Gaming is one of the most competitive, most data-rich, and most misunderstood verticals on Google Ads. The mechanics that make a B2B lead-gen account profitable — clean intent capture, modest creative refresh, last-click attribution that roughly works — break down completely for a free-to-play mobile game where a fraction of a percent of players generate most of the revenue and the median user never spends a cent. Treating gaming UA like generic performance marketing is the fastest way to burn a launch budget.
This is a vertical playbook for gaming and esports advertisers. We cover App campaigns and the install learning phase, the progression from cost-per-install bidding to value-based bidding, player LTV modeling, the YouTube and creator layer that gaming audiences live in, esports event timing, re-engagement of lapsed players, and how to compete for installs in a saturated genre. It is written for gaming UA managers and growth leads who already understand their game's economics. For the cross-platform app-promotion fundamentals, our Google Ads app promotion guide is a useful companion, as is the Apple Search Ads and ASO guide for the iOS side.
The most common mistake in gaming UA is optimizing toward the lowest cost-per-install. A €1.50 CPI looks great on a dashboard and can be a disaster if those installs are low-intent players who never reach the tutorial, let alone a purchase. A €12 CPI can be wildly profitable if it delivers players with €40 lifetime value. The number that matters is the relationship between what you pay to acquire a cohort and what that cohort is predicted to be worth — CPI divided by predicted LTV, watched as a payback curve. Every bidding and creative decision in this guide flows from putting LTV next to CPI, never CPI alone.
Why gaming UA needs its own Google Ads playbook
Three structural features of gaming make generic Google Ads tactics fail and demand a vertical-specific approach.
Revenue is extremely skewed. In free-to-play, the distribution of player spend follows a steep power law — a small cohort of high-spending players (often called whales) drives the majority of revenue, while most users monetize little or not at all. According to industry monetization data tracked by firms like Sensor Tower and AppsFlyer, a low single-digit percentage of payers commonly accounts for the bulk of in-app purchase revenue. This skew means that average revenue per install, the metric most performance marketers reach for, is actively misleading. You are not buying average players; you are buying a distribution, and the value lives in the tail.
The market is enormous and crowded. Newzoo's market sizing puts global games revenue well into the hundreds of billions of dollars, with mobile the largest segment. Big numbers attract big spenders: the top-grossing genres feature studios with eight- and nine-figure annual UA budgets who have refined their LTV models and creative pipelines for years. A new entrant cannot outspend them, which forces a strategy built on efficiency and creative rather than raw budget.
The audience is native to video and community. Gamers discover, evaluate, and discuss games on YouTube, Twitch, Discord, and creator channels. This makes YouTube App campaign inventory and gaming-creator integrations unusually powerful, and it makes pure text-based intent capture unusually weak relative to other verticals. Demand is created visually, in community, before it is captured.
The combined implication: gaming UA is a value-modeling and creative discipline first, and a bidding-and-keyword discipline second. The studios that win treat their LTV model as core IP, run a creative pipeline that produces dozens of fresh assets a month, and exploit timing and community in ways a generic account never would. The rest of this guide builds that playbook step by step.
App campaigns: install volume and the learning phase
App campaigns (formerly Universal App Campaigns) are Google's primary vehicle for driving game installs. They are heavily automated: you supply assets, targeting geographies, a bid, and an optimization goal, and Google's machine learning distributes your ads across Search, Play, YouTube, Discover, and the display network, optimizing toward your goal.
The asset mix is the lever. Because App campaigns are automated, your creative assets are the largest controllable input to performance. Supply the full range: portrait and landscape video, gameplay-footage video, static images, and HTML5 playables. Playables in particular punch above their weight for games because they let users sample the core loop before installing, raising install quality. The algorithm mixes and matches assets across placements, so breadth and freshness of creative directly determine how efficiently the campaign performs.
Respect the learning phase. App campaigns need conversion volume and uninterrupted time to optimize. The single most common operational error is editing campaigns too soon — changing bids, swapping assets, or pausing during the learning period resets optimization and wastes spend. Seed enough budget that the campaign gathers meaningful conversion data quickly, then leave it alone through the learning phase before judging performance or making changes.
