The retention numbers nobody puts in a pitch deck
Start with what the free-to-play model actually achieves. Good mobile game retention in 2026 sits at roughly 27% on day one, 8 to 14% on day seven, and 3 to 7% on day thirty. Those are the benchmarks studios aim at.
The medians are considerably worse. Broader 2026 figures have slipped to about 22% on day one, 4% on day seven, and under 1% on day thirty. Even among the top quartile of games, day-thirty retention typically sits between 1.6 and 1.8%. Only the top 1% reach 13 to 15%.
Revenue reflects the strain. Sensor Tower put total in-app purchase revenue at $81.7 billion in 2025, a rise of just 1.3% on the previous year. Install-to-purchase conversion for most mobile apps runs between 1 and 2%. The dominant monetisation model of the last decade is growing at barely above flat while converting a low single-digit share of the people it reaches.
The incentive under advertising and in-app purchase
A free-to-play mobile game earns from two sources: advertising impressions and in-app purchases. Both scale with time spent. Neither scales with whether the player is enjoying themselves, improving, or would describe the experience as a good use of an evening.
This produces a well-documented set of design consequences. Progression is paced to create friction that a purchase resolves. Sessions are structured around return triggers rather than natural conclusions. Difficulty is tuned to sit just past the point of comfortable play, because that is where conversion happens. None of this is malicious. It is what the revenue model rewards, and studios that ignore it lose to studios that do not.
Crucially, skill has almost no commercial value in this model. A player who masters the game stops encountering the friction that generates revenue. The system is, in a narrow sense, optimised against expertise. A player who becomes genuinely good at a free-to-play title becomes less valuable to it, which is a strange thing for a game to be built to do.
It also explains the retention figures. A design tuned to monetise frustration produces frustration, and frustration produces the under-1% day-thirty median. The model is not failing at what it optimises for. It is succeeding, and the retention numbers are the cost of that success.
The incentive under entry fees
Competitive entry-fee formats invert this. The operator takes a service fee on a match between two players. Revenue scales with matches entered, which scales with how many players stay active, which scales with whether the competition feels worth entering.
First, mastery becomes commercially valuable. A player who improves enters more matches, not fewer, because improvement is the reward loop rather than an obstacle to monetisation. The operator has a direct financial interest in players getting better, which is not a sentence that can be written about the advertising model.
Second, fairness becomes load-bearing. If the matching system routes beginners to experts, beginners leave, and the platform’s revenue base contracts. Free-to-play can survive an unfair difficulty curve because the curve is the monetisation. Paid competition cannot survive unfair pairing, because unfair pairing is simply a transfer from new players to established ones with the platform taking a cut of the extraction.
Third, session length stops mattering. A match has a defined end. There is no commercial reason to extend it, and no mechanism by which extending it would generate revenue. The entire apparatus of energy systems, daily login streaks and session-stretching mechanics has no function.
What the money says
The market data tracks the shift. The skill gaming market was valued at $46.39 billion in 2025 and is projected at $52.71 billion in 2026, heading toward $121.57 billion by 2034 at 11% compound annual growth. The narrower real money skill segment sits at $25.27 billion in 2026 with a projected 13.92% annual growth rate through 2035.
Set that against mobile gaming as a whole, which Newzoo projects at $121.1 billion in 2026 growing at 6.8%, inside a total games market of $213.9 billion growing 6.1%. Set it especially against in-app purchase revenue growing 1.3%.
Competitive entry-fee gaming is compounding at roughly double the rate of the market it sits inside, and at roughly ten times the rate of the monetisation model it competes with. That gap is the entire argument, and it is not a projection about the future. It is a description of what already happened between 2024 and 2026.
Player preference points the same way. More than 58% of United States players say they prefer games that offer real-money competitions. In a market where the operative problem is depth rather than reach, that is a large expressed appetite for a format most publishers do not offer.
How it works in practice
The design differences are concrete rather than theoretical. Backspin Games operates ten mobile titles, including 21 Jack, Bingo, Solitaire and Cannon Blast. A player selects a game and an entry level with the fee shown before anything is deducted, is paired with an opponent of comparable demonstrated ability, and plays from starting conditions equivalent to their opponent’s. The advertised prize goes to the better performance. The platform takes a service fee on entry rather than holding a position in the match.
Every title also has a free practice mode running the same rules and interface as the paid version. In a free-to-play context that would be a leak in the funnel, since the free tier is the monetisation surface. In an entry-fee context it is a recruitment tool, because a player who has learned the game in practice is a player who might reasonably enter a paid match.
The trade-off, stated honestly
The entry-fee model is not strictly better. It carries a cost that free-to-play does not: the player can lose money. A free-to-play game can waste a person’s time and can extract more spending than intended, but it cannot produce a negative balance on a given session in the way a paid competition can.
That is a real difference and it should be stated plainly rather than buried. Returns in competitive formats depend on performance against the opponent and are not guaranteed in either direction. The category’s defenders sometimes present the model as unambiguously more honest than free-to-play. It is more honest about what it is. It is not risk-free, and the two claims are frequently conflated by people who should know the difference.
Games that pay real money are not multi-level marketing and are not gig work. But any operator in any category making representations about what participants will earn is making a claim it needs to be able to stand behind. The commercially sensible position and the honest position are the same one: describe the mechanism, not the outcome.
Frequently asked questions
Why are game companies moving toward paid competition?
Mobile gaming has largely run out of new players, with around 3.10 billion mobile gamers representing about 84% of all gamers, so growth now depends on deepening engagement with existing ones. In-app purchase revenue grew just 1.3% in 2025 to $81.7 billion, while the real money skill segment is projected to compound at roughly 13.92% through 2035.
Do skill-based games make money differently from free-to-play games?
Yes. Free-to-play revenue scales with time spent, through advertising and in-app purchases, and converts only 1 to 2% of installs. Entry-fee platforms take a service fee on each match, so revenue scales with matches entered and therefore with player retention rather than session length.
Can you buy an advantage in a skill-based competition game?
In properly designed formats, no. Both players face equivalent starting conditions and pairing is based on demonstrated ability, so a purchase cannot change the competitive position. The commercial model provides no incentive to sell one, since the operator earns from matches entered rather than from advantages purchased.
Is the entry-fee model safer for players than free-to-play?
It is more transparent about cost, since the fee is disclosed before each match. It is not risk-free: outcomes depend on performance and a player can finish a match down. Free-to-play cannot produce that outcome, though its own record on retention and on monetising frustration is poor.
What does mobile game retention actually look like?
Median 2026 figures are roughly 22% on day one, 4% on day seven and under 1% on day thirty. Even the top quartile of games retain only 1.6 to 1.8% at day thirty. Strong performers reach 27% on day one and 3 to 7% at day thirty.
The design question behind the revenue question
Every monetisation model is an argument about what a game is for. Advertising and in-app purchase treat attention as the asset and design accordingly, which is how an industry arrived at a median day-thirty retention below 1% and decided that was normal. Entry-fee competition treats the contest as the asset and has to make the contest worth entering, because there is no revenue in a player who has stopped entering.
Neither is inherently virtuous, and the entry-fee model carries a real risk the other does not. But as the acquisition era closes and the industry turns to depth, the models that reward mastery rather than working around it start to look less like a niche and more like where the growth actually is.
One practical caveat applies to the entry-fee model that does not apply to free-to-play. Paid competition is regulated state by state in the United States, roughly 12 states restrict cash skill gaming as of 2026, and Pennsylvania’s Supreme Court changed the position there in June 2026. App store distribution is national; paid entry is not. Availability should be confirmed before depositing.
