VGC 2026 · Scale Track
Scaling
What global studios do differently
Khuc Moc Thao · UA & Monetization Summit
Hands up — who shipped a playable in the last 90 days? Keep it up if it was a top-3 creative. That gap is the talk. Most studios MAKE playables; few SCALE them.
The networks are voting with their spend
The state of playables
+49%
a top advertiser's IPM lift from playables— Moloco
30 sec
avg time a user spends in a playable (vs 5–10s for video)— Liftoff 2025
30% → 56%
of the top-30 highest-spend creatives, in 1 year— AppLovin (via Sett)
8–16×
the install rate of non-playable ads— Liftoff 2025
And adding UGC lifts impression-to-install +152% on average — same format, compounding leverage. (Liftoff 2025)
Lifts measured vs the same campaigns' non-playable (video / static) creative. IPM = installs per 1,000 ad impressions.
This isn't theory — it's the networks' own numbers. Moloco saw a top advertiser get a 49 percent IPM lift from playables. Users spend 30-plus seconds inside a playable versus five for a video. AppLovin: playables went from 30 to 56 percent of the top creatives in a single year. Liftoff: 8 to 16 times the install rate of non-playables — and adding UGC lifts conversion another 152 percent. The format won the argument; the only question is who'll be good at it.
And the quality holds
Playables lift the downstream numbers too
Day-7 retentionLiftoff +25% Day-0 ARPUUnity / ironSource +30.5% Day-14 ARPUUnity / ironSource +15.2% ROASMoloco — top advertiser +25% Day-14 ROASUnity / ironSource +5.4%
Each bar is a different metric's % lift vs non-playable creative — read the labels, not the bar lengths. ARPU = revenue per user · ROAS = return on ad spend · Day-0/7/14 = days after install. iOS data.
Slide 2 showed playables get more installs. This shows the installs are BETTER. Day-7 retention up 25 percent. Day-0 revenue per user up 30 — and still up 15 at Day-14, so the lift sticks. ROAS up 25. Even 14-day ROAS — the single hardest number to move — is up. You're not buying a spike — you're buying users who stay and pay. (ironSource figures are iOS.)
Why it works
The creative is the
Networks automated bidding & targeting — your old edge moved into the creative.
Privacy (ATT · SKAN · Privacy Sandbox) killed the signal — the creative now carries the weight.
A playable doesn't describe your game.is your game.
When the algorithm decides WHO sees the ad, the creative decides whether they're the RIGHT person. No format qualifies a user like a playable — they try before they install, so they self-select. Fewer installs, better ones. Structural advantage, resilient to the signal loss breaking everyone's UA.
The real gap
Making one is easy.
One playable: brief, build, ship, decent number. Easy.
Every week · dozens of variants · measured properly · feeding back into UA and the game — and getting better over time.
That's not a creative task. It's a system.
First — how big has this already gotten?
If playables are this good, why did so few hands stay up? Because making one is easy; doing it as a repeatable, measured, self-improving system is the hard part. The winners treat it like a system. But first, let me show you how big this has already gotten.
Top-creative share is exploding
Playables took over the top of the charts
Share of the top-30 creatives by spend that are playables — nearly doubled in a year.
50×
growth in playables produced per month, over 24 months≈1,000/mo (Oct '23) → 50,000–60,000/mo (Oct '25)
And at the top: the top-20 grossing games launch a median of 64 new playables/month — leaders hit 1 playable per 16 videos.
Source —
sett.ai "State of Playables Q4-25" (AppLovin / AppMagic data)
A year ago, 30 percent of the top-30 highest-spend creatives were playables. Today it's 56 — more than half. Production grew 50 times in two years, and the top grossing games now launch a median of 64 new playables a month. The biggest spenders moved their budgets here. If you're not building playables at volume, you're competing against people who are. So — here's what the winners actually do differently. Seven things.
1
A product,
Average studio: the playable is the last thing that happens — handed off, shipped, done.
Global studio: a product with a roadmap, versions, a backlog of hypotheses — always thinking 3 iterations ahead.
