Developers Club geek daily blog

2 years, 10 months ago
Competent segmentation of users — one of cornerstones on which the success of game is based. In this post we will talk about how it is possible to select correctly target audience of games on any platforms: client, browser, mobile.

This article does attempt to make the overview of different criteria of segmentation and is aimed more at the beginning audience of developers of games and the interested geymdizayny.

Segmentation of audience when developing games


By what criteria game designers, analysts and marketing specialists divide audience and what segments select?

  • The simplest — by the demographic principle. Here everything is clear without explanations.
  • The preferred genres. This point is characteristic of games, cinema, of the sphere of entertainments. If we speak about users of online stores, then it will be difficult to define their genre preferences.
  • Behavioural features. Here so-called psychotypes are meant. There is a certain quantity of models on which players are divided into psychological types.
  • Kazualnost. Very popular term which opposes "hardkorshchik" and "kazual". It is the qualitative characteristic which is difficult for expressing in number.
  • Acceptance of innovations. Relation of players to new features, opportunities and types of a gameplay. It is important at the choice of target audience and determination of a gameplay.
  • Solvency. As you understand, it is an important point for any developer. :)

Numerical characteristics

Each segment of target audience can be described by means of different numerical characteristics. What size of audience; what share of conversion; how easily players overcome obstacles to be included in our project; degree of loyalty (retention); an opportunity to hold audience (retention and sticky factor) — numerical parameters which we select at each segment, quite a lot. I will not go deep into their description. It is already made here. In the same article I will pay attention that key characteristics of each segment are its ROI (return of investment), LTV (lifetime value) and the size of this segment. Anyway they partly comprise such important criteria as the average income from each payer (ARPPU — average revenue per paying user), a share of the paying audience in each segment (PU) and retention (commitment of audience).

Actually it is much more characteristics, and it is possible to select them on each segment with the most different methods. Only one retention is calculated by tens of different methods. For example, there is classical retention describing a share of the players who came into game in specific day. Classical "retention of the seventh day" conditionally means that now the user who was registered seven days ago came. There is rolling retention when the user registered a week ago came now or into any of later days. Return retention — when the person was registered seven days ago, and came from the second to the seventh day. There are still bracket dependent return retention and different terrible methods of calculation of this characteristic, and under each project usually it is considered in own way. The subject of calculation of indicators of deduction deserves separate article.

As it was noted above, all these characteristics are consolidated to two main indicators: lifetime value (a lifetime income from the payer) and ROI (profit on the invested capital). We are interested more the second parameter because it considers not only quantity of the money earned from the player but also amount of potential investments into each segment of audience. If the audience size very small, and LTV huge, then, probably, there is no sense to invest in this audience because return will be small just because of the small amount of target audience. For example, you want to purchase very hardkorny audience interested in black-and-white mangas. You posted banners on the thematic websites and in shops and now are going to start up TV advertizing. In such cases of attempt of additional attachments in marketing lead only to decrease in efficiency because there is not a lot of people interested in your subject and it is necessary to work with them more pointwise.

Qualitative characteristics

Except numerical characteristics, is such which cannot be measured by digits. For example — the amount of necessary content. Let's say there is a certain inveterate player in World of Warcraft, it needs one thousand vaults, huge battlefields, arenas and many other things. Also there is a fan of Candy Crush Saga. It needs one thousand levels, the truth enough simple and similar each other. For these two players, naturally, the amount of the made content is necessary absolutely different.

Graphic style. I will not speak about the Asian market where love animeshny girls with big eyes and a breast of the eighth size, but in different games, depending on their genre and target audience, graphic style can be a miscellaneous too. And it very strongly affects development cost. The cost of production of graphics for League of Legends, World of Warcraft and Candy Crush Saga different. Often the setting and a genre exert impact on graphic style, imposing the requirements. Agree, Call of Duty in animeshny stylistics will be aimed absolutely at other audience, than Call of Duty executed in style of realism.

Game opportunities. There are people who will play with the great pleasure any kinds of "kraft", that is in games where it is necessary to create something, to make. To someone it is interesting grind and sharpening — mechanics rather monotonous. And there are people who want to kill, destroy and dominate. We have to remember that game opportunities — features — too have the different cost of development and if we defined target audience, then have to know also what features suit each group.

Segmentation of audience when developing games


Now let's remember about demography. With it everything is very simple: sex, age, region. It is curious that among the general audience of Mail.Ru Group not so many people are younger than 18 years. Generally more adult, more solvent audience prevails.

