Many games look at daily active users to assess their health and performance. However, not all active users contribute in the same way to your bottom line. Furthermore, “active user” can be a very generic term that doesn’t reflect any specific level of engagement and commitment.
When you look at all active users indiscriminately, then you are not focusing on what really matters and impacts your success. Colopl reports its metrics while taking into account only users who have displayed a minimum level of engagement – in their case, a user who has returned to the game at least once after 7 days of install. What I am arguing here is that an even better indicator of engagement is being an active customer. Instead of considering the DAU of your game, considering DAC (daily active customer) might be the right alternative for you. Of course, this will make more or less sense for you depending on your business model. If you are managing a hypercasual game that monetizes primarily through video ads, then DAC is probably not the metric for you. If you are managing a PvP game that relies on having a high number of concurrent active users, you should probably find another measure of activity – perhaps something more gameplay based. The point here is not that you go out of your way to ignore or aggravate non-payers. The point is that if you want to track metrics that are related to success and focus on measuring what matters, then focusing on customers can make a lot of sense. The fluctuations in DAU don’t always impact a game’s overall performance. Looking at active customers provides a much more stable reading on the game’s status. “Customer” reflects a much more stable quality of users and engagement.
Looking at DAU means you are not assessing only the health of your game. You are also assessing the impact of external factors. Specifically, when you look at DAU, you can observe great fluctuations when you have a feature on the App Store or on the Play Store. But more often than not, such an influx of new game users is a temporary – and not meaningful – change. Just in the same way, the decrease in arpdau you observe when you get featured doesn’t reflect game performance, but the fluctuations in active users and their quality. The same can happen if there is a shift in UA strategy – either an increase in overall spending, or a change in geos you invest in. If for some reason you start massively investing in lower tier countries, then that will indeed translate in an increase in DAU. But that increase in DAU most likely won’t be followed by an increase in revenue-contributing customers.
One of the biggest advantages of considering DAC is that it makes things more predictable for you revenue-wise. Past behavior is always the best predictor of future behavior: your customers are your users with the highest propensity to spend. When you consider an active user indiscriminately, then you are bound to see fluctuations in arpdau that reflect these changes in userbase – not changes in game performance. On the other hand, being a customer is a very good indication of engagement. And the average revenue per daily active customer (arpdac) is much more stable. And any fluctuations you’ll observe on that front will truly reflect changes in game performance and content.
Again, this doesn’t mean you need to go out of your way to aggravate non-payers. You don’t and you shouldn’t. But it does means that if you focus your efforts on your active customer base, you will be in a better position to identify the levers that matter and optimize your prices and value proposition. And from that perspective you should embrace the fact that you shouldn’t operate your game with everybody in mind, but your most valuable users.
There are multiple reasons for that. The first one that comes to mind is that entertaining non-payers in the long run simply doesn’t pay off. Monetization doesn’t derive directly from engagement. Your best customers convert quickly, and the longer a non-payer is active in a game, the less likely s/he is to ever convert. If you agree that spending is the result of a high affinity with your game – more than something you create out of nothing – then that also means that a user that doesn’t convert quickly has a low potential to monetize well in the long run. Acknowledging and accepting that is the best way to optimize your resources and development efforts. There will always be a much bigger ROI in trying to get a customer to spend again or spend more than there is to try to get a 30-day old non-payer to make his first purchase (the redeposit rate of late converters is always low). I’m not saying you shouldn’t try to get users to convert later in their lifecycle – but I’m saying you should prioritize your monetization efforts.
Another thing to keep in mind is that most of your revenue comes from a very small percentage of your payers. If you group your payers per LTV percentile brackets (your top 1% customers, your top 10% customers, your bottom 50% customers), it’s very likely that you’ll observe that over 50% of your revenue comes from your top 5% customers. It’s also likely that 50% of your customers (who are already a very small subsection of your installs) don’t even contribute to 5% of your title’s lifetime revenue. Think about it: half your paying users don’t even account for 5% of your revenue.
If you have access to a user table that has your installs and their LTV, then you can run something along the lines of the below query to see what things look like for you (Redshift syntax used throughout).
There is a pervasive idea that you should keep players in your game long enough and eventually they will spend. But as seen above that’s simply not the case. And when you look at things even closer, then you can see that those users who account for most of your revenue convert even faster. When you look at the median time it takes your customers to convert – and more specifically see how long it takes customers in different LTV percentile brackets – then you will observe that the higher the LTV bracket, the faster players convert. And we’re talking a few days at most. It’s very likely over 70% or your IAP revenue is coming from users converting within a few days after install.
If you notice here that the median time to convert for customers in the lowest LTV bracket is slightly less than for that of customers in the bracket above: that might be the result of greater churn for players converting on install day (relatively speaking). But even in that case, the overall assessment remains the same. If your user table also includes install date and convert date, you can run something along the lines of the query below to see what the pattern is for your game.
Because of all this, focusing (both your development and analytics efforts) on your good customers makes the most sense. The reality is the majority of the users installing your game – and simply active on a given day – will not be very engaged. At the very least not engaged enough to actually impact your bottom line. And if you focus on all users, then you are bound to have a lot of noise. Worse, if you design and implement features with everybody in mind, then you risk not doing things for your biggest fans – who are also the ones spending the most in your game. And you risk compromising the value you give and get from your best players to cater to a group of users only moderately engaged. Determining the threshold for relevance is something critical in order to maximize your efforts and implement features that appropriately resonate with your most important players. And the most important players you should be focusing on are your active customers.
Alexandre Macmillan 
11 comments