Start with install volume goals. For a launch, begin with an install-focused goal (target CPI) to build a base of users and seed the algorithm with conversion data. This generates the install and early-event volume you need before you can credibly optimize toward value. We will move beyond installs in the next section, but installs are the necessary first rung.
Geography and budget structure. Structure campaigns by tier — tier-1 high-monetization markets separate from emerging markets — because CPI and LTV differ by an order of magnitude across them and mixing them muddies optimization. A soft-launch in a smaller, representative market is the classic way to validate creative and LTV assumptions cheaply before scaling to expensive tier-1 geographies.
App campaigns reward patience and creative volume. Get the asset mix broad and fresh, fund the learning phase properly, and resist the urge to micromanage — then layer value optimization on top once the foundation is producing data.
From tCPI to tROAS: bidding for in-app purchases
The progression from install bidding to value bidding is the defining arc of a maturing gaming UA program. Move through it in sequence, not in one leap.
Why the sequence matters. Each stage requires denser conversion signal than the last. tCPI optimizes on installs, which are plentiful from day one. tROAS optimizes on revenue events, which are sparse early because most users do not purchase, and which arrive with a delay. Jumping straight to tROAS on a fresh campaign starves the algorithm of the purchase signals it needs, leaving it stuck in the learning phase and spending inefficiently. The intermediate target-CPA-on-event stage bridges the gap by optimizing toward a more frequent quality signal (a progression milestone or first purchase) before you ask the system to optimize toward revenue directly.
The volume threshold for tROAS. As a practical rule, a campaign should generate purchase events consistently — typically dozens per day — before tROAS will optimize reliably. Below that, value bidding behaves erratically. If a campaign cannot reach that purchase volume, keep it on target CPA for a high-value in-app event instead, which gives the algorithm a denser quality signal to work with.
Day-window ROAS versus lifetime ROAS. Early tROAS necessarily optimizes on a short revenue window (day-0 to day-7) because that is the data available in real time. The risk is that short-window ROAS under-credits games whose monetization builds over weeks. The fix is predicted LTV, covered next: instead of bidding to the small amount of revenue observed in seven days, you bid to the forecast lifetime value those early signals imply. This is the single most important upgrade a gaming UA program makes, and it is what separates studios that scale profitably from those that plateau.
The throughline is signal density. Match your bid strategy to how much conversion signal your campaign actually produces, advance through the stages as volume grows, and never ask an automated bidder to optimize toward a goal it does not have enough data to learn.
Player LTV modeling and value optimization
Player LTV modeling is the core competency of profitable gaming UA, and it is what lets every other tactic in this guide work.
Why average revenue lies. Because spend is skewed to whales, the average revenue per install of a cohort tells you almost nothing about whether a specific acquisition source is good. Two sources can have identical average revenue while one delivers a few high-value players among many non-payers and the other delivers uniformly mediocre ones — and the first is far more valuable because high-value players retain and re-spend. You have to model the distribution, not the average.
Predictive LTV from early signals. The practical technique is to use early-life behavior — day-0 and day-1 session depth, progression speed, early purchases, retention at day-1 and day-3 — to forecast a cohort's value at day-30, day-90, and day-180. Studios build these models on their own historical cohort data: take mature cohorts whose true long-window value is known, learn which early signals predicted it, and apply that mapping to fresh cohorts whose long-window value is not yet observable. The output is a predicted LTV per user or per cohort, available within days of acquisition.
Feeding predicted LTV back into bidding. A predicted LTV model is only useful for UA if it closes the loop into bidding. Two mechanisms do this: value rules, which adjust the conversion value Google sees based on segments you know correlate with value (geography, platform, audience), and server-side conversion values, where you send Google a predicted-LTV value for each user rather than only the observed short-window revenue. With predicted LTV flowing into tROAS, Google optimizes toward forecast lifetime value — it will pay more to acquire a player whose early signals predict high value and less for one whose signals predict churn.
Every studio that scales profitably eventually stops bidding to installs or to seven-day revenue and starts bidding to predicted lifetime value. The ones that plateau keep optimizing toward cheap installs, win the CPI race, and lose the LTV race — they acquire enormous volumes of players who never monetize while a disciplined competitor quietly pays triple the CPI for the players who actually matter. LTV modeling is not a reporting nicety; it is the bidding signal.
