Expectations and reality - our experience in creating and promoting mobile games

Hello to all. My name is Jaroslav, today I will tell you about how we created the mobile game Crazy Shapes Run Adventure - which we have already lived through, as well as the most interesting and useful things that will help you avoid many mistakes.

Prehistory

The idea of ​​the game appeared in 2016. It turned out to be good, because Denis and Victor were the founders of the project - the first was already involved in creating games, and the other one understood in building business processes and in management. In such a bundle, and began to work.
')

The origin of the idea

At first they wanted to choose something else, but they stopped precisely on the idea of ​​an adventure runner for smartphones and tablets, because they are in great demand in all countries. And they began to implement the ideas. Based on the following postulates.

The game should be:

• Splat;
• With good graphics (good found a positive designer with good taste);
• With your story!

They also thought about monetization - they decided to apply advertising (and where without it), and in-game purchases - according to all the traditions of the genre.

Began to do beta.

Development

The first version was assembled quickly, but the appearance of the game in the App Store was delayed. That one did not like it, then the other. As a result, users saw the game only a year later. I wanted to do it qualitatively, which I did in principle. While doing - tested on a narrow group of about 7 people - they gave the game in their hands and watched the reaction, after making edits and repeating the process. There were 3 people in the team (including the designer), and also from time to time attracted specialists in freelancing.

First launch on the App Store and promotion

The launch took place in May 2017. It can be said that high expectations did not materialize and the results were much more modest. However, the installations were and people played.

Methods of attracting users

They attracted people for the first time with the help of 3 tools - they brought them to the TOP, did reviews on several sites and ordered advertising in Vkontakte groups.

In the end, this is what we got:

• 19,755 downloads required for output to the TOP (US + RF)
• For the output in the TOP in the Russian Federation it took 8,000 downloads
• To withdraw to the TOP in the United States took 11,647 downloads
• Total downloads - 25,239 downloads (motivated and unmotivated)
• Unmotivated downloads - 5,592 downloads

The graph shows bursts at the time of withdrawal to the TOP and further falls. With targeted advertising (the last point), this does not happen - the installation goes smoothly, facilitating analytics.

Reviews on sites

Reviews ordered on the Russian sites - Prostomac, AppTime, AppleInsider in the maximum package placement. A big surge from placement did not receive. Why - is not clear. It is possible that there were few of them, or they came out with a great delay, or maybe for some other reason. The cost of reviews, by the way, is on average from 10 to 30 thousand in the Russian Federation (the price depends on the publication itself, as well as on the placement package).

Public

Advertising in Vkontakte groups was placed not very much. It turned out that it is best to invest in “millionaires” on a narrow topic - it is better to buy advertising in public on the topic “Apple” with 2 million subscribers than in the public publication “Prikoly” with 6 million. For smaller publics (up to 1 million) you can not waste time and money.

Publicity for placement is very convenient to search in Sociate, and only then write off directly with the administrators - it turns out cheaper.

Conclusion in the TOP

The conclusion in the TOP carried out by category "Arcade". The procedure is controversial and does not always give a good result - someone works, someone does not. We can say that we have worked - from 20 thousand motivated installations, we received a "train" in the form of about 5 thousand organic. That is, the percentage of organic matter for us was about 25 (minus about 20 other channels), which can be considered a good indicator. By the way, who cares, you can read about this method in this article - vc.ru/11905-store-top

When withdrawing to the TOP, always ask for a guarantee, otherwise the agreed number of downloads may not be enough, and your application will not fall into the TOP, which is almost equivalent to the loss of this money.

Work with publishers

We had thoughts about finding a publisher, but did not grow together. Some ask for rights to the project, others - indicators. Our retention rate for 27 weeks was 27 percent, and the publisher started working from 30. As a result, the cooperation did not work out (maybe for the better).

Ico

“HYIP” with ICO did not bypass us and we decided to try our hand at this. We created a website (thank God there was another person, Artem, who helped with this) and launched several advertisements. Here, our expectations were also not justified - instead of the expected millions of sales, we received a rather modest number of applications. Despite all the modesty of their number, they continue to flow to this day in the absence of active promotion. In general, ICO is a separate project that requires well, a lot of attention, effort and money.

