The following article is based on a previous one we published on the GameAnalytics blog: 16 reasons players are leaving your game, which you can find here, written by Nathan Lovato. This article has been trending on our blog, and it’s a hot topic among developers who are looking to optimize their user experience. Considering this, we’ve thought to complement Nathan’s piece with examples of how you’d use analytics to spot and fix the problems listed.
Share the post with this ready-to-go tweet 🙂
To make sense of it all, we’ll follow a similar structure to what Nathan used: listing the reason and the way you can investigate it through the use of analytics. Note our numbering will differ as we won’t be discussing all the reasons.
If you didn’t get a chance to read the 16 reasons…, do so now – we’ll wait for you to get back. However, if you want to cut right to the analytics chase, we made it so that this can also be read and understood as a stand-alone.
Here we go!
1. Your game intro sucks
First things first, you can check one of Nathan’s previous articles for a detailed rundown on How to create immersive intros.
After you think you’ve nailed your intro, start playtesting while making use of a funnel tool, such as the one available for you in GameAnalytics to figure out how players progress through the first time user experience (FTUE). You’ll be able to identify where exactly your users are dropping off, and fix the issue. Check out the funnel below:
It represents players’ progression through the first three levels of a game, with events being sent when players start and complete each level. In this case, the first level coincides with the tutorial.
What is striking at a first glance is that only 13% of the people that started the tutorial got to complete it. It’s a huge drop off, also considering that the levels of completion moving onwards seem stable. This should definitely set off your alarm.
The first thing to check after noticing this abnormal drop off is the Average Conversion Time. In other words, the average time it takes a user to go through each step of the funnel. In GameAnalytics, this is displayed in the Funnel tool underneath the chart. In the case we’re analysing, from the moment players start the tutorial it takes them more than 25 hours to complete it. Note that this doesn’t translate into gameplay time, but actual time elapsed between the start and the completion event being sent.
What is happening in the tutorial level? To dig deeper into this issue, you could start by sending a design event (custom event) appending the actual gameplay time it takes users to play through. That value will reveal whether or not the tutorial is too long, and players need to complete it over more than one gameplay session. If that isn’t the case though, and those who finish the tutorial do so in a short amount of time, then one assumption worth investigating is that there’s maybe something in the level where people get stuck. This probably makes them come back to it later, hence the long number of hours between starting the tutorial and completing it. For more on how you can make use of design events in GameAnalytics, check out our documentation.
Either way, in order to see exactly how players are moving through the tutorial, you would need to build a different funnel, breaking the tutorial into several conversion points/milestones to figure out what causes the drop-off.
2. Your game’s sessions are too long
As Nathan argued in his article, nowadays people are looking more and more for bite-sized games that they can play on their way to work or when they have 5 minutes to kill. Long gameplay sessions could make busy players churn. Here’s how you can figure out whether or not you game has this problem.
Recently, GameAnalytics has introduced Benchmarks. This lets you compare the overall KPIs of your game to that of others, either in specific genres or at large. We currently have 6000 active games in our network, so you can be sure that your conclusions are based on statistically relevant samples, no matter the genre.
Let’s take the example above that benchmarks a game’s Average Session Length to that of all the games in our network pertaining to a certain genre. What this tells you is that your Average Session Length is almost in the 90th percentile. In layman terms: compared to your genre’s average, your game’s sessions are longer than most.
This should make you wonder: is there a certain part of my game driving this average upwards? Take a look at the chart below:
The graph above shows the session length distribution for two levels in a game. A quick look at this makes it obvious that in level 2 (green) sessions take longer compared to level one. Of course, this is normal: as it’s an indicative of the fact that the difficulty level increases from one level to the next. You don’t want your players to get bored, and not experience the feeling of achievement and skill improvement that keeps them playing. However, sudden rises in session length from one level to the next, might indicate an unbalance in your difficulty curve. So this is something you should be keeping an eye on.
3. You are not targeting the right audience for your project
As Nathan mentions, targeting 60 year-olds with a brawler is ludicrous. Though this 3rd reason relates more to game design decisions and choosing the right distribution/marketing channel, what analytics could help you identify here is where you should concentrate these marketing efforts. Looking at your new user’s distribution per country, will indicate where your most users come from, geographically.
