Retention 101

In The Player Lifecycle post I wrote about the story I want to tell. That story starts with Acquisition so that would be the perfect starting point… except it isn’t! The reason being that if it exists one and one king only, that king is Retention.

So… what is retention?

Retention is the game’s ability to keep the players returning. It addresses only their return. It is not concerned if what they did in the game is meaningful.

In quantitative terms, the retention on day X is the percentage of players of a given cohort that returned to the game X days after they started playing it. As an example, let’s say that on January 13 we had 100 new players. In game analytics jargon we say that on January 13 the cohort size is 100. On January 14, 35 players that started to play on January 13 returned. We calculate 35/100, so 0.35.

In this case we say that day 1 retention is 35%.

Several things are going on here. In case you need, here’s a handy glossary! Let’s dive in.

First and foremost our population of interest is the cohort. A cohort is defined by their acquisition date, which is the day they ran the game for the first time. The cohort size is the number of players of that cohort. When we count the number of daily active users (DAU) in subsequent days and divide by the cohort size, we get the retention rate which we usually express in percentage for ease of reading.

So we can say that:


Where X is the difference in days between the date of interest and the cohort acquisition date. DAU day X is the number of daily active users on day X and the cohort size is the total number of players in the cohort.

This is “classic” retention. There are other formulas and I’ll return to them on a later post but this is how retention rate is commonly calculated.

Why is retention important?

From a Player Lifecycle point of view a player that doesn’t return to the game is the end of the story. It means that her relationship with the game is over. The player has churned.

For games that don’t have extra monetisation elements (paid DLCs and IAPs come to mind) not retaining a player means that all the hard work that was put into the content of the game was not fully appreciated by the player. It is sad and it can have consequences to the studio’s reputation if the game is overall bad but commercially speaking the immediate impact is low.

For games that do have these elements, namely freemium titles, this is much serious. It means no players to purchase the game’s products and content and no players to bring other players into the game. It means, quite simply, no monetisation and no virality. It is disaster.

A low retention is always a sign of a game with problems.

“Good” retention standards

When someone says “good retention”, what are they talking about? How high is the retention rate to be considered good retention? If you want my opinion good retention standards are like statistical models, all of them are wrong but some are useful. The reason why I think this, is that no retention rate is high enough. If you have a D1 Retention Rate of 99% you are still losing 1% and that’s 1% too many! On the other hand I read a white paper (would link it but it was more than two years ago) that stated that the average D1 Retention Rate was 11%. Picture that… 11% means that, on average, 89% of the users of mobile apps don’t return to an app they were interested the next day.

We can argue that there are players that didn’t like the game or expected a different experience and those will always churn on day 1. I wonder to what degree this is still our fault, be it the way the game is presented in the app stores or the players that we acquire via UA. But assuming these are unknowns that we cannot control for, what are acceptable retention values?

Some companies state that if a game doesn’t reach a 50% D1 Retention Rate, they won’t invest in further development after launch. This is mathematically understandable. Even if the game loses all its players on day 2, the game will maintain the number of daily active users as long as the number of new users is sustainable, which, with UA, it is. Even if it isn’t, no game with 50% retention rate on day 1 will have 0% on day 2 unless is doing something so wrong that I cannot even imagine what it could be.

My personal favourite is the 40-20-10. What this means is 40% on D1, 20% on D7 and 10% on D30. The reason I like it it’s because I’m a big defender that what makes the difference on any successful game is that players play it for a long time, meaning a high long tail retention rate. Unlike 50% D1, 40-20-10 allows us to understand if the retention is up or down a number on 3 specific milestones, first day, first week and first month.

Closing thoughts…

Retention is a metric that you can observe throughout the player lifecycle and the game’s life too. Although it is quite simplistic, it is quite powerful to assess how well the game is doing. Retention correlates strongly with many KPIs, including monetisation ones which allows to build quite strong prediction models.

I hope this was a good intro to mobile retention analytics. There’s much more to cover and I intend to add more content regarding King Retention.

5 thoughts on “Retention 101

  1. Ricardo, might you recommend a book or some other resource on the topic of game analytics?
    I am mostly interested in success/fail stories, in which are thoughts and methods (to gain insight) of analysts on how, why and where the analytics helped or didn’t at all 🙂
    I also welcome any other resource for a complete beginner.
    Thank you for sharing your knowledge!


  2. Hi Aleksandar and thank you for your kind words.

    Game analytics wise, not that I know of. I preordered and read this book

    I thought it was great… Knowing what I know now I wouldn’t really recommend it. Some interesting insights but very little to build on. I’d gladly recommend another if I find one but it is true that I haven’t searched much more.

    To be honest it was the lack of information that led me to write this blog in the first place. What I learned, I learned with other people in the industry, from information my colleagues would bring and that I had to search and learn and ask and build. I tried to create a community of game analytics about a year ago and there aren’t enough of us around the world to keep it moving, not on a forum kind of thing.

    There are some interesting blogs, most from analytics services. I’ll gather and make a post… that actually might be useful for a lot of people! 🙂


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s