Podcast with Zynga’s SVP + Why Misleading Creatives Actually Work

From the Editor

What’s up friend?

This week gave us a full buffet. Market reports, genre shifts, AI debates, and a Zynga SVP telling us why sweating the small stuff still matters when you’re running a thousand people.

Packed issue. Grab your coffee (or something stronger).

---

Sett Featured Content

Unpredictable Hits S2E1: Yair Melamed, SVP Games at Zynga

Yair Melamed - Detail Obsession at Scale

Season 2 is here. And we kicked it off with a heavy hitter.

Yair Melamed started as an engineer at Playtika. Climbed through Product Manager, Director, VP, SVP. Became Chief Games Officer at Wildlife Studios. Now runs multiple studios at Zynga, managing 1,000+ people.

The episode centers on a question that comes up at every level of leadership: do you zoom out or stay in the details?

Yair’s answer is clear. The best executives stay obsessed with details. That granular focus enables better strategic thinking. Detail obsession at scale is a feature.

He also goes deep on optimizing live game portfolios, launching new games against rising CAC, competing with APAC publishers, and the personal decisions that defined his career.

Over an hour of pure gaming leadership. Worth every minute.

Listen now on the Sett Unpredictable Hits Podcast.

Or on Spotify, YouTube, or Apple Podcasts.

---

Misleading Game Ads and Why They Work

Misleading ads in mobile gaming are the most advanced content recommendation system in entertainment today. That’s a big claim. The full article backs it up.

The core insight: ad networks are recommendation engines. Misleading ads work because they change the aesthetic signal your game projects, which changes the audience the algorithm targets, which unlocks entirely new player pools that would never see your game otherwise.

The article introduces a Game DNA framework (six axes to map any game), a creative space model for finding adjacent audiences, and a practical step-by-step exercise for generating new playable concepts. All grounded in how Kingshot’s famous winning ad turned a 4X strategy game into a casual idle hit.

👉 Read the full breakdown right here

---

Industry Buzz

1. AppMagic’s Mobile Market Landscape 2026: The Execution Era

The definitive state-of-the-market report just dropped. The headline: mobile gaming has hit maturity.

The numbers that matter:

• Revenue growth flatlined at +0.2% in 2025 (down from +3% in 2024)
• New game releases surged +25% YoY to 1.4M+ titles
• Only ~10% of new releases gain meaningful traction
• Non-gaming apps surpassed games in IAP revenue for the first time (September 2025)
• 56% of top 100 grossing games use AI in ad creation

Genre winners and losers: Strategy games dominated with +16% revenue, +15% downloads. 4X Strategy specifically saw +31% downloads, +21% revenue. RPG took a hit: -16.6% revenue, -9.1% downloads. Casual revenue stayed flat at ~$21B. Match-3 flat at $4.8B. Merge puzzles grew from 14% to 20% of total puzzle revenue.

LATAM signal: First saturation signs appearing. Colombia downloads -10%, Ecuador -9.8%, Peru -8.9%. But revenue is surging: Argentina +33.2%, Colombia +23.6%. Downloads down, monetization up. The region is maturing fast.

The thesis: “Winners in 2026 will be the publishers who execute better, not those who produce more.” 25% more games launched, but revenue barely moved. Discovery pressure is at an all-time high.

👉 Dig deeper and read more here

---

2. Gossip Harbor: $34M/Month and One-Third of the Merge Genre

Microfun’s Gossip Harbor owns a third of the merge genre. George Tsomaev broke down just how dominant it’s become.

The breakdown:

• $34M/month in revenue
• $550M in 2025 annual revenue
• Owns ~33% of the entire merge genre
• Runs 9,000+ ads weekly but only ~2,800 unique creatives
• Less than 10% of UA budget goes to experiments
• Creative lifespan extends ~6 months (most ads burn out in weeks)

The creative strategy is a machine. One system extracting maximum value from proven winners through detail variation. 9,000 ads weekly, but less than a third are net-new. They found what works and built a machine to milk it systematically.

The broader story: match-3 has stalled. Merge is capturing its audience. Gossip Harbor is the proof point.

👉 Dig deeper and read more here

---

3. Pixel Flow: 3,000+ Creatives/Week and the Clone Economy

Tracy Phan wrote one of the deepest breakdowns of Pixel Flow’s success. And why copying it keeps failing.