Evaluating the quality of the cohorts of players you pay for it’s also important, as you need to make sure the users coming in are quality. This doesn’t only mean how they retain or monetize, but also how they consume the content of your game. By using the integrations with one of our attribution partners, you can filter all dashboards by attribution source, or even to campaign level. This way you can truly evaluate the performance of your campaigns, and make sure you’re acquiring the right users.
Let’s say you are acquiring users, via multiple campaigns and each of the are targeting different audiences, filtering by that dimension you’d be able to see which of them are bringing in the most users – like in the example above.
4. You didn’t run enough tests
Bugs can destroy a player’s experience, and as Nathan points out, you should be having testers early in the game’s development cycle. In GameAnalytics we offer several ways of testing for errors. One is the Quality dashboard shown below, which not only counts the number of errors and users affected by them, but also lists them together with their severity level and appended messages.
However, some errors need to be identified quickly, so you should also keep an eye on the errors reporting in our Real-time and Live feed, to make sure there are no major issues with your games that might cause players to churn.
5. Your tutorial slows down the player
Like we talked about above, the best way to get an idea if your tutorial works is to create a funnel that measures the progression of your players through it. Here’s a very good example of a tutorial that is slowing players down, to the point of them churning before completing it.
Notice the considerable drop off between the percentage of players that start the tutorial and that of players that complete it. You can also see that the number of hours players take to complete each milestone of the tutorial is very high, amounting to almost 12. As with our first funnel example, to dig deeper into the issue, you should break the tutorial into more steps and see which poses a challenge to most users, causing them to drop off.
6. Your game is too hard to pick up & it presents sudden rises in difficulty
For both these reasons, you can use the dedicated player progression events in GameAnalytics. Instrumenting them will automatically populate the progression dashboard that will indicate how players are progressing through your game’s content and trigger any signals regarding difficulty and level balancing.
In the example above, we’re shown the Win percentage of a game to be 98.37%. This is a huge win percentage and it indicates that the game’s difficulty is too low. Players might be getting bored. Opposite, a very low win percentage will indicate that the game is too hard – your players might get frustrated.
7. Resources are too scarce… or too abundant (we would add)
Either way this points to an unbalanced in game economy. Instrumenting our resource event type will give you rich information about the way in which your users sink (spend) and source (gain) the resources in your game.
In the chart above we are looking at the total flow of one of the currencies in a game: Coins. What this tells us is that the number of coins owned by the players increases almost constantly. This can happen for two reasons: players are earning too many Coins, more than they see a reason to spend, or they are hoarding them for the same reason: they don’t have anything valuable for them to spend them on. Either way, the inflation of coins thus created, is causing this currency to become redundant in your economy design, and create an unbalance.
To uncover what’s causing this, you should be looking at which items Coins are sinked on. In the graph below you can see the flow broken down by item type, and the way they are sinked and sourced.
So, your players are winning Coins in Slots and as Rewards, spending them on Accessories and Boosters. As mentioned above there are two reasons for which this is happening: you are giving away too many coins in your games and/or you do not have enough content for your users to spend on. Depending on your economy design and development plans, you should either reduce the amount of coins you give out as rewards or push more content updates, with potentially more expensive items for your users to spend their Coins on.
This also ties in with the last reason mentioned by Nathan in his article: Updates aren’t coming fast enough – social and multiplayer games need to be kept alive. So beside pushing content for your users to sink Coins on, also release some more playable content!
8. Your game punishes inactive players
A question we get very often when working with developers on analytics is: “How do I track churners and their actions before leaving the game?”. Our recommendation here is to change the perspective. Instead of trying to figure out what your churned players were doing before never coming back, as this is not really an event you can send us (ComingBack: [Never]), track why users have come back after a long absence. Was it a Push Notification, being challenged by a friend, a Facebook share? You’ll thus find out what are the mechanics that create friction or the desire for the players to come back to your game.
We hope this runthrough has given you some points to start your analysis from, and also some ideas surrounding what the thought process should be. To synthesize, keep an eye on your high level KPIs, and identify any trends that might reveal problems in your game. Use tracking of deep game concepts (progression, in-game economy, etc) and analytics tools to drill down into what exactly drives these (downward) trends.
Identify trend, isolate the cause, fix the problem, rinse and repeat until your game runs smoothly. Simple, right?
Have a question? Reach out on twitter @Geekymango