Speaking about education, we first of all think of distinction in retention, solvency and the preferred genres of our audience. I will note that we did not find certain essential distinctions between people with an average, average special and incomplete higher education. From the point of view of profit it is approximately equivalent groups. And here owners of the higher education and school students are two different groups both on revenue, and on the size and loyalty of audience. To the word about education, everything, stated in this article, I in detail sort on a course a game design of an educational program Management of game online projects at the Higher school of business information science.

Segmentation of audience when developing games

When we speak about regional accessory, we think first of all of the market which developed in this or that region, its economic features, conditions of business and cultural aspects. So, in Africa the playing audience small. Middle East too not the most interesting to us region.

Just recently there was the most significant event: the amount of the game industry of China exceeded the amount of the game industry of the USA. I will remind that the States one of the first opened to the world the industry of video games and many years held superiority. ARPPU, the average income with paying, in China and in general in Asia above, than in America, however percent of paying is much lower.

Earlier browser games were one of noticeable segments in structure of revenue of the Chinese game industry. There is an opinion that the population of China had insufficiently powerful "iron" that development of browser, but not client games pushed. Also legislative bans connected with consoles till the last moment interfered with forming of this segment in China. And here cheap mobile phones based on Android came to them quickly therefore from browser games they switched practically at once to mobile. Now it is a huge segment of the Chinese market, and it continues to grow in absolute expression.

The preferred genres of games

There is a number of classifications of games by genres. Some prefer the names which historically developed in the game industry. Someone uses the terminology developed by players: shooter games, balls, stratezhka …

It is important for us that different genres can show different numerical characteristics. Two most interesting — deduction (retention) and ARPU (the average income from the player which is usually calculated in a month).

Segmentation of audience when developing games

The chart reflects a situation with mobile games. In the left part the share designating "seven-day retention" that is how many people continue to play for the seventh day after registration is postponed. The red line belongs to the right scale, it is a share paying. As you can see, in a genre of arcades, for example, minimum levels and retention, and paying. But it does not mean that an arcade — a genre, bad or unnecessary for us. In them can play more people, and it is possible to earn enough money for them.

Segmentation of audience when developing games

Casino — very profitable genre. And on social networks it is much more profitable, than in a mobile segment. Now many large companies, for example BigFish or Zinga, actively earn money on social casinos, despite stagnation of social networks as game platforms.

Strategy and RPG show very stable results. At them both a share paying, and retention rather high, but also the competition in this niche quite serious. On the diagram average data on the market, and at many mobile games Mail.Ru Group indicators others are submitted. For example, from this point of view "Evolution" — something between action and RPG with strategy elements, there "retention of the seventh day" reaches 20–25%. Is almost twice higher, than on average on the market, as well as the size of the average check.

Psychological types

For certain many of you are familiar with segmentation according to Bartle. Richard Allan Bartle, professor of Esseksky university, was the first who received a doctor's degree in the area a game design. For us it sounds a little bit strange. He offered system of segmentation of players on four psychotypes: killers, careerists, sotsialshchik, researchers.

In the context of this post first of all numerical distinction between these psychotypes is interesting to us.

Segmentation of audience when developing games

On the chart the ratio of different parameters of four psychotypes is given above. Killers are ready to invest most of all money in game. They will not stop before anything if only to purchase the Sword Thousands of Truth and to kill all noobs.

On the second place careerists. It is very good audience, in their diagram there are no pronounced sags. At them both high ARPPU, and the big size of audience. On average to 40% of all players of the project are careerists. In some games, the Perfect World type, their more than 70%.

The most devoted audience — researchers. At them the maximum "retention of the 28th day".

The most viralny audience — sotsialshchik. They have huge social ratio — the attraction coefficient describing as far as this audience actively invites in game of friends. These are people who pay a little, ARPPU is lower only at researchers. It is not enough of them. In general they are especially selected with nothing, except for virality coefficient.

The virality is very important for us because each sotsialshchik can invite several other players in game. For example, the friend killer who will pay us money. Therefore when developing features we try to cover all range of psychotypes of Bartle.

Segmentation of audience when developing games

As you know, Bartle was a person kind and cheerful: it did not stop on model from four psychotypes. After it he developed system from eight psychotypes. He added a perpendicular axis to the so-called plane of decision-making by the player, and the three-dimensional space turned out. To "action — interaction" and "players — the world" it added one more orientation: "it is conscious — unconsciously".

