Dragon Boss Fight
Tactical board battle: drag one of four heroes (knight, archer, mage, healer) and the round plays out automatically. A dragon intro with an alpha video, four levels and a retry on defeat.
Playable ads developer
Hi, I'm Ilya
I've been building playable ads for 2+ years: mostly 2D with Spine animations and real game logic. Some of my playables scaled to $20k+ in ad spend.
I also work on game design: I plan reskin grids and read retention and CPI to decide what to test next.
- 2+ years
- $20k+ spend on top playables
- 2D + Spine
- Phaser 4
- 10 ad networks
- Analytics inside playables
AppLovin
Unity Ads
ironSource
Mintegral
Meta
Google Ads
TikTok
Pangle
Liftoff (Vungle)
Moloco
Chartboost
Digital Turbine
InMobi
Snapchat
Bigo Ads
What I can do
- Turn a brief into a playable with a clear flow: hook, challenge, win or fail, end card
- Game logic in Phaser 4 with clean modular code, Spine animation, attack and movement VFX that match the game
- Plan a reskin grid: new settings, characters and simplified mechanics from one strong concept
- Read retention and CPI and pick what to retest
- Portrait, landscape and desktop layouts in one build
- Fit builds into network limits (often 5 MB) without visible quality loss: VP9 webm, webp, compressed audio, trimmed engine
- Fast reskins and mechanic reworks from existing gameplay, After Effects and PSD files: UI slicing, video prep
- Add analytics to playables: events for start, interactions, 25/50/75% progress, win or fail, end card and CTA clicks
Tools I use every day
- One-command builds for all 10 ad networks, inline and split, with recompression scripts that keep them under size limits
- PlayHubs, my own playable workspace: many devices at once, logs, pause, live position edits for every screen size
- Size and FPS profiling on real phones
- Playable analytics: AppLovin Playable Analytics funnel events and Luna Analytics engagement and drop-off dashboards
- Phaser 4, Spine, Vite, After Effects, Figma, ArtCraft, Cursor
Playables
Tap to play. Store links are disabled in these demos.
Choose Your Hero
Pick a knight or a mage, learn the moves in a short tutorial, then clear the final level in 3 moves. Only the chosen hero stays in the game, each with its own enemies and finale.
Rockstar Rise
A choice story about getting rich: the hero loses the girl, then taps to play the guitar and earns money with flying notes and bills. Upgrades change his skin, and she comes back in the finale.
Rich or Poor
Opens with a full-screen video, then the hero picks a car, a house and food from cheap or expensive cards in an isometric scene. Every pick changes his reaction and look.
Pump Up Riches
Pump the handle to raise the pressure gauge: the needle passes rewards and the hero gets a car, a house, a helicopter and more. Pressure drops when you stop, so the player keeps pumping to the end.
Win Her Back
Short version of the choice story with character reactions right at the start: the hero loses the girl, taps to repair electronics and earns money. Then he buys a house or a business, changes his outfit, and she returns.
Hero Showcase
A hero carousel: swipe or tap the arrows to switch between three heroes with stats, abilities and a short bio. The animated heroes are alpha videos keyed on the GPU, and only the visible one is decoded.
Metrics I work with
I build playables for results, not just for looks: I track events inside the ad, read CTR, IPM and CPI after launch and change the playable based on the numbers. Happy to discuss your metrics.
| Metric | What it shows | Typical range | Live |
|---|---|---|---|
| CTR | Clicks on the ad per impression | 2.0% – 7.0% | 4.5% |
| IPM | Installs per 1000 impressions, the main creative metric | 5.0 – 25.0 | 15.0 |
| CVR | Installs per click: how well the ad matches the game | 20.0% – 40.0% | 30.0% |
| CPI | Cost per install | $0.30 – $2.50 | $1.40 |
| Engagement | Share of users who interacted with the playable | 60.0% – 85.0% | 72.5% |
| Completion | Share of users who reached the end card | 40.0% – 70.0% | 55.0% |
| Time in ad | Average time spent playing | 20 s – 45 s | 33 s |
| D1 retention | Installed users who came back the next day | 25.0% – 40.0% | 32.5% |
Live example values within typical ranges. Real numbers depend on genre, geo, network and creative.