A deliverable is finished.
The moment you treat the playable as a product, you stop asking "is it done" and start asking "what are we learning, what do we test next." That question is the engine of scale.
2
Engineer the
Almost everyone drops off in the opening. Your funnel is decided before the user reaches your "real" gameplay.
Don't design the hook — engineer it. Test 10–15 different openings, same game, let data pick.
Amateurs explain the game. Experts give the user something they can't resist doing.
📌 If you write down one thing — write this.
Users give a playable 30 seconds — but only if the first 3 earn it. A strong hook = a near-win they must finish, a "wait, I can fix that" moment. Rebuild only the first 3 seconds, ten ways, test them — your biggest lever is usually right there.
3
Velocity beats polish
Concept velocity — many different ideas/week. Kill losers without sentiment, scale winners. Refresh cadence > batch size.
Delivery speed — idea → live in days, not weeks. The studio that completes more loops wins.
Always-on. A Friday-night idea is live Monday.
Two dimensions of velocity. Baseline: 10–15 creatives per ad group, constantly refreshed. A brilliant playable in 3 weeks loses to a good-enough one shipped in 3 days now on its 4th iteration. If production stops when the office lights go off, your velocity has a ceiling — and competitors know where it is.
4
The engine room:
They're not making these by hand. They run a platform.
Building the engine is harder than making the playables — which is why the smartest studios run on one that already exists.
"That's impossible, we can barely finish one." Right — by hand. The secret: a platform. Variants are cheap, so testing is cheap, so learning is cheap. In 2026 it's cheaper to GENERATE a localized variant than translate one. Months of platform work before a single ad ships.
5
Honest hooks
$0.25→$0.05
lifetime value (LTV) per user — honest ad vs misleading ad (Segwise)
>30%
of all playable impressions are misleading (Segwise)
58%
of users unlikely to try a playable again after a misleading one (Segwise)
The downstream cost: Day-1 retention 32% vs 14% , Day-7 11% vs 1.5% — honest vs misleading. You buy a spike, not a business.
Misleading ads also violate Google Play's Deceptive Behavior policy — risking app suspension or removal.
Temptation to mislead is everywhere — over 30 percent of impressions (Segwise). Short term it posts great numbers. But tricked users churn day one; lifetime value craters from 25 cents to 5; honest ads hold Day-7 retention at 11 percent versus one-and-a-half for fakes. 58 percent of users are unlikely to try a playable again after a bad one. And misleading ads also violate Google Play's deceptive-behavior policy, which can suspend or pull your app. Being honest isn't an ethics lecture, it's a performance strategy. The honest hook is the one still there on Day 7.
6
The feedback loop
A playable measures every tap, every drop-off, every choice. That's not just creative data — it's product data.
Playable → data → game design → playable
When a mechanic out-converts, that's the market telling you something about your game , not just your ad.
The highest-leverage difference — and almost nobody does it.
Most studios think one direction: game first, playable advertises it. The best run a loop. You're market-testing your mechanics with real money and millions of impressions. The mechanic that hooks in the ad often becomes the one they lead with in the game. This is the engine behind the UA–Product Feedback Loop.
7
Networks & devices:
File caps differ — 5 MB (Meta/Google/AppLovin) vs 10 MB (Unity/ironSource). Miss the wrong one → rebuild per network.
Channels aren't equal — AppLovin ≈ 59% (est.) of playable traffic; Mintegral strong on Android; TikTok wants a different shape.
Then × thousands of devices — a 60-fps flagship build can crawl on the $60 Android that is most of your installs here.
This knowledge doesn't close with effort. It closes with volume.
A living catalog — this network rejects that, this device chokes on this — built only by shipping across everything, over and over. You won't see device problems in the office; you'll see them in retention, after you paid for installs. A studio shipping a few/year never accumulates it; a team shipping daily has it as muscle memory. (The 59% AppLovin share is an industry estimate, not AppLovin's own number.)