The measurement foundation. None of this works without clean tracking. A mobile measurement partner (AppsFlyer, Adjust, Singular) integrated with Google Ads, a complete in-app event taxonomy with revenue values attached, and correct SKAdNetwork and Android measurement configuration are the prerequisites. The LTV model is only as good as the event and revenue data feeding it. The same data foundation that powers LTV modeling — joined acquisition cost and downstream revenue by cohort — is exactly what a warehouse like the one in our BigQuery data pipeline tutorial is built to maintain at scale.
Treat your LTV model as proprietary IP. It is the asset that lets you out-bid competitors on the players who matter while out-saving them on the players who do not.
YouTube and gaming creator strategy
Gaming audiences live on video, which makes YouTube and creators disproportionately powerful for game UA compared with other verticals.
YouTube inside App campaigns. App campaigns already distribute video creative across YouTube, so strong video assets directly improve App campaign efficiency. Gameplay-footage video, trailer-style creative, and short hook-driven clips perform because they show the core loop the audience cares about. Investing in video creative is not a separate channel decision — it is the highest-leverage input to your existing install campaigns.
Demand generation versus capture. Beyond App campaigns, a dedicated YouTube and Demand Gen presence creates the demand that makes capture cheaper later. When players have seen gameplay, recognized the brand, and formed intent, your App campaigns convert them more efficiently — effective CPI falls because the audience is warm. This is the gaming version of the demand-creation-feeds-demand-capture dynamic, and it argues for sustained top-of-funnel video investment rather than pure performance spend. Our Demand Gen campaigns guide covers the format mechanics.
Gaming creators. Creator integrations — sponsored playthroughs, channel integrations, branded content — tap directly into the communities where game discovery happens. Their installs are real, but their larger value is the warmed audience they leave behind. The measurement challenge is that last-click attribution systematically under-credits creators: a viewer watches a playthrough, thinks about it for days, and installs later through a different touchpoint. Measure creators with deep links and promo codes for direct attribution, and run incrementality tests — turning creator activity on and off in matched conditions — to capture their true lift. The methodology in our incrementality testing guide applies directly.
Practical creator workflow. Brief creators with the hooks and moments that convert (the core loop, a standout feature, an event), supply trackable links, and concentrate creator pushes around timing windows — launches, content drops, esports events — where their reach compounds with everything else you are running. Treat creators as a demand-generation engine measured by incrementality, not a performance channel measured by last click.
The studios that get gaming UA right understand that video and community are not adjacent channels — they are where the audience forms intent, and they make every install campaign cheaper.
Esports event timing and seasonal pushes
Esports and live-service games run on a calendar of sharp demand spikes, and timing budget to those spikes is a distinct competency.
Demand is event-driven and short-window. Major tournaments, season launches, expansions, and patch drops each create a concentrated surge in search interest, YouTube viewership, and player intent. The window is narrow — interest peaks in the days around the event and decays quickly. An always-flat budget either underspends during these spikes (missing the cheapest, highest-intent acquisition of the quarter) or overspends in the troughs between them.
Treat the esports calendar like retail treats Black Friday. The discipline is the same as seasonal e-commerce: map the calendar in advance, pre-build event-themed creative, raise budgets and value targets ahead of the spike, and pace spend to front-load the window. The studios that capture these windows plan them weeks ahead; the ones that miss them are still building creative when the event is already peaking. Our budget pacing guide covers the pacing mechanics that make front-loading work without exhausting budgets early.
The post-event recapture. Events do not just acquire new players — they reactivate lapsed ones who tune back in for the tournament or the new season. A pre-planned re-engagement burst in the one to two weeks after an event captures this returning population efficiently, because these are players who already know the game and are re-entering with fresh intent. Pairing acquisition during the peak with re-engagement after it extracts the full value of the window.
Seasonality beyond esports. General seasonal patterns — holiday periods, back-to-school, regional events — also shift gaming demand and competition. The same pre-build-ramp-peak-recapture rhythm applies. The broader point is that gaming demand is lumpy, not smooth, and a UA program paced to the calendar consistently outperforms one that spends evenly across it.
Build the next 90 days of relevant events into your plan, pre-load creative and budget for each, and treat every major event as both an acquisition and a re-engagement opportunity.
