For better or worse, much of the games market is moving to games-as-a-service. Once upon a time, this was known as the MMO business model, because all MMOs were games-as-a-service, and virtually nothing else was.

Obviously, a lot of GaaS games won’t be MMOs. Candy Crush Saga is a service, but it’s not an MMO. Nor does GaaS mean the game has to be free to play with upsells via microtransactions. GaaS is a business strategy, and F2P is a revenue model.

In the past we’ve seen services of all sorts drive revenue in lots of ways:

  • free sampling to get you in the door, supported by other revenue streams such as other paying members subsidizing free users, or advertising
  • a la carte offerings which let you dip into a service without an ongoing commitment
  • periodic upgrade offers to take you from that free tier to something with recurring payments
  • subscription tiers — a low basic offering and a few richer ones catered to specific demos, which retain the ability to a la carte for specific features
  • individual one off upsells for special events that are never covered by the sub tier

In fact, cable companies and phone service both do just about all of the above at once. So there’s lots of ways to make money, if you have a service, and arguably, mature service businesses use as many of them as they can.

But underneath it all is one inevitable underlying truth.

There is no ongoing service without retention.

This is why some game genres work better than others as services. If you can figure out how a given genre can retain, then you can make it work for GaaS. Whether or not it works for F2P is a separate, secondary question. And a lot of game types are built to be consumable, snackable, or played very intermittently. By their nature, they will work poorly as the sole anchor for a service (they might work great in the context for a service that has multiple offerings).

Retention boils down to two key measures:

  1. The game holds people a long time — meaning, across sessions.
  2. The game is something people want to play again at least once a month.

Even something like a Dungeons and Dragons campaign meets these criteria. Heck, many classic tabletop games meet this criterion, or else people wouldn’t still be playing chess and go.

But an awful lot of games don’t. What we’re talking about is the line between a singular experience and something that can become a hobby, a regular pastime.

What’s the difference, in practical terms? Take two games. They sold or gave away the exact same number of copies; the blue is the cumulative number of users who tried the game. Game A was popular, but for a short period of time; each player lasted a day. Game B retained players; 90% of players stick around to the next day. The revenue potential is the area under the red curve, which is the game’s active users. Game B has access to a lot more passion, a lot more interest, and therefore also a lot more money.

The difference in active users between 100% and 90% churn shows a huge increase under the curve for the lower churn rate.
The difference in active users between 100% and 10% churn

OK, so how?

The following mechanics are proven to drive retention. This is a non-exhaustive list, but covers most of the bases, I think. These are mostly things you do with a game over time, in order to drive people to play it again, or turn it into a hobby.

A steady content trickle

Pro: Consuming content as you move through a system can be very compelling. It could take the form of defeating, collecting, plot advancement or more. It gives a clear aspirational goal for the session, and likely for the next session, and can hold people as long as the content keeps ringing changes on the basic mechanics. Leveraging this is how puzzle games were “saga-fied” and became viable as services. Even blowing up the current game balance and shifting it around can accomplish this (though it’s risky!). Also easily monetized; most F2P games basically rely on constantly offering users more “stuff” without actually changing the game very much.

Con: Expensive. Like, very expensive. Like, people underestimate how expensive. And if you are solely dependent on this and miss an update you can actually lose your audience. Regular updates is best practice nonetheless.

Persistent profile investment

Pro: A familiar approach that is easy to implement, usually via RPG-style mechanics. Even piling likes into a profile counts here; it’s basically about having a persistent profile that accrues over time. Leaving would mean the loss of the accrued value, and more importantly, the standing in a community.

Con: Works best if there’s a community to display the profile to, which you then have to build and run. Also, you tend to add new content at the top, which can alienate new players, gradually over time, via a host of mechanics ranging from mudflation to game overcomplexity. Lastly, real community prominence is more valuable than the level number proxy you give it; if the player can move to another setting and retain that prominence, this may not serve as a retaining anchor.

In-world investment (building, housing, etc)

Pro: This is what save games in building games do. Unlike character investment which can be sort of portable, this can never leave the game. It is also even more powerful if collaboratively built, because it then implies also building a set of social ties. It also serves as an outlet for a couple of other methods listed below.

Con: Expensive to design and implement, data-intensive to store, and it introduces a host of design challenges around space, public display, and more.

Social connections, such as teams, guilds, etc

Pro: Social groups are the primary glue in games in general. Even single-player games have huge social characteristics to them around widely shared experiences and common ground. Social ties introduce a host of extremely powerful things like mutual obligation, economic exchange, group identity, and so on (see my old talk on social mechanics for tons more). Looser connections and community can often work better than a tight-knit community.

