Ask someone outside the industry to picture the games business and they will imagine a shelf of boxed titles, each sold once. That mental model is a decade out of date. The dominant economic unit today is not a product you buy but a service you keep returning to: a game that ships, then updates for years, monetising attention rather than a single transaction.
The arithmetic of retention
The whole thing rests on one idea. A premium game earns most of its lifetime revenue in the weeks around launch, then decays. A live-service game earns a smaller amount at launch and then keeps earning — for months, sometimes years — as long as players stay engaged. Stretch that tail far enough and the areas under the two curves are not close.
Revenue over time: premium vs live-service (illustrative)
Illustrative. Dashed: premium, a launch spike then a fast decay. Solid: live-service, a gentler curve that keeps paying.
Why it is hard to copy
If live-service is so lucrative, why doesn't everyone do it? Because the model is a capability, not a genre. Running a game as a service means a live-ops team, a content pipeline that never stops, data infrastructure to read player behaviour, and the balance-craft to monetise without alienating. Studios built around shipping-and-moving-on cannot bolt this on cheaply. Many that tried produced a hollow live-service shell around a game that did not need one, and players noticed.
Live-service is not a feature you add. It is a company you become.
The share picture
Put engagement and monetisation together and you get concentration. By our illustrative estimate, live-service titles now account for roughly two-thirds of the revenue among the top hundred grossing games, and that fraction has been climbing steadily. The premium slice is not shrinking in absolute terms so much as being outgrown.
Share of top-100 revenue by model (illustrative)
Illustrative. Live-service dominates the top of the chart.
Where the model runs out of room
Dominance is not destiny. Three pressures are building against the live-service machine. The first is time: a player has a finite number of hours, and every successful service game is competing for the same evenings. The market cannot support an unlimited number of games that each demand to be someone's main hobby. Acquisition costs climb as the good slots fill up.
The second is fatigue. Battle passes, seasons and daily logins were novel once; now they are ubiquitous, and a share of players is quietly opting out of the treadmill. The third is concentration risk for studios themselves: when one live game funds everything, a single bad season can threaten the whole company. That is a fragile way to run a business, however profitable it looks in a good year.
What comes next
Our read is not that live-service fades — it is too economically sound for that — but that the field bifurcates. A handful of enduring titles will keep their grip on the top of the chart. Below them, expect a return of interest in premium and in tighter, session-friendly designs that respect a player's limited time rather than demanding all of it. Studios that can not run a service will stop pretending to, and specialise instead.
For anyone planning a slate, the practical takeaway is to be honest about which game you are actually making. A live-service ambition without the live-ops capability is the most expensive mistake in the current market. This connects directly to the wider structure we set out in our state-of-the-market briefing, and to the financial red flags we cover in how to read a studio's financials.
As with everything on the marketing side of Skelora, the numbers here are illustrative. In Pro, the model breakdowns are sourced and dated, and the estimates are labelled as estimates. The argument, though, holds regardless of the exact percentages: the revenue chart in games is a chart about retention.