Research · November 2026 edition · updated monthly

State of mobile game ads

12,339 ads from 23 game brands, launched between 16 October 2024 and 8 November 2026 (11,893, or 96.4%, since July 2026), watched by Flixr Studio since 2 October 2026. What game advertisers are running, read ad by ad. Data as of 9 November 2026.

1. Video dominates: 84.0% of 12,311 read game ads are video, and gameplay-led ads are the largest single type at 34.9%.
2. Fake gameplay is common but concentrated: it appears in 30.8% of read ads and at 21 of 23 brands, yet one brand accounts for 44.5% of those ads, and in social casino it is only 4.4%.
3. Ads open on people more than on play: a character or emotion hook leads 29.6% of read ads, against 23.5% for a plain gameplay opener.
4. Most new ads are short-lived: of 3,732 ads Studio saw launch between 2 and 20 October, 38.6% were still running at day 7 and 23.1% at day 14.
5. Shared ideas are still rare: 38 of 351 creative concepts run at two or more brands, and 28 of those 38 are in social casino.

What we watched

12,339 public ads from 23 game brands, almost all launched since July 2026.

• 12,339

  Ads tracked

  of all public Ad Library ads of game brands

• 23

  Brands tracked

  of brands with ads in the set

• 12,311

  Ads read by the model

  of 12,339 ads. 99.8% have a video or image reading

• 96.4%

  Launched since 1 July 2026

  of 12,339 ads. 11,893 ads; launch dates run 16 Oct 2024 to 8 Nov 2026

• 68.7%

  Launched in October 2026

  of 12,339 ads. 8,475 ads

Video, static and how ads are made

Five in six game ads are video, and about a third are carried by gameplay.

• 84.0%

  Video

  of read ads (12,311 with a video or image reading). 10,340 ads

• 15.6%

  Static image

  of read ads (12,311 with a video or image reading). 1,923 ads

• 19.0%

  Dynamic creative (DCO) placements

  of 12,339 ads. 2,349 ads, as delivered by Meta

• 0

  Playable ads

  of 12,339 ads. The Meta Ad Library set holds no playables

Real vs fake gameplay

Nearly a third of read ads show fake gameplay, and 21 of 23 brands ran at least one.

• 44.5%

  Real core gameplay

  of read ads (12,311 with a video or image reading). 5,480 ads

• 30.8%

  Fake gameplay

  of read ads (12,311 with a video or image reading). 3,794 ads

• 24.6%

  No gameplay

  of read ads (12,311 with a video or image reading). 3,031 ads

• 21

  Brands with at least one fake gameplay ad

  of 23 brands

• 44.5%

  Share of fake gameplay ads from the single largest brand

  of 3,794 fake gameplay ads. 1,689 ads; fake gameplay is concentrated

• 40.0%

  Fake gameplay showing a hero running to build

  of 3,790 fake gameplay ads with an ad-level reading. 1,516 ads

What the video is cut from

Stylised gameplay fills more screen time than raw capture by about three to one.

• 42.5%

  Stylised or recreated gameplay

  of all seconds of read video footage. In 49.6% of 12,311 read ads

• 13.2%

  Raw gameplay capture

  of all seconds of read video footage. In 20.5% of 12,311 read ads

• 8.9%

  AI-generated or AI-enhanced footage

  of all seconds of read video footage. In 1,248 of 12,311 read ads (10.1%)

• 2.2%

  Creator or UGC footage

  of all seconds of read video footage. In 5.9% of 12,311 read ads

UGC and creators

Almost one in three read ads uses a creator, a skit or real people.

• 31.5%

  Any UGC or creator concept

  of read ads (12,311 with a video or image reading). 3,873 ads

• 14.0%

  Scripted skit story

  of read ads (12,311 with a video or image reading). 1,725 ads

• 8.5%

  Celebrity or pro endorsement

  of read ads (12,311 with a video or image reading). 1,050 ads

• 3.8%

  Streamer reaction

  of read ads (12,311 with a video or image reading). 463 ads

• 2.6%

  Review to camera

  of read ads (12,311 with a video or image reading). 325 ads

The first three seconds

Characters and emotion open more ads than anything else, ahead of plain gameplay.

• 29.6%

  Character or emotion

  of read ads (12,311 with a video or image reading). 3,645 ads

• 23.5%

  Plain gameplay opener

  of read ads (12,311 with a video or image reading). 2,893 ads

• 10.6%

  Reward or offer

  of read ads (12,311 with a video or image reading). 1,300 ads

• 9.2%

  Text question or claim

  of read ads (12,311 with a video or image reading). 1,138 ads

Why the ad says you should play

Luck, power and finishing things are the most common primary pitches.

• 22.8%

  Luck and jackpot chase

  of read ads (12,311 with a video or image reading). 2,808 ads; an ad can have one or two primary motivations

• 13.6%

  Power (growing stronger)

  of read ads (12,311 with a video or image reading). 1,676 ads

• 11.7%

  Completion

  of read ads (12,311 with a video or image reading). 1,441 ads

• 8.0%

  Status

  of read ads (12,311 with a video or image reading). 987 ads

Offers and who the ad is for

A third of game ads name an offer, and one in eight speaks to lapsed players.

