Product Craft

How products are built is itself becoming policy: the EU Digital Fairness Act (expected mid-2026 to Q4 2026) names addictive design, dark patterns and loot boxes; India has banned all real-money games; the US is fighting over who performs age verification at the state level; Japan leans on self-regulation; and child-safety-by-design is going global.

1. Engagement maximization is now a policy issue: the EU Digital Fairness Act sits on the Commission's 2026 work programme (Q4 2026), explicitly targeting dark patterns, addictive design and unfair personalization — turning design assumptions into a regulated category.[10]
2. Some observers now expect the EU DFA proposal as early as mid-2026: President von der Leyen has named addictive design, dark patterns, influencer marketing and online profiling, with infinite scroll, autoplay, reward loops and loot boxes listed as targets, and ~70% of consultation respondents backed binding rules on in-game spending — putting design assumptions squarely on the political table.[32][33]
3. 🇮🇳 Monetization design can itself be outlawed overnight: India's Promotion and Regulation of Online Gaming Act 2025 (passed Aug 22, 2025; rules notified Apr 22, 2026; in force May 1, 2026) imposes a blanket ban on all online real-money games — skill or chance — hitting a ~$3.7B market (real-money ~86% of it, ~$3.2B) and showing how fast a state can criminalize a business model.[21]
4. 🇺🇸 In the US, who performs age verification has become a political fight: starting with Utah's App Store Accountability Act (SB 142), several states now require app stores to verify ages and tie minors' accounts to a parent who must approve downloads and purchases — shifting compliance burden from platforms onto app developers and bearing directly on family-product design.[40]
5. 🇯🇵 Japan governs gacha and monetization through industry self-regulation rather than statute: JOGA's odds-disclosure guidelines and the Premiums Act (the 2012 kompu-gacha ruling) set de-facto product rules, constraining design with a comparatively light-touch political posture.[12]
6. Child-safety politics are hardening minimum design standards: the UK's AADC was extended to mobile games in 2025, while US COPPA amendments and age-assurance momentum push high-privacy-by-default — directly governing design freedom for family products.[13][20]

Coming soon

Rising CPI and falling retention squeeze unit economics, shifting the contest to live-ops and retention; D2C web shops reclaim the ~30% store fee (~62% of the top-100 grossing titles ran one in 2026), which MIXI is already proving with Monster Strike. The US is the priciest market and biggest D2C battleground while India is high-volume/low-CPI/low-ARPU — and MIXI is building overseas sports (PointsBet/TIPSTAR) into a second pillar to cut yen and Monster Strike dependence.

1. Acquisition costs are climbing: 2026 mobile-game CPI averages ~$4.22 on iOS and ~$2.97 on Android, with North America the most expensive at +31% YoY. High-monetizing genres (RPG/strategy) run hottest, eroding the economics of buying growth by volume.[2]
2. Retention is broadly declining (GameAnalytics 2026): even top-quartile D1 is ~26–27%, median D7 ~3.4–3.9%, and 75% of titles fall below 3% by D28 — so live-ops and retention design now drive LTV more than install volume.[1]
3. 🇯🇵 D2C web shops reclaim the ~30% store fee: MIXI's Monster Strike Web Shop (PayPal, card, convenience-store payments, exclusive deals and bonus orbs) grew share 'beyond expectations' in its first year, while peers report DTC at 10%+ of revenue growing ~92% YoY.[5][3]
4. D2C has moved from experiment to default: in 2026 ~62% of the top-100 grossing titles run a web shop, mobile D2C revenue is up +46% YoY, and ~80% of first-time web-shop buyers repeat — with live-ops-driven shops outperforming. Reclaiming the fee is now a contest of operational quality, not whether to do it.[38]
5. 🇺🇸 The US is the most expensive and most valuable market and the biggest D2C battleground: North American CPI is +31% YoY[2] but US ARPPU (reportedly ~$48) towers over the global average (~$17),[39] and since the Apr 2025 external-payment ruling several titles including the Supercell Store have exceeded $1B via fee-free external links — making the US where reclaiming the store fee delivers the largest profit uplift.[29][4]
6. 🇮🇳 India is a high-volume, low-CPI, low-ARPU market: the mobile-game market was ~$3.5B in 2025 and Q1 2026 app IAP topped $300M (+33% YoY), but the real-money-gaming ban froze ~86% of the market overnight (~$3.2B, once projected toward ~$9B by 2029) — so MIXI-style IAP/D2C titles must win on volume and casual genres (puzzle), not stakes-based monetization.[22][21]
7. 🇦🇺 MIXI is diversifying away from Monster Strike dependence and yen exposure via overseas sports: it acquired ASX-listed PointsBet (~66.43% stake), its TIPSTAR betting platform grew +87.2% YoY (cumulative through Q3), the Sports segment doubled in Q3 FY2026 and full-year revenue came in at ~¥171.3B (up ~10.6% from ¥154.8B) — porting its social-betting craft from Monster Strike/TIPSTAR into Australia and Canada.[42][43][27][31]
8. Monetization is converging on hybrid designs: battle passes, mini/seasonal passes, dual-track free+premium paths and non-random purchase options alongside gacha are now standard, with hybrid-casual the fastest-growing category (~37% in 2024) and seasonal live-ops/battle passes estimated to drive ~20% of IAP revenue.[11][39]

