gaming 2026 says that "creative wears out faster" as volume concentrates. That's a direction, not a day count.
- TikTok is the only ad network we found that publishes a day count: upload 3–5 new creatives when fatigue appears, "around every 7 days".
- Aarki reports, from its own internal data, that click-through rate can drop about 45% after four exposures without a refresh.
- In the other direction, Google advises giving new Demand Gen assets at least 14 days before judging them. Replace faster than the learning window and you throw the data away.
So treat lifespan as your number: the days from launch until a winner's install rate or cost per install crosses the line where you'd replace it. It depends on spend, audience size, geo and network, which is exactly why no universal figure holds.
A planning model: slots, lifespan and hit rate#
Forget flat rules of thumb. A game's creative need is set by its account, and it can be worked out.
- Count your ad slots: the distinct ads you want live and spending. The networks give a frame. TikTok recommends 3–5 creatives per ad group and 3–5 ad groups per campaign; Google App campaigns take up to 20 images and 20 videos each; Meta's Advantage+ app campaigns run up to 50 creatives.
- Divide by lifespan. A slot whose winner lasts 14 days needs a new winner about twice a month. Slots × 30 ÷ lifespan is how many ads must win each month.
- Divide by your hit rate. In Motion's 2026 benchmark of 578,750 Meta creatives (cross-vertical, mostly ecommerce; a winner there is an ad spending at least 10× the account's median, and at least $500), about 5% became winners. An older analysis of Facebook app and game tests found only one in twenty new ads beat the control. At 5% you test twenty to replace one.
| Network | What it says about volume | Sources |
|---|---|---|
| TikTok | 3–5 creatives per ad group, 3–5 ad groups per campaign; refresh around every 7 days | Help Center, Creative Accelerator |
| Google App campaigns | Up to 20 images and 20 videos; add assets rather than pulling "Low" ones, and replace gradually | Assets, Optimising assets |
| Meta Advantage+ app | Up to 50 creatives; diversify concepts rather than only swapping CTAs | Advantage+, Diversification |
Meta's rules matter most here: the same Sensor Tower report puts Instagram (35.6%) and Facebook (33.6%) together at about 69% of social gaming ad impressions, well ahead of TikTok (12.7%) and YouTube (12.0%).
Now set the sliders to your own account. The defaults are starting points to move, not benchmarks:
That's about 100 new creatives every week.
slots × 30 ÷ lifespan = replacements; ÷ hit rate = tests. Your inputs only.
The hit rate is the lever you control#
Lifespan and hit rate carry equal weight in the formula: double either and the tests halve. The difference is control. Lifespan is set by your spend, audience and network; the hit rate moves with what you choose to test.
- Winner lasts 7 days
- Winner lasts 14 days
| Period | Winner lasts 7 days | Winner lasts 14 days |
|---|---|---|
| 5% | 1,714 | 857 |
| 10% | 857 | 429 |
| 20% | 429 | 214 |
| 40% | 214 | 107 |
At the ~5% hit rate the benchmarks describe, 20 slots with two-week winners need about 850 tests a month, or about 200 a week. That's the output of the very top of the market. A plan built only on brand-new concepts at a 5% hit rate is out of reach for almost everyone.
Plan slots in concepts, volume in variations#
The way out is to split the work in two. Both 5% figures describe new creatives competing with everything else in the account, and not every new creative has to be a new idea.
- Exploration: new concepts. Distinct ideas, personas and angles. They win rarely, but they find the next winner, and they are what Meta means by creative diversification: "distinctly different pieces of creative", as opposed to two ads that differ only in their call to action.
- Exploitation: controlled variations of a winner. Change the character and keep the hook; change the hook and keep the character; new first frame, new size. Each one is a clean test (one change, read against a control), so you learn which element carried it.
We'd expect variations of a proven winner to beat the control more often than new concepts, because they start from something that already won. No public dataset measures that yet, so track the two separately and measure your own rate for each.
Meta's distinction is a useful guard-rail: a hundred CTA swaps are not a hundred concepts. Count your slots in distinct concepts; spend most of your production on variations of the ones that win.