In practice we very seldom use this expanded model because without polls, without tests, without laboratory researches it is impossible to define, the player or unconsciously consciously arrived. When we want to understand, a sotsialshchik it or the killer, we watch what features it uses. For example, sits in a chat much — +1 point to "sotsialshchik", 20 times killed on the arena — +1 point to "killer". I strongly exaggerate, but an essence of this division such is. In reality heuristic models are used. And in expanded system of psychotypes the heuristics does not work.

There are also many other methods of segmentation, for example the BrainHex model. This model is sometimes used in Asia, it considers the emotions experienced by the player. Here six key psychotypes and the seventh center — certain "greed" as far as they are actively developed are selected.

Kazualnost degree

How to define by numerical parameters whom is the player — a kazualshchik or a hardkorshchik?

Segmentation of audience when developing games

First of all we evaluate skill of the player, that is how fast he accustoms to game as far as it is ready to go on more difficult game levels. At a hardkorshchik a S-shaped learning curve. In the beginning it needs a certain acceleration to begin to play game about which he knows nothing. But then it shows the high potential of training. At a kazual this potential rather low. It easily begins to play, but cannot reach great skills. Or, we will be frank, usually does not want.

Fundamental difference of a hardkorshchik from a kazualshchik is in as far as they are ready to plunge into game process and "to suffer", understanding difficult game mechanics.

Still we select a so-called midkor, an intermediate class. Around the world 80% of all audience — a kazuala, about 15% — a midkor, and the others — hardkorny audience.

Why it is interesting to us to make games including for these remained 5%? Because they have extremely high retention and ARPPU. It is not enough such players, but they pay much and differ in high degree of commitment. Midkorshchiki less devoted is also paid less, but they are three times more. Kazualov a huge number, but profit from each of them it is very small, and they change games as a glove. They can be monetized advertizing or by other methods.

For example, бóльшую the part of audience of Clash of Clans makes a midkor. There is a share of hardkorshchik who there are a lot of forces and time give to game, and there is a share of kazual, but they for a long time in game are not delayed.

With Candy Crush Saga absolutely other situation. The main share of profit is brought by kazuala, there is a lot of them. On the second place hardkorshchik. In Candy Crush Saga there is a tough hardkorny audience which plays many hundreds of levels and donatit good money.


Segmentation of audience when developing games

Let's look at training cycles. Such models used even for training of staff of KGB. When the person for the first time sees something new, in our case — new features in game, it gets to the area of unconscious incompetence, unconscious misunderstanding. He does not know as how to do, at all about game knows nothing. After a certain shakedown period to it conscious incompetence comes. He understands what knowledge is not enough for it. It is that moment when active training begins. Here also the difference between kazuala and midkorshchik is shown. Kazuala speak: everything, I do not know and I am not going to develop further. On it their training comes to an end, and they actively play at the level.

Midkorshchiki go further. Let's say you never played MMORPG and came into World of Warcraft. Then understood that there are abilka, talents, fractions, quests and a lot of everything but to understand to you it for the present laziness, or just there is too much information. Thus, you passed to the second stage. If you understood all this, learned all features of game, constructed the tree of talents and feel like the good player, then you passed into the area of conscious competence. You understand as how to do.

At last, when you to automatism work all the skills, pass into the area of unconscious competence as fighters of special troops, cyberathletes or school students in Dota. :)

On learning curves practically all balance of any game in which there are a development and content is under construction. Initially all thought that the learning curve has to look as a direct line, that is constantly growing pressure upon the player. Then developers changed character of constantly growing pressure up to logarithmic.

How the learning curve is under construction? Roughly speaking, this relation of complexity of mastering to the spent time. And complexity can be expressed also shortage of resources, and lasting achievements of the goal, and many other methods.

Developers gradually left from model of permanent pressure increment on the player and just began to bring it to some maximum level. But it turned out that to people it is uninteresting. Then the new learning curve was tested.

The person gets to one area, trains, perfects skills, meets new complexity and so on. But some games went further away. For example, in Candy Crush Saga the sawtooth curve of complexity is used.

