Case Full HP: how to get featured from Google Play and adapt ASO to different countries

Full HP Ltd is an international mobile game development company with 40+ employees and offices in Rostov-on-Don and Cyprus. The portfolio of 8 games, including Mad GunZ (the choice of the editors of Google Play) and Blocky Cars (the choice of the editors of Catappult). Mad GunZ has over 12 million downloads, Blocky Cars has over 42 million downloads at all sites.

The company is actively involved in the life of the IT community and is the organizer of the Sunflower game-festival.

The Full HP Ltd team translates texts for Blocky Cars and Mad GunZ in the online translation service Nitro and agreed to share life hacks with us:

• how to get the most out of ASO optimization
• how to get to the main page of Google Play
• how to monetize games for children
• and that gives the output of games on alternative sites.

- Mad GunZ and Blocky Cars are available in 12 languages. Tell us how is the localization process going on?

- The process we have is structured this way: the first game in early access is received by players from Russia. After the game passes certain tests, we open early access for MULTI-5 countries (English, French, Spanish, German and Italian) and begin to localize pages in these languages. For some time in all countries screenshots and page descriptions are displayed in English. As soon as the game enters the world, we immediately add the standard languages ​​in which all our games go: in addition to the languages ​​from MULTI-5, these are Portuguese, Arabic, Asian languages ​​(Korean, Japanese, Chinese, Thai).

Gradually, the pages of the game are localized in the store into new languages, as this allows you to reach more users. It’s more convenient for us to translate these texts through Nitro, because often the translations are ready in a few hours.

While the translation is being carried out, we analyze the keywords and select the most relevant keywords for the game - taking into account that these words can be organically embedded in the text.

- And where do you localize the game itself?

- Previously, we worked with an agency that deals with translations, but because of the long work with the documentation (billing, payment through the bank and only after that the transfer was made) the process was very delayed. Plus, some terms in the game require clarification - in the same Mad Gunz, for example, very specific names of weapons, and translators had to explain what exactly we had in mind.

We do not have much text in games, so we completely switched to Nitro both for translating the page in the store and for localizing the game itself. We make games for children, and we don’t want to load them with a lot of text. The most voluminous in our game is the user agreement. When we were featured on Google in Korea, it was necessary to translate this user agreement into Korean. The text is long, legal, and we translated through Alconost managers .

We have a life hack: standard phrases such as “play”, “menu” wander from game to game, so we store them in translation memory and do not translate each time again, but use ready-made translations.

New texts appear when game designers introduce new content to the game. There are, say, 10 phrases that need to be translated to the release. We drop them on Nitro, write a comment for the translator. As a rule, there are no problems, but if the translator misunderstood something, we explain it through the manager.

- You mentioned keyword analysis. Tell us how you work with ASO?

- We always rewrite ASOs for specific countries if we consider it necessary. Here, for example, is a screenshot with data on Turkey. At that time, the page was not localized in Turkish, screenshots, descriptions - everything was in English. This shows the conversion rate from the page before we localized the application page. A very small number of users reached us - 15.8% of the 22 thousand users who visited the page in the Play Market.


Conversion to page localization (Turkey)

As a rule, after page localization, conversion increases by 3-5% of the total number of users who visit the page. We translated part of the description and all screenshots into Turkish. In the screenshots, the phrases are short, and users primarily focus on the image, only some of them read the description itself.


Conversion after page localization (Turkey)

But what was the conversion 2-3 weeks after page localization in the store - the increase was 18%, (100% = 3 614, this is 15.8% who installed the game before localization, the increase is calculated from this number because there has been an increase in users visiting the page). We found that localized screenshots work, but we need to find keywords for Turkey.

The same thing with Vietnam - the conversion rate was 18.3%, it became 19.5%. This is an increase in a few weeks. ASO accelerates for a long time. To see how conversion will grow, you need to observe and constantly work on keywords and screenshots, and then you will notice the first result in 2-3 months.

During the ASO optimization of application pages in markets and after regular A / B tests, we found out that for our games page conversion in countries such as India, Vietnam, Malaysia is 5% higher if the text and screenshots are localized.

- Can you show how the conversion rate changed a few months later?