Retention and re-engagement campaigns
Acquisition gets the attention, but for games with meaningful churn — which is nearly all of them — re-engagement of existing players is often the more efficient source of revenue.
What re-engagement campaigns do. App campaigns for engagement target users who have already installed your game, bringing lapsed players back and deepening the engagement of active ones. The mechanism that makes them work is deep linking: instead of dropping a returning player on a generic launch screen, the ad links directly into a relevant in-game destination — a live event, a sale, a new season, a tournament — so the player lands where the value is. This dramatically improves the odds that a re-engaged player actually re-engages rather than bouncing.
Why the economics are favorable. Lapsed players already understand your game, have an account and progress, and convert to purchase at a higher rate than cold installs once reactivated. That makes the cost per revenue euro from re-engagement frequently better than from fresh acquisition. For a mature game, re-engagement should be a standing line item, not an afterthought.
Segment by lifecycle stage. Different lapsed populations need different treatment:
- Recently churned active players — bring them back to whatever they were last engaged with, plus a reason to return (new content, an offer).
- Long-dormant players — re-introduce major changes since they left; the game may have evolved substantially.
- Active non-payers — deep-link into monetization moments and offers timed to events.
- Active payers — deepen engagement and surface premium content; protect this high-value cohort.
Tailoring the deep-link destination and offer to each segment materially outperforms a one-size-fits-all re-engagement campaign.
Scale around events and content. Re-engagement compounds with the event timing from the previous section: a content drop, season launch, or esports event is exactly when lapsed players are most receptive to coming back. Run re-engagement continuously as a baseline and scale it up sharply around these moments.
Retention-focused spend is where many studios find their best blended efficiency. A player you already paid to acquire is cheaper to reactivate than a stranger is to acquire, and deep-linked re-engagement is the tool that captures that advantage.
Competing for CPI in a high-competition vertical
Gaming is brutally competitive, with deep-pocketed studios bidding aggressively in the most lucrative genres. Winning is possible, but not by outbidding — by out-modeling and out-creating.
You cannot win the raw CPI race against scale. A studio with a nine-figure UA budget and years of model refinement will outbid a new entrant on undifferentiated install auctions. Trying to match them euro-for-euro on CPI is a losing strategy that drains budget for low-quality volume. The competitive edge has to come from somewhere other than spend.
Edge one: bid to LTV, not CPI. Predicted-LTV bidding is the great equalizer. It lets you pay above-market CPI for the specific players your model identifies as high-value, while paying below-market for everyone else — so your blended economics beat a competitor who bids the same CPI for all players indiscriminately. You are not winning every auction; you are winning the right auctions. This is why the LTV modeling in section 4 is the foundation of competitive UA, not a reporting luxury.
Edge two: creative velocity. In automated App campaigns, creative is the single largest controllable lever on CPI. The studios that sustain efficient acquisition run a high-velocity creative pipeline — dozens of fresh assets a month across video, playables, and images — constantly testing and retiring. Creative fatigue raises CPI relentlessly, so the pipeline is not a one-time effort but a permanent capability. Out-producing competitors on creative volume and quality directly out-competes them on CPI.
Edge three: timing and thinner auctions. Competition is not uniform. Esports event windows, seasonal moments, soft-launch geographies, and emerging genres all offer auctions where the giants are less present or less optimized. Concentrating spend where competition is thinner — and where your timing and creative give you an edge — beats fighting head-on in the most contested placements. The event-timing discipline from section 6 is partly a competitive tactic for exactly this reason.
For studios weighing platform mix, the install economics differ between Google's network and Apple's ecosystem; our Apple Search Ads and ASO guide covers the iOS side, and the Google Ads app promotion guide covers the cross-platform fundamentals.
The summary of competitive gaming UA: accept that you will lose the raw-CPI auction to scale, and win instead on the dimensions scale does not automatically buy — sharper LTV modeling, faster creative, and smarter timing.
If you want AI-driven optimization that watches your value-based bidding and creative performance across campaigns so your team can focus on LTV modeling and creative production, SteerAds runs a free 14-day audit on Google and Microsoft Ads accounts.
