AppSamurai: You’ve worked across different areas of mobile gaming, including casual titles as well as social casino and iGaming. From a user acquisition perspective, what are the biggest differences you’ve seen between these categories? Are the main differences in the channels you use, the metrics you optimize for, player behavior, creative strategy, or something else?
Kentaro Sugiura: I would say the biggest difference between mobile gaming genres is not necessarily the acquisition channels—it is the underlying business model and how paid and organic acquisition work together.
Some genres have a very broad audience, which can result in lower CPIs and a larger volume of organic installs. In these cases, paid users may have stronger average quality than organic users, so the payback model can be evaluated primarily on the performance of paid acquisition.
Other genres have higher CPIs and a smaller, more specialized audience. Paid acquisition alone may not appear profitable because there is a limit to how efficiently that audience can be reached. However, paid campaigns can also generate additional organic discovery. When paid and organic users are evaluated together, the overall economics may become profitable. In niche genres, organic users can sometimes be particularly valuable because they already have a strong interest in that type of game.
This changes how I measure performance. I look beyond attributed ROAS and consider the relationship between paid spend, organic uplift, total acquired-user value, and the appropriate multiplier—or “K-factor”—for each title. Incrementality testing is important because the same attributed result can have a very different business impact depending on how much organic demand the campaign generates.
The core advertising channels are generally similar across genres. What changes is the audience size, achievable scale, payback expectations, and the way we combine paid and organic impact when making investment decisions.
AppSamurai: Meta has grown its share of impressions considerably over the past year, while a couple of other major platforms have pulled back. Does that concentration actually make your job easier, since there's less channel-testing overhead, or does it worry you as a UA lead to have that much budget dependent on one platform's auction dynamics?
Kentaro Sugiura: Channel diversification is essential when managing a large portfolio. Concentrating spend on Meta may make day-to-day execution simpler, but it also creates significant risk. If performance declines because of auction pressure, an algorithm change, creative fatigue, or a measurement issue, you need somewhere else to allocate the budget.
That is why I encourage teams to test different channel types regularly, including social platforms, SDK networks, DSPs, preloads, and rewarded or incentivized channels. A channel that did not work one year ago may perform differently today because its technology, optimization models, inventory, or targeting capabilities have improved.
Testing does create additional operational work, but I see it as maintaining future options. Not every experiment needs to receive a large budget immediately. The objective is to build reliable benchmarks, understand where each channel can contribute, and identify alternatives before they become urgently necessary.
Meta can remain a major part of the portfolio when it delivers strong marginal returns, but concentration should be an intentional investment decision—not the result of having stopped testing. A diversified portfolio gives the team more flexibility to move spend, protect overall performance, and take advantage of improvements across the wider advertising ecosystem.
AppSamurai: You've spoken before about the importance of challenging assumptions in re-engagement strategies, particularly around inactive windows and retargeting rules. In your experience, what are some of the most common assumptions that mobile game teams still make about lapsed players, and what kind of an approach leads to better reactivation outcomes?
Kentaro Sugiura: One common mistake is assuming that the same inactivity window should apply to every game. The appropriate window depends heavily on the genre, the competitive environment, and how easily players can replace the experience with another product.
For example, I worked on a casual game with very distinctive gameplay and no close direct competitor. In that situation, the inactivity window could be longer. Players were less likely to find an equivalent experience elsewhere, so there was less urgency to retarget them immediately. The main question was whether and when they would naturally return.
I have also worked on games in niche but highly competitive categories, where several products offer a similar experience. In that situation, waiting too long creates a greater risk that players will form a new habit with a competitor. Retargeting therefore needs to begin earlier, with a shorter inactivity window and messaging connected to relevant content, events, or player progress.
The key is not to define a player as “lapsed” using one universal rule. I would test different inactivity windows by genre, player value, previous engagement, and competitive intensity. I would then use holdout groups to determine whether the campaign genuinely caused players to return rather than simply claiming credit for users who would have returned organically.