- No, because there it was not only the page localization that affected the numbers. In July, we had phishing, so the indicators increased significantly due to phishing and the influx of audience. Mad GunZ has been in Google's top selection on the homepage for three weeks. There were many more users who visited the page than usual, so conversion rates were significantly different from regular metrics. We have to show a small gap, but there is a gain due to the localization of the page, and this is visible. And so with any country.

But with Hindi, we have a different situation. For them, we also localized the page in the store, but the effect was lower than expected. Most searches in India are in English. Most likely, this is due to the fact that few companies localize the page in Hindi.

In Germany, the effect of page translation was more noticeable. There, we generally had a minus conversion (a percentile below 50 indicates that we are not getting users). That is, they went to the page, but something pushed them away: maybe not a localized page, incorrectly selected screenshots - because for different countries even the order of screenshots can matter.


Conversion before page localization (Germany)

When we translated the page into German and selected the keywords in German, not English, not only the conversion increased (by 25%), but also the number of users who visited the page.

Locating the page and especially screenshots is important. If you do not want to spend a lot of money on translating the description, then at least translate screenshots and a couple of phrases from the description into the user's native language, and this will save the situation.


Conversion rate after localizing a page into German

- Our customers say that when you only translate a page, and the game itself is in English, users can be upset and give low ratings. Have you encountered this?

- No, because our game is localized in all these languages, except for Vietnamese, Turkish and Hindi, but users of these countries did not complain.

Perhaps the players are upset if the game does not have localization in the MULTI-5 languages ​​(English, German, French, Spanish, Italian) and in Russian. Still, these are very popular languages, and if the page is localized to one of them, then the players expect the game to be in that language too.

We always add these languages, and of course Arabic, Thai and Asian languages ​​- because this is a very large part of the audience that I want to reach.

- You say that you sometimes revise, change keywords. And how often do you need to do this?

- Most often, ASO has to be rewritten in Japan, in Korea, since they are a very specific audience and they need a certain approach.

We check ASO regularly, working out each country. But there are countries that do not need this, such as the United States. We make sure that the conversion remains at a good level, but we can conduct A / B testing of new keywords, icons or screenshots.


In the USA, we have the highest conversion rate, and the percentile is above 75.

An interesting point: if ASO works fine in the States, and we take these words and use them in England, Canada, Australia, it will not work. Same thing with Portuguese in Brazil vs. in Portugal, as well as with Spanish-speaking countries: what works in Spain may not work in Mexico, Argentina, etc.

We often do A / B testing with Google, it helps a lot to understand in which direction you need to move in the description, which icon has led more users to the page, etc.

We use a marketing tool that allows us to see where the game is on one or another key request. Also, using it, we can check what chance the game will have on the desired key.

Using the example of Blocky Cars, it can be seen that after we optimized the page, the conversion increased by 6% - from 19.3% to 25.3%, which is equal to an increase in users of 50% (100% = 3021 - 19.3 % who installed the game before localization; the growth is calculated from this number, since there has been an increase in users visiting the page), and the number of users has increased from 15,000 to 33,000.


Conversion to Blocky Cars before fixing ASO


Conversion to Blocky Cars after fixing ASO

- Let's talk about profitability. Your games are translated into 12 languages. Was it worth it? Maybe 5 popular European languages ​​would be enough?

- We have many paying users from those countries for which we have translated the game. I can give an example of localization in Portuguese: this language is not included in the MULTI-5, however, players from Brazil are among the most paying in our games.

When we released Mad GunZ on an alternative site, Catappult, the game came out localized only in English.


Growth after localization in Portuguese. The chart shows the featured and localization in Portuguese. Top countries: Brazil, Mexico, Portugal, USA and Vietnam.

We had featured (the biggest jump in the screenshot), and after it the indicators fell back. Since most of the players are from Brazil, it was decided to localize the game and the application page in Portuguese, after which the installation indicators crawled up.
Here are the 5 most profitable countries for us:


The same countries are leading for both games, and the indicators do not differ much.

- I look, all the USA in the first places.

- The United States is always in first place, because they are one of the most paying countries.

- And also Japan.

- Yes, Japan and Korea. But with them there is a difficulty even in making them interested in installing the game, and after installation - in inducing them to make a purchase. Given their specific taste, it is quite difficult to lure them.