Sources
- support.google.com/google-ads — Google Ads App campaigns documentation
- thinkwithgoogle.com — Think with Google gaming and app insights
- newzoo.com — Newzoo global games market data
- appsflyer.com — AppsFlyer gaming UA and LTV benchmarks
- sensortower.com — Sensor Tower mobile game market intelligence
FAQ
What is a realistic CPI for mobile games on Google Ads in 2026?
It varies enormously by genre and geography. In tier-1 markets (US, UK, Western Europe), casual and hypercasual games often see CPIs of €1-3, mid-core titles €4-10, and hardcore or strategy games €8-25 because they target higher-value, harder-to-reach players. Emerging markets run far lower, sometimes €0.20-1.00, but with correspondingly lower monetization. CPI alone is the wrong number to optimize, though — a €15 CPI that delivers €40 LTV players beats a €2 CPI that delivers €1 LTV players. Always pair CPI targets with day-0 to day-7 ROAS expectations from your own cohort data.
Should I bid on installs (tCPI) or in-app purchases (tROAS)?
Start on tCPI to build install volume and seed the algorithm with conversion data, then graduate to tROAS once you have enough purchase events to model value reliably. Most studios run a sequence: launch on tCPI for installs, switch to target CPA for a key in-app event (tutorial complete, first session milestone), then move to tROAS once daily purchaser volume is sufficient for Google's value model. tROAS too early starves on sparse purchase signals and never exits the learning phase. The transition usually happens once a campaign generates dozens of purchase events per day consistently.
How important is player LTV modeling for gaming UA?
It is the single most important capability separating profitable from unprofitable gaming UA. Because monetization in free-to-play games is heavily skewed — a small share of whales drives most revenue — average revenue per install is misleading. You need predictive LTV: from early signals (day-0 and day-1 behavior, first purchases), forecast a cohort's day-30, day-90, and day-180 value, then feed that predicted value back into tROAS bidding via value rules or server-side conversion values. Studios that bid to predicted LTV consistently outperform those bidding to install volume or short-window ROAS.
Do YouTube and gaming creators actually drive installs, or just awareness?
Both, and the line between them blurs in gaming specifically because the audience is native to video. YouTube App campaign inventory and creator integrations drive measurable installs, but their bigger value is creating the demand that makes your tCPI campaigns cheaper. Gameplay footage, creator playthroughs, and trailer-style creative on YouTube warm an audience that then converts more efficiently across the App campaign network. Measure creators with promo-code or deep-link tracking and incrementality tests rather than last-click — their contribution is systematically under-credited by attribution windows, especially for longer consideration mid-core titles.
How should esports advertisers time their campaigns around events?
Esports demand spikes sharply around major tournaments, season launches, and patch drops, and the window is short. Build budget pacing that front-loads spend in the two to three weeks before and during a major event, when search interest, YouTube viewership, and player intent all peak together. Pre-build creative referencing the event, raise budgets and tROAS targets ahead of the spike, and prepare re-engagement campaigns to capture the post-event surge of returning and lapsed players. Treating esports calendars like retail treats Black Friday — planned, pre-loaded, and tightly paced — is the difference between catching the wave and missing it.
What does a re-engagement campaign do for a mobile game?
Re-engagement (App campaigns for engagement) targets users who already installed your game to bring lapsed players back and deepen engagement of active ones. It uses deep links to drop returning players directly into a relevant in-game destination — a new event, a sale, a tournament — rather than a generic launch screen. For games with meaningful churn (which is nearly all of them), re-engagement is often more cost-efficient per revenue euro than fresh acquisition because lapsed players already understand the game and have a higher conversion-to-purchase rate. It should be a standing line item, scaled up around content drops and events.
How do I compete on CPI in a saturated, high-competition genre?
You rarely win a saturated genre by outbidding on CPI — you win on LTV efficiency and creative. The studios that sustain profitable UA in crowded categories do three things: they bid to predicted LTV so they can pay more for genuinely valuable players without overpaying for the rest, they maintain a high-velocity creative pipeline because creative is the largest lever on CPI in App campaigns, and they exploit timing windows (events, seasonality, soft-launch geos) where competition is thinner. Trying to match a deep-pocketed competitor euro-for-euro on raw CPI is a losing game; out-modeling and out-creating them is not.