--- title: How mobile UA changed (2015 → 2026) — the remaining levers, team structure and channel mix by budget slug: mobile-ua-evolution-and-team-structure summary: From 100-person manual media-buying teams (Machine Zone era) to algorithmic buying on 4–5 consolidated networks — what's left to control (bid, campaign type, budget, creative, signal), how a modern UA team is structured, and how channels and process change from six- to eight-figure monthly budgets (gaming lens). domain: marketing tags: [user-acquisition, mobile-gaming, team-structure, channel-mix, creative, automation, history] related: [signal-engineering-and-conversion-event-audit, evaluating-new-ad-channels, measurement-stack-attribution-incrementality-mmm, paid-social-creative-testing-framework, app-ad-creative-concepts, ua-team-and-stack-by-spend-level] sources: [yt:3Uj6p1N8VJA, yt:iD_YbEW0Y3w] updated: 2026-10-24 --- # How mobile UA changed — levers, team and channel mix ## TL;DR - **2015.** Advantage came from people and data. Machine Zone ran 100+ media buyers, 5–10 per Facebook/Google, custom publisher- and placement-level reports, and supply-path optimisation on programmatic buys. - **2026.** Buying is algorithmic: Google App Campaigns (UAC), Advantage+, AppLovin. After consolidation, 4–5 networks cover ~95% of reach. You control four levers: **bid/target, campaign type (optimisation event), budget, creative**, plus the signal you feed. - **Execution is easier, winning is harder.** Competition is global (Asian studios scale in the West via AppLovin), so UA amplifies great products rather than finding arbitrage. Nebo's rule of thumb for games: "if your product is not working on AppLovin, don't launch it." - **Team structure.** Three pillars: media buying (a few senior, AI-fluent "super ICs"), creative (strategist as bridge to production) and analytics (incl. creative analytics and predictive LTV). - **Gaming channel mix by monthly spend:** - $100–300K: Google on Android plus AppLovin on iOS only - $0.5–2M: add Unity and Meta - $1–5M: add geos and localisation - 8 figures: audience segmentation, brand/IP partnerships, and incrementality + MMM ## Details **The manual era** ([[people/nebojsa-radovic]], ex-Machine Zone). Daily spend reports were pulled and stitched by hand. Better reporting, especially creative- and publisher-level, was the competitive edge.[^1] Nebo ran DSP buying at a level of granularity "we don't really get… even today": which network and which impression in a given app's waterfall was cheapest. This is supply-path optimisation, plus direct publisher deals.[^2][^3] A marketing-technology team automated anything taking more than 60 minutes.[^4] In 2015 Facebook and Google took ~60% of spend (Google ~20%), programmatic and video networks 20–25%, with incentivised traffic and preloads (~5%) making up the rest.[^5] **Automation and consolidation.** Nobody thought UAC would work, and it underperformed for about two years. Then it, and later Advantage+, democratised sophisticated buying.[^6] Network mergers (ironSource+Unity, Liftoff+Vungle, AppLovin's acquisitions of MAX and MoPub) left a handful of channels with ~95% coverage.[^7] What remains is bid, campaign type (CPI, retained user, D1/D7 ROAS), budget and creative. Creative is "the only one kind of high touchpoint item".[^8] With global competition, profitable UA is hard.[^9] Test new games on AppLovin early in development: if they can't run there at scale, they're unlikely to succeed.[^24] Today the top spenders are simply the best products (Royal Match, Monopoly Go).[^10] **Creative operations.** Refresh monthly for games spending a few hundred thousand dollars a month; almost daily at millions per month. Top AppLovin spenders launch thousands of ads a month with creative teams of 300–500.[^11] The success metric is **hit rate or "hero rate"**: the share of new creatives that win a meaningful slice of spend (10–30%) at goal.[^12] Licensed IP adds about a week of approvals on top of 3–4 days of production.[^13] [[people/thomas-petit]] warns that volume is often fake: "you've launched a thousand ads… no, you've launched 18". Count distinct concepts, not variants.[^14] **Team structure** (Zynga today).[^15][^16] - **Media buying.** T-shaped, AI-fluent UA leads who can run a game alone; changes every other day, not hourly. - **Creative.** Strategists sit between UA and production. They read IPM, CPI and retention data and write weekly briefs. - **Analytics.** Marketing analytics (opportunities, predictive LTV) plus creative analytics at the centre of all three pillars. Signal engineering is everyone's job, not a dedicated role.[^17] **Scaling stages (games).** $100–300K/month: Google Android and AppLovin iOS, "and do it well"; Google on Android is "truly incremental".[^18] $0.5–2M: add Unity and Meta; Meta isn't a top-3 game channel now. One UA lead with shared creative and analytics resources.[^19] $1–5M: more geos, localisation and culturalisation. 7→8 figures: a dedicated pod, audience segmentation beyond the "golden cohorts", brand (TV, IP licences, ambassadors), and measurement complex enough to need incrementality and MMM.[^20][^21] **AI and the job.** Platform automation (Advantage+, PMax, Axon) already acts as the agent. The human job is handling the steady stream of edge cases and broken data pipelines ("I'm a plumber") and finding non-obvious patterns that off-the-shelf setups miss.[^22][^23] ## Disagreements / caveats - This is a **mobile gaming** lens with large budgets. For subscription apps Meta is usually an early channel (see [[evaluating-new-ad-channels]], [[paid-social-creative-testing-framework]]). [^1]: yt:3Uj6p1N8VJA @ 03:16 — Nebojsa Radovic [^2]: yt:3Uj6p1N8VJA @ 06:45 — Nebojsa Radovic [^3]: yt:3Uj6p1N8VJA @ 09:31 — Nebojsa Radovic [^4]: yt:3Uj6p1N8VJA @ 10:18 — Nebojsa Radovic [^5]: yt:3Uj6p1N8VJA @ 11:53 — Nebojsa Radovic [^6]: yt:3Uj6p1N8VJA @ 15:04 — Nebojsa Radovic [^7]: yt:3Uj6p1N8VJA @ 16:39 — Nebojsa Radovic [^8]: yt:3Uj6p1N8VJA @ 18:12 — Nebojsa Radovic [^9]: yt:3Uj6p1N8VJA @ 19:05 — Nebojsa Radovic [^10]: yt:3Uj6p1N8VJA @ 21:43 — Nebojsa Radovic [^11]: yt:3Uj6p1N8VJA @ 29:43 — Nebojsa Radovic [^12]: yt:3Uj6p1N8VJA @ 30:35 — Nebojsa Radovic [^13]: yt:3Uj6p1N8VJA @ 31:24 — Nebojsa Radovic [^14]: yt:iD_YbEW0Y3w @ 14:33 — Thomas Petit [^15]: yt:3Uj6p1N8VJA @ 35:24 — Nebojsa Radovic [^16]: yt:3Uj6p1N8VJA @ 37:17 — Nebojsa Radovic [^17]: yt:3Uj6p1N8VJA @ 38:08 — Nebojsa Radovic [^18]: yt:3Uj6p1N8VJA @ 42:10 — Nebojsa Radovic [^19]: yt:3Uj6p1N8VJA @ 42:58 — Nebojsa Radovic [^20]: yt:3Uj6p1N8VJA @ 43:43 — Nebojsa Radovic [^21]: yt:3Uj6p1N8VJA @ 44:32 — Nebojsa Radovic [^22]: yt:3Uj6p1N8VJA @ 46:54 — Eric Seufert [^23]: yt:3Uj6p1N8VJA @ 47:41 — Nebojsa Radovic [^24]: yt:3Uj6p1N8VJA @ 20:45 — Nebojsa Radovic