Days to convert and churn

Generally speaking, users converting on install day will be your best payers. There is another general phenomenon about monetization on install day. When you look specifically at you users converting and the day they convert, churn is the highest for users converting on install day. From a methodological point of view, I have some issues with the very notion of churn (I’ll try to discuss that in the near future). But if you go along with it and define churn as not returning to the game for a given period of time (say 2 weeks) you will see a higher % of users who convert on day 0 will not return to the game within 14 days of their first IAP purchase. A much higher percentage.

What the above graph is telling us is that close to 10% of the paying users who are making their first IAP purchase on install day don’t return to the game for the following 2 weeks. That percent drops, then is relatively stable. Converting users churn the most when they convert on install day. Not only is there a higher churn rate for users converting day 0 – compared to users converting later. The churn rate itself can be actually quite high in of itself – depending on the game, the 14 day churn rate of users converting day 0 can hover around 10% in tier 1 countries.

The important thing to keep in mind is that relatively to other user groups, the churn rate is not necessarily different on install day (although it can be). Users converting on install day churn more than users converting on day 1 or day 7. But on the other hand, churn is the highest for all users on install day – a large portion of your userbase will install your game, give it a chance, and never return. So, if you consider there are 3 mutually exclusive types of users who can be active on any given day (active nonpayers, active users who convert on that day, and customers – active users who have converted in the past) then churn of users converting day 0 is not necessarily worse than the churn of nonpayers (again, relatively speaking).

What strikes me as interesting here is that having close to 10% of users spending on the day they install churning does not fit with the current narrative regarding monetization in the mobile F2P industry. The prevailing assumption is that spending in a mobile game is the highest level of commitment possible. It might even seem at odds with my previous posts which discussed how install day is the day most conductive to conversion, and how users converting day 0 are the best payers. But it’s not. I think that a) day 0 being the day where users convert the most, b) having users converting on day 0 being your better payers, and c) observing a high churn rate of users converting on install day should be brought back to a same dynamic that defines key monetization patterns in mobile games. And it has to do with common motivations and behavior patterns among users converting soon after install – it’s not a game-specific thing. The decision to spend in a mobile game soon after first playing it is primarily emotional, based on enthusiasm. The decision to spend real money in a free to play game is not the result of a rational cost/benefit calculation. a) users are likely to be the most enthusiastic when they just download your game. They are still discovering it and everything is new, b) users who have a strong positive emotional connection to your game are likely to be the ones who will engage with it the most – and that also means they are more likely to spend and continue spending. It therefore seems like an inevitable corollary that c) enthusiasm will at times be accompanied by changes of heart. If we say that users converting on install day are more enthusiastic than users converting  later, then you are likely to come across more users who decided to spend in your game, and soon after decided they didn’t like it as much anymore.

What are users converting on?

One thing to keep in mind is that all users don’t convert on the same bundles. Specifically, if your starter pack has a 24-hour timer on it, you shouldn’t have any users converting on that bundle 7 days after install. Looking at churn per conversion bundle can be a good way to identify the bundles that are not satisfying for your users – and optimize what you are selling.

In the above graph you see that users converting on bundle 1 churn much more than users converting on bundle 4 (or any other bundle in that example). Sometimes it can be due to the price point of the offer – the lower the offer, the more likely you are to attract low potential payers. But even when comparing at same price points, you see that some bundles are more associated to users churning than others. That’s a good indication to start looking at content and/or the placement and timing of the offer. When you look at churn per purchase, then you might be able to identify relatively underperforming offers.

The important thing to keep in mind here is that a higher churn on install day doesn’t have to do with the type of bundle purchased. If you look at churn for users converting on a same bundle, you’ll clearly see that churn is highest on install day. So while there are relative performance factors (between one bundle and another), the most important dynamic at play here is the time it takes to convert. Converting users churn the most when they convert on install day.

What to make of this?

What are the scenarios that can help us make sense of this higher churn rate for users converting on install day? One potential interpretation could have been that users who converted on day 0 were disappointed with what they purchased or had buyer’s remorse. And their response was to not return to the game. However, when you look at churn rate per conversion day – and look at users who converted on a same item – you see day 0 payers consistently churn more. So that pretty much invalidates the “disappointing value” hypothesis. There is no apparent reason to be more disappointed by a same purchase earlier on rather than later. If anything, relatively speaking a same purchase should provide a greater advantage earlier on. In an RPG getting any given character on install day is likely to have a much greater impact than getting the same character 14 days after install. In a simulation game, buildings are cheaper earlier on – your $1.99 will get you more on install day than 7 days after install. The same applies for pretty much any game or genre that has a progression system.

The other potential explanation – more simple, intuitive and accurate in my opinion – is that in the grand scheme of things, spending $5, $10 or $20 (or even more) on entertainment is probably not experienced by paying users as being such a commitment or investment. So out of those fans who installed the game and quickly liked what they saw – and wanted to spend money to engage with the game further – some of them maybe did not feel they liked it enough to come back and play it again. That’s probably the inevitable and inconsequential corollary of generating enthusiasm and high emotions in a free-to-play mobile game. Those users have a level of engagement and enthusiasm high enough to spend on install day, but not high enough to come back to return to the game the next day. More importantly, the existence of such a large segment of payers churning on install day strongly suggests we shouldn’t presuppose that spending in a mobile game is the highest form of engagement. Many users probably value their time more.

Finally, the existence of such a large segment of payers churning on install day – combined with the facts that users convert more quickly after install and that users converting earliest end up being your best payers – strongly suggests that enthusiasm (a positive emotion) is the main driver of conversion on install day (which is associated to higher quality customers). So not only do you need a monetization strategy as soon as your users are finished the tutorial. Every monetization strategy for install day should focus on generating as much enthusiasm as possible – by showcasing what makes your game unique and great – and capitalizing on this enthusiasm. Not creating friction.