The scale:

• Seven-figure daily revenue within three months of launch
• Top-25 US App Store grossing
• Estimated $10-13M/month in IAP revenue (Sensor Tower estimates)
• Reportedly produces 3,000+ creatives weekly (playables + UGC)
• 2-4 clones launch every week. Most fail.

The distinction that matters: behavioral metrics vs. raw metrics. Raw metrics (CPI, retention, playtime) tell you what happened. Behavioral metrics (completion rate, win rate, session depth, FTUE) tell you what will happen. The behavioral layer is where real scalability signals live.

The clone economy proves the point. 2-4 copies launch weekly and die. The mechanic is copyable. The obsessive player psychology underneath it is not.

👉 Dig deeper and read more here

---

4. First Scaler Advantage: Proven Core + Deep Meta x Velocity

Lemon Choi laid out a framework that challenges how Western studios think about speed.

The formula: Proven Core + Deep Meta x Velocity = Market Win.

What it means:

• Proven Core: FPP camera mechanics. Higher player immersion. Lower CPI.
• Deep Meta: Full RPG progression with monetization integration. High LTV.
• Velocity: Ship commercial builds in 4-6 weeks. Lean teams of 5-8 developers.

The split: Western studios spend months perfecting core gameplay. Then months building meta systems. Methodical. Craft-focused. Slow. Eastern studios ship fully commercial titles fast. Rapid iteration on proven mechanics. Aggressive scaling.

Voodoo’s Path of Kings validated FPP RPG core mechanics as a Western prototype. But validation isn’t the same as capturing the market. By the time the First Mover finishes polishing, the First Scaler already owns the category.

The velocity component is the multiplier. Same framework applies to creative strategy. The teams that test, iterate, and scale creative assets fastest win.

👉 Dig deeper and read more here

---

5. Truck Star: $230K/Day on Misleading Ads

Century Games’ Truck Star scaled to $230K/day in under two years. Jakub Remiar broke down how, and it’s uncomfortable.

The most successful new match-3 game outside the Big 4, built on a UA strategy centered on deceptive creative funnels. Parking mini-games in onboarding. Trending idle lumber concepts. None of it reflects the actual gameplay. It’s a match-3 with a male-focused car restoration theme.

Their sister title Tasty Travels runs identical tactics. Company-wide strategy, not a one-off.

The revenue is more than double its predecessor Chrome Valley Customs at peak. Century Games, known for strategy titles, is expanding into casual with the same aggressive playbook.

The numbers don’t lie. But neither does the creative integrity question this raises for UA teams.

👉 Dig deeper and read more here

---

6. Sort Puzzles: From $2M/Year to $40M Per Title

Ahmetcan Demirel broke down subgenre creation for Deconstructor of Fun. The sort puzzle category is the textbook example.

The evolution:

👉 Stagnation (Pre-2023): Water Sort dominated. Basic UI. Ad-heavy. Minimal IAP. 👉 Hybrid Pivot (2023): Block Jam 3D proved IAP scalability. Peaked at $2.5M+ monthly. 👉 Consolidation (2024): Hexa Sort dominated. $4M monthly IAP at peak. 👉 Diversification (2025-2026): Magic Sort hit $40M+ in Year 1. $30M Series A funding.

The mechanic didn’t change. The packaging, polish, and monetization strategy did. Magic Sort took a basic Water Sort mechanic, applied Royal Match-level production values, and built a $40M/year business.

The 2026 prediction: conveyor-based mechanics (like Pixel Flow) are the emerging dominant archetype. Artwork-centered levels enable endless creative variation for UA.

👉 Dig deeper and read more here

---

7. Gamigion Launches Deconstructions: 300 Levels Deep

New recurring section worth tracking. Gamigion’s Yagiz Gur is systematically deconstructing hybrid puzzle games by playing through 300 levels per game. Raw data shared as Google Sheets.

The first two: Pixel Flow (free) and Arrows (paid subscribers).

The core thesis: “Players come to play levels, not chase endless mini events.” Most analysis focuses on monetization and meta layers. This goes the opposite direction, studying the core loop and level design as the real retention driver for games with 10%+ D30 retention.

Nobody else is doing this level of systematic, data-driven level design analysis. For UA teams, understanding which level design patterns correlate with strong D30 retention changes how you evaluate the games you’re acquiring users for.

👉 Dig deeper and read more here

---

AI & Games: The Big Debate

Five different sources this week tackled the same question from different angles. The consensus is clearer than the headlines suggest.