We have a partner in Japan who places our games on Japanese sites. There is a certain number of downloads, but with Google Play does not go to any comparison. Among the sites on which we are represented, there is ONEStore - this is a Korean site, and, as you can see, there are also not many downloads there.

- What about China? From China, usually at least a lot of downloads.

- In China, we have 22 million downloads. We also work through a partner there, because in China we need a license to distribute the game. Easier to interact with the publisher. But despite the fact that there are a large number of downloads, the yield from there is less than it could be, since the income is shared with a partner. The publisher has a specific set of sites with which he collaborates. In China, we are represented at 30 venues where the publisher distributes us.


In China alone, 22.6 million downloads are even more than on Google Play!

- Have you had to adapt the games themselves or advertising campaigns for users of Asian countries? What features in their preferences have you noticed?

- Of course, Asian players have very different preferences. When translating to
these languages ​​we use other texts than for users from Europe. For Asia, it is necessary to adapt not only the text, but also the visual part: the icon, screenshots, videos ... For example, in the screenshots for Asia you need to observe high brightness of colors, and place the text differently than in the European and American versions of the game.

We tried to increase the brightness and contrast, once we made the skin of one of the characters under the anime ... As a result, we spent a lot of effort, but the result was not as impressive as we had hoped.

They tried to replace the standard horizontal screenshots with vertical ones - they heard that in Korea they are better for conversion from the page. The bright colors in the vertical screenshots went well, but the fact that they were vertical and the game horizontal was bad.

Asia is a very picky audience. Recently, we conducted an advertising campaign on iOS in a number of countries, including Japan - the conversion in Japan was the worst.

Adaptation in the game itself is not performed, with the exception of some unacceptable points for the children's audience of China. There was an isolated case in Mad GunZ: when the game was first launched, there was a training that a fairy (a character with a beard, but in a fairy dress) conducted for a European audience. But for China, a man in a dress is not acceptable. They see this as unnatural. We added pants and a shirt to him, and he began to look like a Chinese mid-level manager. There were no similar problems with Japan and Korea. Mad GunZ is also present in Iran, and for them we also cleaned up a fairy man in a dress.


Oops: the Chinese did not like the bearded fairies in the dress. On the right is an adapted version for China.

Speaking about ASO, we try to use Asian trends in the description. For example, we noticed that our target audience in Asia has the most popular query when searching for a game - “robots”. In Blocky Cars, we just introduced robots during the update, so in the European description of the game we emphasized that there are cars in the game, and in Asian countries that there are robots.


Blocky Cars

- You have been featured for three weeks. Is this in every country? Or how does it work?

- It was the editorial choice in the Battle Royale category, in particular, Mad Gunz was in it. Blocky Cars and Mad GunZ were featured in the world almost at the same time, but in different categories. Mad Gunz is three weeks and Blocky Cars is a week.

- What does it depend on - several weeks, one week? How to stay there longer?

- It depends on how often the category changes. Battle Royale was one of the categories for the summer. The editors have selected several categories, for example: "Best Arcade", "Battle Royale", etc. The games selected for the category were posted on the main page. These categories change every three weeks.

And in the category where Blocky Cars was, there the games change every week. It is called "Novelties and Updates."

What would you advise to those who also want to be “fixed”? What contributes to this?
Mad Gunz just noticed, we tried to keep a certain level of profitability, the minimum number of failures (in technical terms: the game crashes and other things that prevent the user from playing comfortably), to make high-quality content, and this contributed to what Mad Gunz chose.

As for Blocky Cars, here we already contacted specialists: we submitted an application that we have a good game, and we would like to attract more users through phishing. The same requirements: a certain level of profitability and a small number of failures. Then they sent their requirements from Google, and already under them, our technical experts corrected the game a little.

We also had to add screenshots of the gameplay without inscriptions and without processing - the scenes in the game as they really are. If the company wants Google to notice it, then you must definitely add at least 1-2 clean, not photographed screenshots.

- How long do players linger in your games? Share statistics?

- Blocky Cars came out exactly 5 years ago. And we have users who play all 5 years. They are few, of course, but they are. The same thing with Mad GunZ - it was released in July 2017, and there are players who play from the very beginning (for 2.5 years) and are waiting for updates.

