5 Common Pitfalls When Setting Up Mobile Game Analytics
As a games analyst for W3i, I have the unique experience of going through a plethora of analytic instrumentations. I have witnessed nearly every analytics provider’s capabilities, strengths, and weaknesses. More importantly, I have seen how many game developers choose to implement event tracking and common places where they go wrong.
1) Know What You Want from Your Analytics Provider
There are many analytics providers in the mobile space, and they all offer different products. There are free services with a limited tool set, and companies that charge monthly fees but provide a more robust product. Understand what each provider is capable of and take advantage of those features during implementation. For example, if you want the capability to A/B test elements of a game, you’re going to have to pay. On the flipside, A/B testing can provide tremendous revenue increases.
Let's say Game A earns $30,000 in revenue per month. If you create two tests that generate 3% revenue increases, that service has already increased revenue a little over $1,800/month and that can be a permanent change for every month moving forward. The effect of A/B testing is theoretically cumulative, so a few positive tests each month increases your knowledge of what works in your game, and furthermore it could quickly snowball into huge revenue increases.
2) Set up Your Tracking with Goals in Mind
The term “mobile analytics” is sometimes bandied about as a cure-all for games. In reality, analytics and event tracking are tools those in the mobile space can utilize to learn what players appreciate and discover the pain points for others. How you choose to use this tool depends on what you want to accomplish.
For example, if you want to improve retention in your game and you want to be able to take action at the points where players lose interest. To accomplish this goal, you would want to track each event in the tutorial and how often users make it to each level. If you find that 80% of all people who download the game make it through to Level 3, but only 20% reach level 4, you can dig deeper and find the root cause. Maybe users are not receiving enough currency at level 3 which discourages them and they ultimately quit playing.
3) Don’t Track Everything
I cannot stress how important this point is. Many game developers work in a small studio and don’t even have an analyst on staff. Not only does tracking everything take much longer for a developer to instrument, but you are going to have an overwhelming amount of data.
If you’re making a city building game, there’s no reason you need to know how many times the user reorganizes their city or changes the name of City Hall. Generally the goals of implementing an analytics solution are to increase engagement, retention, and monetization. Don’t focus on little details with analytics. It’s a much better use of your time to improve the points of friction for your users, or iterate on the parts of the game they enjoy.
4) Do Track (Almost) Everything Related to Virtual Currency
In order to keep a studio open, you have to monetize users. There’s no way around it. Keep track of where users get currency, how much they get, where they spend it. It is essential to understand what users value and what they are willing to pay real money for. Tracking all the sources and sinks of virtual currency will enable you to balance your virtual economy and determine if you are being too generous or stingy with how much currency users receive from grinding. Knowing how often users purchase each item in your storefront will allow you to create new content that users will value. Getting users to monetize is not about tricks or gimmicks; users will monetize because there is something about the game that they value.
4b) Nest Related Events
This is mostly a time saver for whoever ends up analyzing the data, but who doesn’t want to have a little extra time? All analytics providers have some way to nest related events into one variable. There are a few different names out there: sub-events, event parameters, properties.
Running with the theme of a city builder, let’s say there are four different kinds of buildings (Residential, Industrial, Commercial, Decorations) and each type of building has 30 different objects the user can place. The amount of coding required will be roughly the same, but if you create four events with 30 parameters, you won’t have to consolidate any data later. If you have 120 different variables, getting that into Excel or another spreadsheet program will be a headache for your analyst.
5) Pay Attention to Qualitative Sources
Some would argue this last point isn’t really related to analytics. However, I view analytics as a tool used to improve games, and in that sense I think this applies. Getting direct user feedback can be the most powerful and helpful of all, especially since it comes from people with no attachment to your game. Read user reviews, and engage with users on a forum or through any medium possible. Not everything they suggest has to be put on the roadmap, but it’s a safe bet they’ll have some good ideas.
Instrumenting an analytics solution shouldn’t be an arduous process. If you’re like me and love numbers, you might even find it to be a fun way to reflect on your game through a critical and quantifiable lens. If you’d like to discuss these points, or anything related to gaming or mobile analytics, reach out to me at firstname.lastname@example.org.