Coming soon

Community and UGC are now core to retention and growth, even as developer and player sentiment toward generative AI sours and, for family products, trust itself is the product. India is mobile-first and price-sensitive with the collapse of real-money gaming reallocating talent and community; the US is both the vanguard of AI-in-production and the front line of disclosure scrutiny.

1. Community is a retention engine: Discord research finds players spend ~6x longer when a friend is present, and linked accounts see ~25% more launch days and ~16% more play sessions — the social graph beats any feature list for retention.[15]
2. UGC and creator economies keep scaling — Roblox, Fortnite and large creator Discords lead, and AI tools lower the barrier to creation — shifting players from consuming content toward co-creating it.[16]
3. Generative-AI sentiment is souring: at GDC 2026, 52% of devs call it harmful to the industry (up from 30%), and while 52% of companies use it only 36% of individuals do — a culture/trust gap that makes AI-made content a reputational risk.[6][34]
4. 🇯🇵 For family products, trust is the product: MIXI's FamilyAlbum (みてね) passed 30M users worldwide in May 2026 across 175 countries (~40% overseas), a domain where privacy-by-default and child protection are baseline expectations rather than features.[14]
5. 🇮🇳 India is institutionally separating betting from play: the new rules formally recognize esports as competitive sport and auto-approve non-paid social games (registration optional) while banning real-money games outright — so Dream11 lost ~95% of revenue overnight with 100+ exits, and MPL, Gameskraft and Zupee also cut staff, reallocating community and talent from stakes-based games toward casual/esports.[36][37]
6. 🇮🇳 India is mobile-first and price-sensitive: ~590M+ gamers, Android-dominated downloads with iOS driving higher-value spend, and growth led by Puzzle/casual — a market where free-to-play hooks and community matter more than deep IAP whales.[22]
7. 🇺🇸 US developers are the vanguard of AI-in-production yet also of disclosure scrutiny: ~92% report using vibe-coding practices,[30] but Steam-style player-facing disclosure and souring sentiment make 'how AI was used' a US go-to-market consideration.[26]

Coming soon

Generative AI is embedding into production (MIXI cut work hours 98% and built it into evaluation), with a named AI-NPC stack (Inworld, Convai), Unity Muse/NVIDIA ACE, player-facing AI and voice NPCs the 2026 frontier and code generation going mainstream (~60% of new code AI-generated in 2026), while measurement fragments under ATT/AdAttributionKit and the death of Privacy Sandbox.