The average retention of the first day at Mad GunZ is 35%, at Blocky Cars 33%.


D1 - First Day Retention, D7 - Seventh Day Retention

First day retention in the top 5 countries (Russia, Brazil, Thailand, USA, Japan) ranges from 30% to 39%, the highest rates are in the USA and Japan. Both the numbers and the list of the most active countries in Mad GunZ and Blocky Cars are the same.

Stickiness in both games is about 12%, there are ups and downs, but on average about as much.

- And what are you doing to keep the players?

- Blocky Cars 5 years, we have not added new content to the game for a long time, but just recently released a big update, which greatly pleased users. After that, many old players returned to the game.

Mad GunZ released a major content update at the end of December. We try to add something themed to the holidays. For example, Mad GunZ has a winter weapon that can only be purchased during the winter holidays.

Social networks help us to retain old users and attract new ones. Mad GunZ is represented in VK (in Russian), FB (in English), Blocky Cars - in VK, FB and Instagram. We also use social networks to identify the needs of players and their problems.

Thanks to social networks, we are transferring traffic from one game to another, because in Blocky Cars groups we can talk about Mad GunZ and vice versa. Players move from one game to another, the age of the audience of these two games is more or less the same: from 7 to 14-15 years.

We regularly hold contests in groups and give out promotional codes. Competitions are mainly aimed at ensuring that the player receives in-game currency or an item in the game.

Once there was a problem: players were offended that they did not receive gifts from the developer, we solved this issue with the help of promotional codes and contests. Also, from the games themselves, we have a link to social networks and a task for which players receive in-game currency: sign up for a group in FB (for all foreigners) or in VK (for Russian speakers). Our audience likes it.

- How did it happen that you chose a children's CA for your games?

- The choice of the target audience was influenced, including by the trend on Minecraft in the USA. Children really liked this game, and it remains in the tops to this day. Moreover, the audience of the game is only growing, and the franchise itself is developing. We thought it would be cool to give young gamers the opportunity to choose entertainment with a pleasant and familiar visual style, but with fresh gameplay. In particular, there are still no full-fledged Blocky Cars analogues on the mobile market.

The children's niche cannot be underestimated: children spend more time playing games than adults, sharing games in school, in the yard. We often get questions in community messages: “I play Blocky Cars with a friend at school, my friend has a cooler car than mine, tell me how to assemble a cool car?” Or "I wanted to play with a friend at school and could not find him, tell me how can I find him in your game?"

- How then to monetize games designed for children? Children can’t make purchases from their own account, but ask their parents?

- For modern children, the donat in the game is a kind of pocket money that their parents give them, and you can also pay for purchases in games from your mobile phone account. We also have advertising, including offerwalls (i.e., the player is given a specific task that must be completed in another game, for this he receives in-game currency in our game.).

- Recently, the case told how difficult it is to make an application for a children's audience, because the parties are tightening the requirements ... What can you say about this?

- The parties have now tightened their policy on working with children: in particular, Apple prohibits the use of other applications in offerwalls, and Google only allows the use of advertising networks accredited by them.

For us, the only difficulty, perhaps, is the requirement that there is not a lot of blood in the game itself and that the screenshots do not show that the shots are fired at each other. Mad Gunz is a shooter. Therefore, in Photoshop we move the characters a bit, remove the sights. All the same, everyone understands what is happening, but we don’t show blood and violence.

We recently took new screenshots for Battle Royale mode. In one screenshot, we showed how a hero - a girl - stands on a shark, she was wearing high knee-highs and shorts. And Google “stabbed” us with this screenshot, prevented it from entering the market - indicating that this clothing is a “call for sexual action”. And when the designers simply combined golfs and shorts and made pants - that's all, no problem. Although the little man is pixelated, and there is no sexual connotation there.

Thanks to the Full HP team for sharing an interesting experience!

About the author

Article written in Alconost.

Nitro is a professional online translation service for 35 languages. Created by Alconost.
Native translators, order from 1 word, online chat with customer support.
50% of orders are ready within 2 hours, 96% - in less than 24 hours.

PS: Nitro is good for translating small texts. If you need to localize or translate a project more seriously , welcome to Alconost .

Source: https://habr.com/ru/post/undefined/