Cohorted and Live Ops approaches

In some of my previous posts (for example, here) I have discussed how there are different ways to approach game performance.

On one hand you can focus your attention on the journey of a user. When you are looking at cohorts, you are considering an aggregation of users from the install moment onwards. This is a forward-looking perspective. On average, how much revenue will you get per user 7 days after install, 30 days, 360 days, etc.? On average, what percent of users will be playing your game 30 days after install, what percent of users will have done an IAP within 2 hours of install, etc.? In this case you are looking at the high-level picture, and not focusing on the specifics of user behavior or activity. Said differently, when you are looking at LTV, you are looking at the raw output for you, considering some users will be active, and a majority will not. Considering what your input is (most of the time, your UA investment and CPI) what is the outcome you can expect (the LTV of the users)? Whether your user is active or not is part of the black box when you’re at that level of thinking (but of course, the retention of your game will be reflected in your LTV).

On the other hand, with a Live Ops focus you are considering active users. This is a game-centric, “now” perspective. You don’t consider where users come from, or how long ago they installed and where they are on their install journey. Rather, you are just interested in the fact that on a given day, some users show up in your game and interact with the features you have put together. And the users who show up in your game 2 years after launch are probably not the same kind of users showing up in your game 2 weeks after launch. Here, the output you are looking at is actual revenue on a given period. On a same day a 7-day old user and a 30-day old user might spend. What will matter for you is the revenue on that day – not how that spend fits into the user journey.

When you are thinking about your input/investment, you want to make sure you have a positive ROI after a given amount of time. What’s important is the output x days after the investment. The behavior/engagement/retention is part of the black box that ultimately underlies your LTV and the potential profitability of your marketing campaigns. That’s why you are looking at installs. On the other hand, when you are thinking about the revenue generated by the game, you need to be thinking about who is actually active and showing up – you need to be in the game to spend. That’s why in that case you are looking at active users.

 

Why is this distinction important? Growing the game and growing revenues

If you’re managing a live game, you are probably looking to improve monetization by all means possible. But the Live Ops/Cohorted distinction is important because improving the monetization of your active users doesn’t necessarily mean improving the LTV of your users. More specifically, you might be able to increase (even significantly) the revenue generated by your game. But that doesn’t necessarily means you will also be able to grow your game’s userbase – and ultimately grow your game in the long run. If OKRs are important to the way you work, you want to make sure you are identifying precisely the levers you are trying to move (and why those are the levers with the most leverage).

Differently stated, when it comes to optimizing your monetization, you won’t approach things the same way if you’re trying to improve the monetization from a cohorted or Live Ops point of view. Improving your revenue will involve improving your arpdau – and doing things with your active users in mind – while (profitably) growing your game will involve increasing your LTV and giving more ammunition to your UA efforts. And improving the aprdau – even significantly – won’t necessarily have a similar impact on LTV. What that also means is that you won’t be doing the same things if your goal is to grow revenues or to grow the game (read, scale the game in terms of installs).

To illustrate this, you can look at where your revenue is coming from. More specifically, what is the “age” of the players spending in your game. I’ve discussed previously here why average age is probably not the best metric to use. To keep things simple, you can identify the percentage of revenue coming from players who installed 60 days ago or more (of course, look at your data to determine what the appropriate time is – is 60 days the relevant cutoff for you, or 7 days, 30, 90,etc.?). You will probably be looking at something similar to the graph below.

 

This linear increase reflects an ideal situation with no external factors (for example, a huge feature accompanied by a large influx of new users and customers) and a stable UA strategy. The longer the game will have been in the store, the older the players showing up on a daily basis. As a result the longer the game will have been in the store, the more your revenue will come from older players (your biggest fans will be the ones that stick around the longest and spend the most – over time less engaged users will naturally filter themselves out and the quality of your active user base will gradually improve). At some point you’ll most likely stabilize at a very small percent of your revenue coming from new players.

As your title ages you’ll be left with a larger portion of your revenue accounted for by an older user base. Also, important here is that the activity patterns of newer players is very different from the activity patterns of players who have been in the game for longer. An older active users will be more likely to spend than a user who recently installed. This is not because users get better as they age. It’s because less engaged users will naturally filter themselves out of the game (by not playing) as time goes on. In order to illustrate that, you can look at the arpdau of active users grouped by user age (said differently, days since install). I’ve discussed here why you will probably consistently see a spike in day 1 arpdau.

 

Why increasing your revenue won’t necessarily improve your LTV

This “filtering” out process is the reason why growing your daily revenue – even significantly – won’t always impact your LTV and your ability to grow the game by getting more installs in. On the one hand arpdau will improve as an active user ages. However, on the other hand, as time goes on fewer and fewer users will be active in your game. Ultimately LTV is a function of the arpdau per days since installs, as well as the retention per days since install (and I’ve already discussed here which one of the two is the most impactful for you).

So, if you are doing something to improve the monetization of your older users (for example, adding some late game content or balancing an event), you can easily get a sense of how much (or little) it will impact your LTV.

Here, you can retrieve for your game the retention and the aprdau per days since install. You should get something like the table below.

Once you have these numbers, you also have your cohorted revenue per day (retention X arpdau) and by extension your cohort’s LTV (the cumulative sum of your cohorted revenue per day)

With these numbers, you have what you need to assess the impact of an increase in monetization of your active users on your LTV. Say you want to add an end-game feature that will increase by 50% (or even double) the monetization of your active users – who might be mostly older users depending on the age of your game. You can reproduce the above table for your game and see what the impact of your LTV would be. Most likely, increasing your arpdau won’t have that big an impact on LTV (in the below example, I imagined the new feature improved the monetization of users at day 180).

 

Again, this is not to say that you shouldn’t try to improve the monetization of your active users (because you definitely should). The point here is to emphasize the fact that you have to be very clear on what your monetization goals are – and intentional in the monetization strategies your are implementing in your game. Increasing your later revenue won’t always increase your LTV.