Retention is king, right? Not always

(Mark Baldwin) #1

For those who don’t know, my background is in computer games and my current role is with mobile games and the free to play model. We are having a very interesting discussion with a games publisher about retention. I know that this will not be relevant to everyone, but hopefully I can make a point about being obsessed with hitting certain arbitrary targets.

Most free to play games companies measure player retention at day 1, 7 and 30 because the longer you keep your players the more value they add and this theory is broadly applicable to all communities. Retaining members/players should be more important than recruiting new members right? Maybe not.

Take for example a game that gets 10,000 downloads a day and has a 30 day retention rate of 20% over a game that has 50,000 downloads a day and a 30 day retention of 10%.

Now a publisher might be saying, 10% retention is not good enough, it needs to be higher. We are obviously attracting a lot of people, but not enough are sticking around, what can we change? But does change really need to happen when the overall figure is better? In the above scenario, the game with better retention is less profitable than the one with poorer retention.

So the actual questions should be, why is the game with poorer retention getting more downloads? and how can we get more downloads for the game with the higher retention? and in an ideal world we would want to look at how to further boost retention without harming downloads.

Now the reason for posting this is that we are currently chatting with a publisher who is obsessed with hitting retention targets and this obsession is going to end up costing them a lot of money in the long run. Retention is important, but it’s certainly shouldn’t be the only thing to consider. Sometimes we need to step back and see the bigger picture, a lot of people can be obsessed with making sure their community hits certain targets and this obsession could ultimately harm the long term prospects.

(Robert McIntosh) #2

I know nothing about the video game market, other than as a user, but surely all of this misses the point if it is not focused on the commercial measures?

What is the revenue model in this case? Advertising? In-game purchases?

Should the measures not be ‘time playing the game (over time)’, or ‘purchases’?

As a user of many free games, and as a father who rarely allows his kids to purchase in-game extras, I always wonder why anyone would bother trying to keep me playing if, over the course of the last few days, weeks or months I have not given them any money? They’re only continuing to develop and support a game that isn;t generating revenue.

I suppose that as an engaged user I am creating a more attractive space for other users who might be better value than me, but still … I should be being pushed out after a while (if I’m being hard-nosed about it)

The only thing I can think of is that I see the occasional banners that are so small they have ZERO effect and only get clicked on by accident.

If you can show how retention is directly linked (or not) to revenue, then maybe the targets will make sense.

Or am I missing the point?

(Richard Millington) #3

I was playing around with a model of this for fun. I was in a rush and made a few mistakes though I think.


Ultimately it depends if you include the total user base within the churn rate. If you do and you lose 10% of your total user base each time…you have a big problem. If after 30 days the churn rate shifts (as is likely), then you would need change it.

The key thing though is to look past just 30 days (unless that’s a payment date) and consider the long-term. Retention rate is important because if it’s bad you’re burning through a limited asset. The number of people who will download a game (or join a community) is finite.

So when you push on the marketing side at the expense of retention, you’re simply getting people faster. This is great in the short-term. But it’s terrible in the long-term if retention is too low. You simply burn through everyone (see Pokemon Go) and run out of newcomers.

So in the short term, you’re right (I think the numbers are cherry-picked a little). But in the long-term you probably need a more complex model than the above to see. Retention is a pretty critical measure of whether people like the game though.

(Mark Baldwin) #4

The vast majority of free to play games rely on advertising to make most of their revenue, so adverts will play on a regular basis and the developer gets their money this way, so getting high numbers playing is key to making money. Keeping hold of those players is important, but over time if you are growing your player base as in getting more people daily than you are losing then you will make more and more money. This is why the top grossing free to play games spend a lot of money on acquisition, there is a huge churn but they are constantly cumulating.

(Mark Baldwin) #5

I wish I could share some real world data with you, but I can’t. The figures I mentioned were just to make a point. We constantly measure churn on our own titles and are constantly changing strategy, but publishers who constantly turn out games are always looking for short term gain and will concentrate on acquisition. This is how titles at the top of the grossing charts stay there.

You would think so, but people are constantly upgrading their devices, going back to games they used to have. How many people for example as soon as they get a new device go and download candy crush/angry birds. Also with so many young people with devices the market is constantly expanding. One of our games has been out for 4 years now and still gets a similar amount of daily downloads, but yes, when you are making a game that you still want people to play in a year or two you need a long term strategy to bring in new players but also to keep existing players happy and engaged.

We would love a little slice of that revenue right there, even after the drop off of new players. :slight_smile:

(Mark Baldwin) #6

@richard_millington you might find this interesting. We work with flurry on analytics and this is a brief article with some interesting figures. You will see that retention varies massively depending on type of app.