Con: Guilds often migrate games as a whole, so you want your user tied to the community via multiple touch points that aren’t in the same guild. Social connections also bring drama, which means community management, moderation, and much more. There’s a vast amount of expertise involved in engaging in governance here, and even though this is arguably the most powerful tool in the arsenal, it’s also very challenging and causes burnout in staff on a regular basis.

Economic play and arbitrage

Pro: This might happen with real money, or it might happen with play money. The profit motive is incredibly powerful and will keep players engaging in your experience well past the point of actual enjoyment. Regardless of whether it’s real or fake money, though, the fact that it is effectively a self-generating set of ongoing challenges means that it’s a substitute for content. The game keeps refreshing itself in somewhat unpredictable ways.

Con: If using real money, it can quickly chase out more playful ways of engaging. Either way, it requires a robust set of systems that enable trade, sales, and so on, and dynamic virtual economies are much harder to design and balance than static ones.

Extreme game depth

Pro: This is of course the classic method used by chess and go, but also games like StarCraft. A game that has enough depth and complexity to it that players continually see new heuristics ahead, new ways to improve, and effectively never see the game as boring, is in many ways the Holy Grail of game design, particularly when paired with multiple ways to play (variants, restarting with alts, classes, speedruns… lots of ways to do this). A rule of thumb I have used for over a decade is “it’s a good sign if your game merits a player-written strategy wiki.”

Con: Are you a good enough game designer? Few are, frankly; this can feel like capturing lightning in a bottle. Also, be aware that high skill ceiling often doesn’t play well with accessibility.

Player vs player competition

Pro: Other players are a free source of depth. This has also been a default tactic for literally centuries, so much so that it has merited being called the “orthogame,” almost the default historical format.

Con: Watch out for zero-sum play (one winner, one loser) causing players to be chased out. The typical user loses more often than they win, because the most skillful players tend to take up a disproportionate number of the wins. The result is that historically, most services based solely around this managed to only get 1/10th of the population of services based on cumulative character mechanics like RPGs. Today we’re seeing a melding of competitive play with other items on this list, such as content trickle, which hugely ameliorate this. There’s also the current common tactic of offering the game as a non-service and having a service only for the higher echelons of players.

User creativity

Pro: It can be in various forms. It can also be surprisingly cheap to build. Extra powerful if it can be monetized by the player, but powerful even if not, as long as user creativity has a publicity channel that garners audience.

Con: It can also be crazy expensive to build if you’re not careful. Also, creativity depends on audience to a very large extent, so you must have infrastructure to support sharing, showcasing, and so on, in order to drive the social proof and the acclamation that are underlying motivations to engage in the behavior. Because of this, creativity often works best when in the setting of a social network that valorizes it.


Pro: The retention tactic of the soap opera: emotional engagement (typically with characters, not plot or setting) via ongoing narrative. Fortunately, writing, though hard, is also an ancient discipline and the expertise is out there to do this, is games can only be persuaded to leverage it. This also gains leverage when combined with community: if deep enough via lore, easter eggs, and pockets to explore, it can be strongly retentive. Can also play well with regular content trickle; in fact, the classic MMO expansion was both a content dump and plot advancement. Today seasons are a modern equivalent.

Con: It has similar issues to content trickle (which should be defined as consumable gameplay or content, as opposed to narrative). Characters can be hard to shoehorn into a lot of game genres. Just don’t rely solely on episodic unless you are positive you can hit a release cadence, because when the story arc ends, often so does your retention.

Emergent play

Pro: arises out of systemic depth, simulation, etc, and usually interacts with game breadth as opposed to depth. Don’t think having one chess system; think having many smaller systems that are individually not as deep but which interact with one another in surprising ways. It can be surprisingly cheap to implement a system like this.

Con: It can also be brutally hard, particularly to balance, and only some designers seem to have a knack for it. It can lead to enormous community outcries, and your game moving in totally unexpected directions.

In the end

Many of these play with each other very powerfully, but can also exist independently. Like, if you have rich sim, players will create emergent content, which then you can turn around and leverage into your broadcast narrative, and tie content releases to. Emergence tied to a shifting economy and creativity leads to rich places. And so on.

Once you have retention, you can worry about how to make money. If you can’t make money from a userbase that has decided to make your game into a lifestyle choice, well, you’re not trying. Again, it doesn’t imply a particular business model: a service-based game is not a dirty word, doesn’t mandate constant moneygrubbing, doesn’t mean it has to be free to play. It just means that you the developer and you the player are in it for the long haul.