The catch is production. Built by hand, a variation costs nearly as much as a new ad: find the art, rebuild the layout, re-export every size. Generated as a flat picture, it can't be varied one element at a time, because changing one word redraws the whole image. Variation only gets cheap when the ad is built in pieces, so each element can change on its own.
Where the production time goes#
Most of the hours in a week of creative aren't in the idea. They're in the second through twentieth version of it:
- Resizing: one concept laid out again for feed, Stories and Reels, landscape, MREC and banner placements, each with its own safe zones.
- Recomposing: the character moved clear of the store card, the headline broken onto two lines for 9:16.
- Swapping: a new character, a seasonal background, a localised headline.
- Re-exporting: every change, every size, every network.
That's the gap we built Vidscape for: it turns the art a game already has into layered ads, so a variation is an edit rather than a rebuild. The planning model holds whatever you produce with; the only question is what each variation costs you.
Turn the plan into a routine#
- Set the weekly target from the calculator with your own lifespan and hit rate, and revisit it monthly. Both will move.
- Split it between new concepts (to find the next winner) and variations of current winners (to extend and explain them). Keep the split fixed for a month so results are comparable.
- Name every variation by what changed (character, hook, CTA, size) so results roll up by element as well as by ad.
- Feed the learnings back: an element that keeps winning becomes the default for the next round of concepts.
Measure in your ad manager or your attribution tool; Vidscape makes the creatives and doesn't read campaign results today.
If you run UA, the growth-team page shows how this looks in Vidscape. If you build the variations, start with the page for marketing artists. The broader channel playbook is in how to market a mobile game.
Frequently asked questions#
Is there a standard number of creatives a mobile game needs?
No. It follows from your account: ad slots × (30 ÷ how many days a winner lasts) ÷ the share of new ads that win. For reference, the median app advertiser launches about 13 new creatives a week and the top gaming advertisers about 190 (Singular 2026; AppsFlyer 2026).
How long does a winning game ad last?
No universal figure has a method behind it, so measure it per account: the days until a winner's install rate or cost per install crosses your replacement threshold. TikTok's own guidance is to refresh around every 7 days; Google suggests at least 14 days before judging new Demand Gen assets.
What hit rate should I plan with?
Start from your own history. Motion (2026, cross-vertical Meta data) finds about 5% of creatives become spend winners; a 2020 analysis of Facebook app and game tests found 1 in 20 beat the control. Track variations separately: they change one element of an ad that already won, so we'd expect them to beat the control more often. Measure your own rate for each.
Do variations count as creative diversity on Meta?
Not by themselves. Meta distinguishes iteration (same visual, different CTA) from diversification (distinctly different creative). Use distinct concepts to fill slots and variations to extend and explain the winners.
- State of Gaming 2026 (figures on p. 43 of the PDF) — Sensor Tower · Feb 2026
- State of Gaming 2026 (summary) — Sensor Tower · Feb 2026
- 2026 Global Mobile Gaming Marketing Trends White Paper — SocialPeta × Singular × Aarki · Feb 2026
- State of Gaming App Marketing 2026 — AppsFlyer · Jan 2026
- State of Creative Optimization 2025 — AppsFlyer · Apr 2025
- Creative Benchmarks 2026 — Motion · Apr 2026
- Creative Benchmark Report 2026 — Singular · 2026
- Creative testing: why is the control so hard to beat? — ClickZ (Brian Bowman, Consumer Acquisition) · Mar 2020
- Creative best practices for performance ads — TikTok Business Help Center · Jun 2025
- Creative Accelerator — TikTok Creative Center
- About App campaign assets — Google Ads Help
- App campaign asset optimisation guidance — Google Ads Help
- Demand Gen creative assets refresh guidance — Google Ads Help
- Advantage+ app campaigns — Meta for Business
- Demystifying creative diversification — Meta for Business · Dec 2025
- Insights into global mobile game marketing trends for H1 2025 — SocialPeta · Jul 2025
- Mobile game ad creative strategy 2026 (withdrawn lifespan figures) — Game Growth Advisor · May 2026