1. 🇯🇵 MIXI has operationalized AI in production: on one Monster Strike project, combining Gemini, ChatGPT and Adobe Firefly compressed ~14 business days into ~1 hour — a 98% cut in work hours. The Design Division embedded AI use into its evaluation system, and guidelines v2.0 enabled video generation (Jul 2025).[7]
2. 🇯🇵 MIXI is also testing player-facing AI: an 'AI Companion' voice assist in Monster Strike aimed at UX and KPI uplift, plus event AI experiences (an AI ball-art maker and AI-character dialogue) — concrete cases of live-ops x AI on first-party IP.[9][8]
3. A named live-ops AI content stack is consolidating: Inworld AI's Character Engine (personality/memory/emotion/safety) and Convai's Actions (LLMs triggering in-game behavior, not just dialogue) for NPCs, ElevenLabs for voice, and Scenario.gg/Leonardo.AI for style-consistent art — dynamic, no-two-playthroughs-alike content is becoming a measurable retention driver for live-service titles.[23]
4. The AI-tooling map is sharpening: among AI users ChatGPT leads at 74%, Gemini 37% and Copilot 22%, with specialists Inworld (NPC dialogue), Unity Muse (asset generation), NVIDIA Omniverse/ACE (physics and NPC training) and ElevenLabs (voice); adoption skews to code and QA while art/character creation is still resisted, and player-facing AI features remain rare in shipped titles.[35][34]
5. AI is moving from assist to authorship in code: industry estimates put ~60% of new code as AI-generated in 2026 and senior-dev productivity at 3–5x, but up to ~45% of AI-generated code carries security vulnerabilities — so MIXI's 98% time cuts must pair with review/testing/security gates before shipping.[30][7]
6. GenAI is widely used in production (52% of companies) but concentrated in research/brainstorming (81%) and code assistance (47%); voice-driven, improvising autonomous NPCs and agent systems are the 2026 frontier, with content generation the next battleground.[6]
7. Measurement tech is fragmenting: Apple moved from SKAN to AdAttributionKit, ATT opt-in sits ~25–35%, and Google formally retired Privacy Sandbox in Oct 2025. User-level attribution is harder, pushing UA toward creative quality and first-party/D2C data.[17]

Coming soon

Responsible design is now a legal axis: the EU DFA, gacha odds disclosure and minor spend caps (self-regulated in Japan), children's codes, the US Apple v Epic external-payment fight (heading to SCOTUS in 2026), state App Store age-verification laws and Steam's AI-disclosure rules are starting to dictate how products are built.

1. The EU Digital Fairness Act (expected Q4 2026 proposal; public consultation closed Oct 24, 2025) eyes dark patterns and addictive design plus loot-box transparency and real-currency cost disclosure for virtual goods — potentially requiring parental consent for minors' paid loot boxes or restricting randomized rewards.[10][19]
2. 🇯🇵 Japan's gacha rules are self-regulatory at core: JOGA guidelines require pre-purchase odds disclosure and industry guidance also sets monthly minor spend caps (~¥5,000 for under-15s, ~¥10,000 for 16–19s), while kompu-gacha was withdrawn as unlawful under the Premiums Act in 2012 — so odds disclosure, pity timers and spend caps are de-facto design requirements rather than statute.[12][28]
3. 🇺🇸 US platform-fee and loot-box law is in flux: the Ninth Circuit upheld the Apple contempt finding in Dec 2025 (its 27% external-link fee struck down) and Apple is petitioning the Supreme Court in 2026, keeping fee-free external links alive for now;[24] separately, the FTC's loot-box staff perspective and consent orders push odds disclosure, age verification and parental controls without a federal statute.[25]
4. 🇺🇸 US fee and age-verification law concentrates in 2026: Apple formally petitioned SCOTUS on May 21, 2026 (challenging the contempt finding), with a June 25 conference deciding whether to take the case — if granted, argument in the Oct 2026 term and a decision in late 2026/early 2027 — while Utah's App Store law was litigated and its key provisions pushed to May 6, 2027. The premise of the D2C billing surface could shift abruptly.[44][24][41]
5. Storefront-level AI disclosure is now a compliance surface: Valve requires an 'AI Generated Content Disclosure' on Steam store pages and, after a Jan 2026 rewrite, focuses it on player-facing content (not behind-the-scenes dev tools), with live-generated content requiring described safeguards or risking delisting.[26]
6. Kids-product law constrains design itself: the UK AADC was extended to mobile games in 2025, and US COPPA amendments plus age-assurance mandates require data minimization and high-privacy defaults — directly relevant to family products like FamilyAlbum.[13][20]

Coming soon

Generative AI and always-on operations lift the compute/energy footprint, putting performance and efficiency design at the intersection of UX and sustainability.

1. Heavy use of generative AI in content pipelines plus always-on live-ops expand the compute and energy footprint, becoming a non-trivial factor for both operating cost and ESG disclosure.[6]
2. 2026 UX trends emphasize performance, battery/data efficiency and on-device processing — making performance-focused design a point where user-experience quality and sustainability align.[18]

Coming soon