Where the installs are
IPM by genre — 2025
IPM = installs per 1,000 ad impressions. Higher = the format converts views to installs faster.
IPM swings 3–4× by genre — but every genre benefits. The lighter the genre, the more a great hook moves the needle; the hook just changes shape.
Source —
adjust.com Gaming App Insights 2025 (IPM = installs per 1,000 impressions, by genre)
IPM varies hugely by genre — racing and hypercasual convert at 3 to 4 times what RPG does. The lighter the genre, the more a great hook moves the needle, but every genre benefits. Wherever you sit on this chart, the playable is your lever — the hook just changes shape. An RPG hook is a power moment, not a one-tap loop, but it still works.
The steering wheel
Measure the install,
Installs (IPM) × Retention × ROAS— measured per creative, not per campaign · ROAS = return on ad spend
Installs alone (IPM) tell you the playable gets users. They say nothing about whether those users were worth getting.
Concentration is extreme — the top 2% of creatives drive 53% of all gaming ad spend (AppsFlyer). Measure per creative, or you're flying blind.
Measure at the creative level → you know which hook to scale, which mechanic to feed back, and killing a loser becomes obvious instead of political.
Source —
appsflyer.com Creative Optimization 2025 (1.1M creatives)
Average studios stop at IPM — easy number, wrong place to stop (misleading playables win on IPM, lose on everything else). And concentration is brutal: the top 2 percent of creatives drive 53 percent of all gaming spend, so a few winners carry everything — you only find them at the creative level. The winners can say "THIS specific playable brought users who retained and paid." Measurement isn't the reporting step at the end — it's the steering wheel.
The honest math
You don't need to build the factory.
Free Monday moves: rebuild your first 3 seconds 10 ways · tag installs by creative, watch Day-7 retention · velocity target not quality target · treat the playable as a window into your game.
But the engine — tooling, always-on production, cross-network/device knowledge, measurement — is a full-time craft ($5k–$80k and weeks per custom playable). Rent it and you still keep the data: a good partner ships you instrumented creatives and creative-level results, so the read on your own game stays in-house.
Own the learning. Rent the hands.
Cost range — industry production benchmarks (
coinis.com )
Starter moves need a mindset change, not budget. But be honest about the rest: building the engine in-house = artists + engineers + a data person + months of tooling. For most studios that math doesn't work. The quiet thing the global players figured out: keep strategy, data, and the feedback loop in-house; rent the production engine already running at full velocity. The real question isn't "should we make playables" — it's "become a playable studio, or work with one."
The next five years
Stop making playables.scaling them.
The format that converts best, qualifies best, and survives privacy best — treated as one of the most important products you build.
The gap isn't budget. It's mindset — and you can close it this afternoon.
One idea to leave them with. The winners don't bolt the playable on at the end — they build it as a core product with a roadmap, a platform, and a feedback loop into the game. The gap is a mindset gap; close it today. [pause for applause] Happy to take questions. Optional CTV bonus close if time allows.
Appendix · Stats & sources
Key figures
Metric Value Source IPM / ROAS lift (one top advertiser) +49% / +25% Moloco Interactive end cards — IPM / conversion +7–14% / +30% Moloco Playables — D0 ARPU / D14 ROAS lift (iOS) +30.5% / +5.4% Unity / ironSource Playables share of top-30 creatives 30%→56% (1 yr) AppLovin / Sett Playable install rate vs non-playable 8–16× Liftoff 2025 Day-7 retention uplift +25% Liftoff 2025 Creative concentration (top 2% of spend) 53% AppsFlyer 2025 Global avg IPM (2025) 8.62 Adjust 2025 Misleading share / LTV collapse >30% / $0.25→$0.05 Segwise
IPM = installs per 1,000 impressions · ROAS = return on ad spend · ARPU = revenue per user · D0/D7/D14 = days after install. Full confidence tiers + source URLs in stats-pack-2025-2026.md.
Reference slide — skip in live delivery unless asked for sources. The clickable source list is for the public web version so the audience can verify every figure themselves.