Supercell CEO: AI Is the Opportunity, Not the Threat

Supercell Embracing AI

Ilkka Paananen’s message is direct. While Google’s Project Genie demo tanked gaming stocks 10-12%, Supercell is betting hard on AI as an accelerant.

The numbers: EUR 2.65B revenue in 2025 (down 4% YoY). But EBITDA hit EUR 932M (up 6%). Profitability improved. Headcount grew 30% while the rest of the industry laid people off. New game development investment doubled last year. Doubling again this year.

The quality bar: Squad Busters crossed $100M in revenue. They killed it anyway. Didn’t meet their standard.

Paananen’s real concern is that the industry is coasting on optimizing what already exists rather than innovating. AI solves the production bottleneck, freeing teams to focus on what actually matters.

👉 Dig deeper and read more here

Take-Two CEO: “Project Genie Is a Great Sign”

Take-Two Tools vs Entertainment

Strauss Zelnick had the opposite reaction to everyone else when Google showed Project Genie. His framing cuts clean: “AI tools are not entertainment experiences. Creators use tools to create great entertainment.”

Take-Two reported $1.76B in net bookings. GTA Online revenue grew 27%. NBA 2K jumped 30%. Zynga mobile grew 19%. Hundreds of AI pilots running across the company.

Zelnick is “even more optimistic than I’ve ever been” about AI’s role. The market saw a threat. He saw validation.

👉 Dig deeper and read more here

SciPlay Art Director: “AI Never Produces Final Assets”

Peter Franco’s practitioner perspective from inside SciPlay. AI applications span concept exploration, logo development, asset creation, and progressive visual states. Custom AI models with bi-weekly capability updates.

But the key insight: AI never produces game-ready final assets. Human intervention is always required. Photo-bashing, composition, color correction, technical optimization.

AI gives you “an army of artists in a box.” Someone still needs to know what to point them at. The shift is from production to creative direction. Traditional drawing, painting, and composition skills become competitive differentiators.

When asked about AI’s trajectory, his answer was one word. “Terrifying.”

👉 Dig deeper and read more here

NetEase: AI as Core Competence

NetEase reported 27.5 billion yuan (~$3.9B) in quarterly game revenue, up 3%. CEO William Ding calls AI a “core competence for development and operations.”

The difference: this isn’t experimentation. NetEase has integrated AI across the full pipeline. Graphics, design, programming, animation, QA testing. Every stage of the development lifecycle.

When a $3.9B/quarter publisher calls AI a core competency, smaller studios are running out of time to catch up.

👉 Dig deeper and read more here

Games Built AI’s Playbook

Ken Noland (AI Guys CTO) at PGC London flipped the standard narrative. The “AI is coming for games” framing misses the point. Games came first.

The four primitives of generative AI: state, action, memory, feedback. Every single one has been “prevalent in games for ages.” Pac-Man’s ghost behavior demonstrated state management. Chess showcased forward-thinking anticipation. RTS pathfinding maps directly to modern AI search techniques.

“Every single token that’s generated is searching through a massive data set, running through various algorithms and generating the next token.” The same real-time feedback evaluation that game AI has used for decades.

Games didn’t just inspire AI. They built the playbook.

👉 Dig deeper and read more here

---

Also Worth Reading

Mobile Gaming Today: Feb 5 Roundup - Liftoff targeting $5.17B IPO. Reports of Meta iOS revenue surging 5-7x overnight. ChatGPT Ads at $60 CPM with $200K minimum spend. AppLovin’s Axon hit $11B annual advertiser spend. Vietnamese gaming scene deep-dive covering 12+ studios.

👉 Dig deeper and read more here

New Mobile Games: Jan 30 - Feb 5 - 4,022 games launched in one week. Only 30 worth watching. Moon Active’s merge-2 “Reality Bay” and SayGames’ return to puzzle sorting with “Screw World 3D” signal where the smart money is going.

👉 Dig deeper and read more here

---

Wrap Up

The theme this week is loud and clear. Execution wins. Speed wins. Systems win.

The market is saturated. Revenue growth is flat. And the teams pulling ahead are the ones who found what works and built machines to extract every drop of value from it. Gossip Harbor’s 9,000 ads a week. Eastern studios shipping commercial builds in 4-6 weeks. NetEase wiring AI into every stage of development.

The playbook is converging. Build fast, test everything, keep what works, and systematize the hell out of it.

See you next time.

Yours, Jonathan (lover of my daughter, games, and pizza, in this order) Fishman

---