Recently we held a webinar about game analytics, it was co-hosted by Sergey Himmelreich - founder of ORC WORK, Alexander Pashin - Senior Game Designer from Silly Penguin, and moderated by me, Vasiliy Sabirov  - Lead Analyst from devtodev.

During the webinar we came across few tips that might help to enhance game analytics process. 

Find the optimal density of the training in tutorial

Experience confirms that people often fall in churn after the tutorial.

For instance, there was a game in the "Farm" genre and the players churned. After series of experiments, it became clear that the game had quite high density of the explanations of mechanics: the mechanic was demonstrated to the user, then 2-3 approvals, the next mechaniс was demonstrated, then 2-3 approvals, a bit of free time, more mechanics demonstrated. And with the help of analytics the most important moment was caught: there was no need for such a long entry into the game, it's better to give a "piece" of the gameplay and postpone the explanations of mechanics for later.

Can we now state unequivocally that it's better to let the player understand the game himself, reducing the density of training in the tutorial? It totally depends on the game:

Optimize the tutorial technically

It is important to pay attention to the technical issues of the tutorial. We analyze not only the path of the user, but also the download of the tutorial, especially when talking about mobile device.

Case study: in devtodev you have an opportunity to analyze in detail the first session of the player, to build a funnel of more than 100 steps and find the problem areas. One of our customers integrated a lot of steps, including the invisible steps when levels, textures were loading and so on. And it turned out that the biggest churn happened at those invisible steps - user didn’t wait for the download. The developers had to optimize the tutorial technically, that helped them to significantly increase the retention rates.

Count the retention

It is necessary to distinguish between the classical retention and rolling retention.

This is why the rolling retention shows better results, and it is often used to tell the investors about retention, but it is not entirely correct. For most work tasks the classical retention is used (at least because of all the benchmarks on the Internet taking it into account). For mobile games it is often measured over a period of 30 days, as follows: 0-day, 1st-day, 3rd-day, 7th-day, 14th-day, and also 28th-day or 30th-day.

If we talk about the big games that can be called GaaS (game as a service), the long-term retention makes sense, as it correlates more with the revenue. Our research indicates that the longer the player is in the game, the higher is the likelihood of his payment in each subsequent day, and the larger is his every next payment. Thus, long-term retention  increases the audience and grows its monetization potential. Can be considered the retention of 180 or 360 days.

Anyway, retention is a very sensitive metric, it is like no other reflects the loyalty of the players. In particular, here are the factors that retention may depend on, in our opinion:

Analyze core gaming loop

Core loop can be analyzed with the help of a funnel. You must specify the correct sequence of steps in accordance with the objectives and steps of the core loop, set the time of passing for each step and watch the users taking them. With the help of such graph, you can either see that the mechanics do not work, or they are not perceived by users. Also during construction process of such a report it is useful to take into account the levels and the number of attempts spent on their progress.

A bit more:

Track the right events

Here is the list of events that makes sense to track (pass to analytical system) to understand the picture of the user behavior in the game:

It is important that a set of steps allows to build a funnel for all key user actions: payment, level progress and so on.

Real case studies that confirm the necessity of gaming analytics:

Case 1: In one of the games we saw that the players fall into churn pretty early. It was a Trading Card Game, and we realized by analyzing the behavior of the users, that the players did not have a long-term goal. They simply lost interest in collecting cards. As a result, we decided to use the so-called method "to hang a carrot": we invented an award for the set of cards. The retention rates increased significantly, by 20%.

Case 2: We had a game with an internal free market. We realized that our revenue is reduced, but could not understand why. We toiled for more than three days, trying to understand the reasons. We were not able to see how much money did players have on hands, we could not see the goods flow. Finally, we found out that our players duplicated the in-game items, there were becoming more and more. We learned about the existence of the black market, and realized that some people earned in our project even more money than we did. It was the original detective story.

You can find more tips on game analytics reading our blog here.