How to think about retention: From early validation to long-term game health

Retention is often called the most important metric in games, and for good reason. It tells you whether players come back, whether your game keeps its promise, and whether there is a real foundation for monetization and long-term growth.

However, retention is also one of the most misunderstood metrics, and teams often treat it like a single pass-or-fail number, when in reality it needs to be interpreted in context: by genre, by monetization model, by lifecycle stage, and by the kind of game you are trying to build.

This is exactly what Nikolaj Ahlberg-Pedersen (CPO, GameAnalytics) discussed with Michail Katkoff (Deconstructor of Fun) in the third GameAnalytics Masterclass. You can watch the full masterclass here.

One of the clearest definitions from the conversation is also the simplest: retention is “the amount of players installing on a specific day that returns a certain amount of days later.” That sounds straightforward, but what matters is what that curve means over time.

Retention is not just one number

Most teams know the shorthand benchmark: 40/20/10. That means roughly 40% day 1 retention, 20% day 7 retention, and 10% day 30 retention. It is useful as a baseline, but it is not universal.

As Nikolaj points out, different genres and business models produce different retention shapes. A hybrid-casual game may have a much sharper drop and still work as a business. A more social or habit-forming game may start lower but hold much more strongly over time.

That is why one of the most useful lines in the discussion is this: “It is a good baseline to go for 40/20/10 and just see.” In other words, treat it as a reference point, not a law.

Retention only matters in context of the business model

A game does not need the same retention curve as every other game. It needs a curve that makes sense for how it plans to make money.

This is one of the strongest concepts in the masterclass. Retention should not be read in isolation. It has to be connected to:

• ads versus in-app purchases
• scale expectations
• content cadence
• cost of user acquisition
• long-term revenue per player

A game with weaker long-term retention may still work if it monetizes quickly through ads. Another game may need a much stickier curve because its value comes from payer retention and longer-term LTV. That is why retention is both a product metric and a business metric.

Early retention is about validation

In the earliest stages, retention is less about optimization and more about proof. Teams need a flow of fresh users to test whether the game actually works. They cannot just keep retesting the same people, because the experience is no longer new. That is especially important when testing onboarding, tutorial flow, and first-session design.

Nikolaj makes a useful point here: if you cannot move the retention curve early and you cannot build a business case around it, then you likely have a problem that goes deeper than tuning. That is why early retention should not be treated as just another dashboard number. It is one of the strongest signals a team has about whether the core experience is viable.

Cohorts make retention more useful

The masterclass also reinforces an important best practice: retention should be viewed through cohorts, not just as one blended number.

Cohorts let teams compare:

• players from different acquisition channels
• players from different builds
• players entering during different feature or season states
• players exposed to different monetization setups

This is where retention becomes much more actionable. Instead of asking whether the game is retaining “well,” teams can ask:

• did retention improve with this build?
• are these new users better than the last cohort?
• which channels are bringing users who actually stick?
• are new features helping or hurting the retention curve?

That is also why the discussion ties retention so closely to LTV. A cohort is not just about who stayed. It is about what that group became worth over time.

Payer retention matters, but only after conversion

Another valuable point in the conversation is that payer retention should not be viewed as a completely separate universe from general retention.

Yes, paying players are especially valuable. They are often the ones who continue spending and contribute most to long-term revenue. But before a player becomes a retained payer, they first have to become a payer at all.

That is why Nikolaj reframes the issue: the goal is not just to keep payers. It is also to understand “how good are you at converting your players?”

So retention strategy needs to consider both:

• general retention, because players have to keep playing to ever spend
• payer retention, because long-term monetization depends on those players staying active and engaged

LTV and retention should be read together

One of the clearest business takeaways from the masterclass is that retention becomes much more powerful when paired with lifetime value. Retention alone tells you whether players are staying. LTV tells you whether that staying power is economically meaningful. That is why cohort-based revenue tracking is so useful. If a specific week’s install cohort generates stronger long-term value than another, that gives the team a real signal about what changed — whether that was the channel, the creative, the build, the season pass, or the game state.

As Nikolaj explains, “It’s super important to look at the LTV of these different cohorts to see is it going down, is it going up, am I doing something right or wrong?”

That is also why retention is so important for UA. A game may not recover its acquisition cost in the first week or even the first month. But if strong cohorts keep retaining and spending, the business case can still work.

Novelty helps, but proven systems reduce risk

The masterclass also touches on the tension between novelty and proven design. Teams want something fresh enough to stand out, but also stable enough to de-risk development. That usually leads to a middle ground: keeping the proven bones of a genre while introducing a narrower point of novelty in theme, progression, or presentation. Retention is one of the best ways to test whether that balance is working. If the concept feels fresh but players do not come back, the novelty may not be enough. If the game retains well but does not stand out in the market, the team may need to improve how it presents that novelty.

Late-game retention is mostly about content and stickiness

As the game matures, the retention challenge changes. At this point, the question is no longer whether players understand the game or whether the core loop works. It is whether the game keeps giving them reasons to stay.

That usually comes down to:

• content depth
• progression systems
• competitive or social loops
• new things to collect, complete, or buy
• ongoing goals that make the game worth returning to

Nikolaj summarizes this simply: “It is basically content of the game.”

That content can be direct, like new levels or new heroes, or indirect, like async multiplayer, guilds, and competitive systems that make the game socially sticky. In either case, the goal is the same: to stop the retention curve from flattening because the player has run out of reasons to care.

A/B testing helps, but only when the effect is real

The masterclass also gives a useful view on A/B testing in retention work. Yes, testing matters. Yes, even surprisingly simple changes can affect retention. Nikolaj mentions examples like major color changes or changing the order of store packages. But he also cautions against overreacting to tiny differences.

If the effect is within the margin of error, it should not drive a decision. If the effect is real and large, it can be highly valuable. That makes retention testing less about constant motion and more about knowing when a result is strong enough to trust.

Final takeaway

Retention is still one of the best ways to understand whether a game has real potential, but only if it is interpreted properly. It is not a single benchmark to chase blindly. It is a way of understanding:

• whether the core experience works
• whether the business model makes sense
• whether the audience is the right fit
• whether content is keeping players engaged over time
• whether acquired users are actually becoming valuable

The simplest line in the masterclass may also be the most useful: “Retention is a symptom of something not working well” when it starts fluctuating or breaking down unexpectedly. That is what makes it so powerful. It is not just a KPI. It is an early warning system and a long-term health metric at the same time.

If teams treat it that way, alongside cohorts, LTV, and content strategy, it becomes far more than a dashboard number. It becomes one of the clearest tools they have for making better game decisions.

FAQ

How is retention calculated in games?

Retention measures the percentage of players who installed on a given day and returned on a later day, such as day 1, day 7, or day 30.

Is 40 / 20 / 10 the ideal retention curve?

It is a useful rule of thumb, but not every genre or business model should expect the same curve.

Why should retention be viewed in cohorts?

Cohorts make it easier to compare players by channel, build, feature, or time period and understand what changes actually improved performance.

Why is retention important for monetization?

Players have to keep playing in order to spend. Stronger retention usually improves the long-term value of the player base.

What is payer retention?

Payer retention refers to how well paying players continue returning and spending over time.

What affects late-game retention the most?

Usually content depth, progression, social stickiness, and having enough reasons for players to keep coming back.

‍