Motion graphics and 3D ads, built around the game on screen.

What motion graphics and 3D ads are

A motion graphics ad is a video built from animated design rather than filmed footage: type, interface, icons, illustrated characters and game art, moved and timed to a script. A 3D or CGI ad is built from modelled, lit and rendered scenes, so a character, a creature or a world can be shown from any angle, in any light, doing things the camera could never capture. Most of the motion and 3D work we make for games mixes the two with real captured gameplay.

The difference from UGC is simple. A creator sells a reaction to the game. Motion and 3D sell the game itself: the board, the system, the hero, the world.

When motion and 3D beat live action or UGC

Our view is that motion and 3D win when the thing the player needs to understand lives inside the game, and a face on camera would only get in the way. In practice that means five jobs:

The job	Why motion or 3D does it better
Showing a system	A rule (match, merge, outrun, collect) can be slowed, labelled and replayed in a way a screen recording cannot.
Showing progression	A level or a base can go from empty to finished in ten seconds.
Selling a world	Camera moves across a map or an island give scale that UI capture flattens.
Using IP and characters	Licensed creatures and heroes can be rendered at the quality the fans expect.
Explaining a mode	Kinetic labels and arrows can teach a new mode in the time a creator spends saying hello.

UGC is still the better tool when the hook is a person: a reaction, a confession, a challenge between two players. We would run both, not pick a side.

One caution from Google is worth taking seriously. Its help article on gameplay videos separates "gameplay videos", which are "actual gameplay content that help demonstrate what playing the game is like", from "promotional videos", which are "cinematic trailers and skits that show off the game's storyline, characters, or features", and says "most mobile game players are more likely to install a game after watching gameplay videos compared to promotional videos." Google's guidance on App campaign creative also says videos "should get to the app experience as quickly as possible." We read that as a rule for this format: the render is the wrapper, and the game has to be on screen early. A beautiful cinematic that never shows the board is a trailer, not a user acquisition ad.

Formats and specs per network

Specs change, so check each network's page before a batch goes out. As published when we wrote this:

Network	What the network's own page says
AppLovin (apps)	Video "9:16 (Portrait)", MP4 or MOV, maximum file size 1 GB, maximum duration 60 s. AppLovin reprocesses every video into a standard format.
Instagram Reels (Meta)	Ratio 9:16, recommended resolution 1440 x 2560, duration 0 seconds to 15 minutes. Meta suggests leaving at least 14% of the top, 35% of the bottom and 6% of each side free of text, logos or other important elements.
Facebook Feed (Meta)	Recommended ratio 4:5 at 1440 x 1800, duration 1 second to 241 minutes, maximum 4 GB.
TikTok in-feed (non-Spark)	Vertical 9:16 recommended at 540 x 960 or more; 16:9 and 1:1 also accepted; up to 10 minutes; 500 MB or less.
Google App campaigns	Google encourages uploading up to the maximum of 20 videos, and says successful app video ads in its research "on average" contain "2+ cut changes within the first 5 seconds".

Two consequences. Design for 9:16 and treat other shapes as reframes, keeping the game and the type inside Meta's safe zone from the start. And on AppLovin the video is only part of the ad: the playable or end card that follows is a separate asset, and our playable ads and end cards pages cover those.

What good looks like

These pieces from our showcase show the five jobs above. We describe what each ad shows, not how it performed.

• A mode, explained. SEGA's Sonic Rumble "How to Rumble" moves through the game's courses, the player lobby and a running player count, and ends on the characters and a Play Now button. It teaches a mode in under a minute.
• A system under pressure. Kitchen Masters "Tomato Tsunami" puts a 3D chef in a tank filling with tomatoes above the puzzle board, so every move on the board has a visible consequence, before a fail screen and the level map.
• A world, split from the play. Game of Thrones: Dragonfire "Sheepstealer Outrun" runs a rendered dragon in the top half and top-down flight over the sea in the bottom half under the line "Can you escape Sheepstealer?". The IP gets its spectacle and the game still gets the screen.
• Character drama into the board. Empires and Puzzles "Animated Board" opens on a 3D heroine choosing between "Hold" and "Fight", then cuts to an isometric rendered battle with the match board in the middle.
• A story with a world at the end. Paradise Paws "Island Awakens" is illustrated 2D motion with dialogue boxes, a hut restored from ruin and a closing shot of the island and its animals.
• A live opener into the game. Dragon City "DC Party Reveal" starts as a filmed party reveal and turns into animated eggs hatching, feeding, the world map and the dragon roster.

How Flixr makes it

We make video, creators and UGC, motion, playables, end cards, statics and the landing page from the same brief, with generative AI where it earns its place and cast actors where trust has to land. That matters for motion and 3D in one specific way: the rendered hero shot, the playable and the end card should carry the same promise, and they only do if one team briefs them together.

The work runs on a weekly loop: brief, make, launch, read, log. It starts with read-only access to your results and ad accounts, so we are judged on the same numbers as your team, and a first month spent reading 90 days of your ads one by one before the first three hypotheses go live. Flixr Studio, our own tool, tracks every ad running in a category on AppLovin, Meta and TikTok, how long each one lasted and what replaced it; it has read over 40,000 ads across 195 markets. For motion and 3D that is where the brief starts: which rendered concepts in your genre the networks keep funding, and which ones they dropped. Our genre teardowns show what that reading looks like.

Our production habit for this format is to build for variation. A render is expensive to make once and cheap to recut, so we plan the hero shots to be reused across hooks, lengths and aspect ratios, and change the opening seconds far more often than the scene.

How to test it

Our view: test motion and 3D in the same campaigns as your UGC and live action, not in a separate "brand" line, and judge them on the same result. For a game that means installs and payers in your MMP, read inside your own ad accounts, rather than view rate or watch time.

Use what each network gives you. Google's asset report rates each asset as "Waiting", "Learning", "Low", "Good" or "Best", and Google says those ratings "are relative to the other assets in your campaign"; it recommends adding more assets "instead of removing 'Good' or 'Low' rated assets." TikTok's split testing keeps other variables the same, gives each audience group only one ad group, and is designed to reach "a 90% confidence rate" before it names a winner.

Within the format, test one thing at a time: the first three seconds, gameplay early against gameplay late, 2D motion against a 3D render of the same idea. For a deeper method see how to know which part of an ad drives installs and UGC or produced.