• 34.2%

  Any offer (video, title or copy)

  of ads with an ad-level reading (12,331). 4,221 ads

• 32.0%

  Offer in the video or title

  of ads with an ad-level reading (12,331). 3,940 ads

• 2.3%

  Offer only in the ad copy

  of ads with an ad-level reading (12,331). 281 ads

• 24.3%

  Free rewards or gifts

  of ads with an ad-level reading (12,331). 3,000 ads

• 5.7%

  Limited-time event

  of ads with an ad-level reading (12,331). 699 ads

• 15.5%

  Coins named

  of ads with an ad-level reading (12,331). 1,915 ads; amounts are each game's own currency and are not compared

• 12.6%

  Win-back (addresses lapsed players)

  of ads with an ad-level reading (12,331). 1,554 ads

• 85.1%

  Acquisition

  of ads with an ad-level reading (12,331). 10,488 ads

Messaging devices

The big win celebration is the most common device, in more than one in five read ads.

• 22.6%

  Big win celebration

  of read ads (12,311 with a video or image reading). 2,780 ads

• 16.3%

  Social proof

  of read ads (12,311 with a video or image reading). 2,012 ads

• 10.9%

  Tutorial or how to

  of read ads (12,311 with a video or image reading). 1,345 ads

• 7.8%

  Dare or challenge to the viewer

  of read ads (12,311 with a video or image reading). 962 ads

• 6.7%

  Apology or come-back letter

  of read ads (12,311 with a video or image reading). 822 ads

Social casino vs the rest

Social casino ads lean on offers and statics and rarely show fake gameplay.

• 82.2%

  Social casino ads with any offer

  of 3,532 social casino ads with an ad-level reading. Vs 34.2% across all 12,331

• 4.4%

  Social casino fake gameplay

  of 3,538 read social casino ads. Vs 30.8% across all 12,311

• 27.8%

  Social casino static images or slideshows

  of 3,538 read social casino ads. Vs 16.0% across all 12,311

• 20.3%

  Social casino UGC or creator concept

  of 3,538 read social casino ads. Vs 31.5% across all 12,311

How long new ads keep running

Of ads launched 2 to 20 October, fewer than four in ten were still running a week later.

• 38.6%

  Still running at day 7

  of 3,732 ads launched 2 to 20 Oct 2026 and seen within a day of launch. 1,440 ads

• 23.1%

  Still running at day 14

  of 3,732 ads launched 2 to 20 Oct 2026 and seen within a day of launch. 861 ads

• 0.7% to 7.4%

  Still running at day 30

  of 907 to 973 ads launched 2 to 10 Oct 2026 (the only ones 30 days old by 9 Nov). 6 of 907 with a known outcome; 66 more are still flagged live but were last seen on 3 Nov. If all of those reached day 30, 72 of 973

Ideas that travel between brands

Studio found 351 shared creative ideas; 38 of them run at more than one brand.

• 351

  Concepts found

  of concepts across 23 brands

• 10.8%

  Run by two or more brands

  of 351 concepts. 38 concepts

• 5

  Run by three or more brands

  of 351 concepts. The widest is run by 6 brands

• 137

  First launched since 2 Oct 2026

  of 351 concepts. 7 of these already run at two or more brands

• 61.6%

  Ads that belong to a concept

  of 12,339 ads. 7,599 ads

• 14

  Brands that launched a shared idea after another brand

  of 23 brands. 13 brands launched at least one shared idea first

• 28

  Multi-brand concepts in social casino

  of 38 multi-brand concepts

• 36

  Median days before a second brand launched the same idea

  of 36 multi-brand concepts with a dated second brand. Range 1 to 637 days; dates are the Ad Library's reported starts

How we counted

• Only public Meta Ad Library ads are counted (visibility public, source ad_library). No client performance data, spend or results are used.
• Shares are out of read ads: video and image fields (format, ad type, gameplay, hook, UGC, motivations, footage, devices) out of 12,311 file-read ads; offers and lifecycle out of 12,331 ad-read ads.
• Every label is a model's reading of the ad (Gemini, taxonomy v2), checked against a second model's blind labels: lifecycle and offers agree about 95 to 100%, fake vs real gameplay about 95%, mechanics and primary motivation about 80%, format about 70%. On social casino ads, offer amounts about 83%, footage sources about 70%, messaging devices about 75%. Read small differences with that in mind.
• Fake gameplay is the industry term for gameplay that is not the game's core loop as a new player meets it (invented puzzles, minigames). Staged or simplified real gameplay counts as real.
• Competitors have no spend data here. How long an ad runs means a brand keeps paying for it, never that it performs.
• Run length uses a cohort Studio saw from launch: ads launched 2 to 20 October 2026 and first seen within one day of launch (3,732 of 4,542). Each day-N share counts only ads old enough to reach day N by 9 November. Older ads are left out because only the ones still running were there to be seen.
• Genre is set per brand. Only social casino (8 brands) is reported as a genre; the other genres hold one or two brands each, and 3,838 ads from 10 brands carry no genre label.
• Concepts group ads that play the same creative idea across brands, named and confirmed by a model from public ads, updated daily. Start dates are what the Ad Library reports. A later launch is not proof of copying.
• Some brands were last collected on 3 November, so their live flags are as of that date.

Want this read for your own category?

See Flixr Studio