A successful re-engagement strategy reflects how replaceable the game experience is and how quickly the player’s attention may shift elsewhere.
AppSamurai: Playable ads have become a staple of mobile game marketing, but some games are much harder to translate into playables than others. As a UA leader, how do you work with other teams to design playables that maximize campaign performance while still giving players an authentic sense of what they'll experience in the game?
Kentaro Sugiura: I see two main challenges when developing playable ads: technical limitations and accurately representing player motivation.
I previously worked on a mobile game based on a major AAA IP. The game was large and visually complex, so reproducing the full experience within a playable’s strict file-size and technical limitations was difficult. We worked closely with the Product team to identify the essential assets and mechanics, then collaborated with an external playable-development partner to create a simplified experience that still felt recognizably connected to the game.
The second challenge is understanding why players want to play. A playable does not necessarily need to reproduce every system or present an exact miniature version of the product. If the core motivation is collecting, progression, or strategic decision-making, the playable should communicate that motivation through a simple interaction that works effectively within the format.
In my experience, motivational consistency is more important than literal replication. Players may not remember every detail of the advertisement that led them to install, but they will notice if the product fails to satisfy the expectation the ad created. The playable can simplify the mechanics, but it should not misrepresent the underlying reason the game is enjoyable.
My role as a UA leader is to connect performance data with Product, Creative, and technical expertise. We test different simplified expressions of the core motivation, evaluate engagement and install conversion, and then validate the acquired users through retention, predicted LTV, and ROAS. That allows us to maximize performance while ensuring that the playable attracts players who are genuinely suited to the game.
AppSamurai: You've flagged rewarded networks and CTV as channels gaining real traction, and influencer marketing moving from one-off deals toward long-term, performance-driven partnerships. Which of those three bets has actually paid off the way you expected, and which one turned out to be more hype than substance once you tried to scale it?
Kentaro Sugiura: All three channels are developing quickly, but their scalability depends heavily on the game’s genre, predicted LTV, and business model. Of the three, rewarded acquisition is currently the most proven from a direct-performance perspective.
I see two main rewarded-network models. The first is app-based: the network acquires users for its own rewards app and presents advertiser offers inside it. The second is SDK-based, where rewarded inventory is integrated into third-party apps. In my experience, the app-based model can be more scalable. However, its economics are affected by the wider advertising market. Rewarded platforms must acquire their own users, so when market CPIs rise, that additional cost is eventually passed to advertisers. Games with a sufficiently high predicted LTV can absorb the higher CPI, while lower-LTV genres may struggle to make the model profitable.
CTV also has meaningful potential. Previously, its inventory was concentrated among specialist CTV partners, whereas today many DSPs and established UA networks provide access to it. The main limitation is measurement. Traditional MMP attribution does not capture the complete cross-device journey. A user may see an advertisement several times on television and later convert after seeing a Meta ad, causing the final touchpoint to receive most of the credit.
CTV therefore requires incrementality testing, geographic experiments, and broader analysis of the customer journey. If it is evaluated only through traditional attributed ROAS, it can appear either much better or much worse than its true contribution. I would describe CTV as promising rather than hype, but its measurement capabilities still need to mature.
Influencer marketing is also becoming much more performance-oriented. Creator programs can provide revenue sharing through referral codes, bringing a model already common in e-commerce into gaming. Platforms such as TikTok are also making campaigns more programmatic: advertisers can define objectives, audience or creator requirements, and cost expectations, while creators apply to participate. This makes it possible to run influencer activity at a much larger scale without managing every relationship manually.
Overall, rewarded acquisition has produced the clearest direct-performance results. CTV and influencer marketing both have significant potential, but neither is automatically scalable. Their success depends on product economics, technical implementation, and measurement. I expect both to improve as attribution, experimentation, and campaign-management technology become more sophisticated.