U-turn

The game started, there is an installation. They began to look at the indicators in Google Analytics, which were not very good, and began to refine. We came to the idea of ​​connecting new testers, because those old ones did not give any useful information.

As a service for testing, tried UserTesting and off you go! Literally right away (after ordering 5 tests) we saw what was wrong and quickly corrected. Further, several times new features were tested, which eventually led to the release of a large update.

Big update

The second updated version launched in December. The results were better than the first time. There were positive reviews on the page in the App Store. Hold went up, DAU and loyal user percentage did too. This gave us measurements and data for further improvements.

The current project performance is as follows.

Impressions, clicks, inorganic downloads, organic downloads, conversion:


Hold:


Activity:

Optimize the page on the App Store

Before we released the update and ordered the promotion, we made the optimization of the page in the App Store - we updated the keywords and description (this helped the Hindus - this service is very cheap for them). The result was not long in coming - the number of organic plants (through the search) increased by 50-100 percent.


The page after optimization (in fact, the description has greatly changed, the screenshots have become slightly different and from the new version, of course).

Tracking system

In fact, a major role in creating the game is played by the tracking and analytics system. At the first launch, we only had data from iTunes, as well as Google Analytics. Probably at some stage this is enough, but when working with several partners at the same time, their capabilities are not enough (for example, by default it is impossible to see which channels the users come from, which of them deleted the application and when), and the data acquisition process from there it takes more time and skill.

Under the influence of these factors, we decided to use the AppsFlyer system (we received 2 months of a free period instead of 1 because we started working with an Indian company - yes, in addition to cheap labor, there are also such bonuses).

For reference, the tracking system shows where the users came from, the analytics system shows what they do in the application. Now, most solutions like AppsFlyer combine both.

Traffic attraction

For the second launch, we decided to refuse to withdraw to the TOP, as well as from advertising in Vkontakte. Instead, the choice fell on targeted advertising, which gives a guaranteed number of users interested in the game. Among the channels were Instagram, Classmates, Vkontakte, In-App-advertising, Facebook. Placement of articles for support in thematic publications left and added foreign.

Targeting average rates (CPI) turned out to be as follows and depend on the region:

• RF - 1.2-1.5 dollars for the installation.
• US - from 1.8 to 2.5 dollars for the installation.
• Europe - from 1.5 to 2.5 dollars for the installation.

Expensive, but what to do. Want good traffic - pay.

When ordering traffic in Europe, the countries with the highest percentage of the English-speaking population immediately agreed - Great Britain, Malta, Sweden, the Netherlands, Finland (based on data from Wikipedia).

Agreements with partners

This is a big question and better give it enough attention. It is mandatory to specify in it:

• The exact number of downloads;
• Who creates creatives (you or a partner);
• Rates for each region;
• Implications of fraud detection;
• Payment terms (preferably 50-50).

It is also highly desirable to prescribe the requirements for the quality of users (for example, such and such a percentage of retention on day 3, or such a number reached the level 3), especially if you already know the performance of your project.

Anti fraud

Be prepared for the fact that some of the traffic will still be designated as fraud. Nevertheless, you still need to follow the fraud. For this we used the Fraud Score system.


Red values ​​are marked above the rest. Already here you can pay attention.

In addition to this system, fraud can be traced along other lines, among them:

• Retention rates too low;
• Retention rates too high;
• The sharp decline in the retention rate on a particular day;
• Low event flow with high retention rates.

You can also read the article from AppsFlyer .

I compared these indicators between partners, as well as with organic traffic, which I used as a kind of reference (yes, maybe this is not entirely correct, but in my case it allowed me to give at least some starting point).
We found 1, 3 and 4 points.


Suspiciously low value of the indicator "Hold".


A sharp decline in the "Hold" on the third day.

Conclusion

These are the stages of the project and we have such numbers. Now we are engaged in further improvement of the project, optimization of the game balance, monetization. We are also working on porting to Unity 3D, so the game will soon be available on Android devices.

What do you think about the game and our performance?

PS

We are faced with situations that, according to AppsFlyer (as well as Google Analytics and App Metrica), we have about 7,000 downloads. At the same time, in iTunes over the same period there are 3800. Everywhere they write and say that this is due to the difference in tracking methods (Apple ID and unique devices), but this is very doubtful, since the difference is very big. What do you think about this, what to look for?

Write your opinion about the game and everything you think.