Segmentation of audience when developing games

Her developers consider that, except permanent training and growth of skills, there has to be a moment when again just it becomes easy for user and to play. Match-3, clickers and many other games use such approach too. How it becomes in Candy Crush Saga? At a development stage is formed the analyst, responsible both for configurations of levels, and for an order of loss of candies. This order submits to the certain patterns, so opaque that the player will not be able to distinguish it from complete randomness. All parameters of the dropping-out candies are selected on certain algorithms. If the player cannot long reach next "top" of a saw, then candies are combined in such a way that at you "supercandy" by means of which you at last pass level and you get on following turns out. Developers call it "special wow - effect" which gives to players hope. The player always has a hope: "Well pancake, will develop, there now … Pancake, it did not develop. Once again tomorrow I will try, to a zadonach, still I will think up something". And at some point the algorithm decides that it will be enough to press on the player, zheleshka develop, there is a magic, and the player passes on the following level. This solution was very effective.

Segmentation of audience when developing games

How to use learning curves in practice? This illustration is made on the basis of the work of the famous game designer Jesse Shell devoted to game balance. The green line from below — production of a certain abstract resource. The red line — consumption of this resource. As you can see, consumption is more than production. The difference between them is also the complexity expressed in pressure upon the player.

On the light-red line it always grows. But if we want to implement the correct learning curves, then consumption has to grow not linearly. This diagram something reminds a sinusoid. In zones of extrema when the player has maximum deficits, we can receive from him money. That is competently to use pressure curve for game monetization.

Pay attention that in the beginning all curves go almost in parallel each other. It is connected with what at a grade level does not need to be pressed on the player. There the most important — to make good retention that the player became interested in the project, kept, the second, third, fifth, seventh day lived. It is necessary to hold some time of the player that it was possible to consider with a certain share of confidence it the audience. After that it is possible to begin to press on it and to apply means of monetization.

By means of curves of complexity and training it is possible to measure the player's kazualnost in number.

  • First, what sector it reached, on level and/or the spent time.
  • Secondly, how many at it game sessions in day and what average duration of one session. It was purely experimentally established that at kazual the number of game sessions are more. The audience which is considered casual plays "Evolutions" on average 6–7 times a day, for 3–5 minutes. This rather typical behavior for kazual. And at hardkorny audience each session can last one or two hours, but such sessions in day one-two, it is no more.

For example, at two of our famous projects "Mini-games" and "Alodiums Online" very different nature of sessions: their frequency, duration and distribution within a day.

Acceptance of innovations

All our audience is divided into five groups on extent of acceptance of innovations: new game features (opportunities), genres, features of game process.

Segmentation of audience when developing games

The smallest group — innovators, them about 2,5%. These are those who create something new, those games, gadgets and so on. They generate ideas, a half of which is senseless, and the others horrify even experienced specialists. :) But some of their ideas are valuable, and they are selected by the following group — early apologists whom ~ 13,5%. These are those who create a mode.

Apple when advanced the smartphone, were guided by group of early apologists. They needed flock of followers who will advance this idea further. It is very difficult to jump between groups or to hit into everyone, but it is always possible to strike in a segment necessary to you and to motivate it to bear idea further. This approach is used in many games.

Earlier the majority, unlike apologists, will accept your product or idea only when it becomes fashionable. iPhone became fashionable — representatives of the early majority buy iPhone now. The late majority, on the contrary, will not buy until the product becomes mass until it appears at everyone. And the last group, conservatives, begin to buy when the product already got out of fashion. For example, today conservatives are those who go with old phone folding bed or Nokia and do not accept smartphones.

Many of modern games actively used this curve for the marketing. Two most striking examples — Candy Crush Saga and Clash of Clans. The Candy Crush Saga developers very well understand that it is heavy to people to pass from one group into another, for this purpose it is necessary to give them rather strong motivation. Therefore game was guided by the early majority. To Candy Crush Saga the King company made about 200 games, first of all social spheres, not mobile. The market was prepared for Match-3, very large volume of such games collected. And here they release game and speak — "it is a mass masterpiece". They struck in the early majority. And as soon as it plunged into Candy Crush Saga, right there under laws of a genre the late majority began to be connected. Transition between these groups is rather simple if completely to attract the early majority.

It is possible to make it by means of huge marketing investments. For example, in 2013 the King company earned 2,2 billion dollars from Candy Crush Saga. For comparison: the amount of the Russian game industry for today — about 1,6 billion dollars. From these 2,2 billion they spent about 1,7 billion for marketing. You represent what low marginality? The lion share of profit is spent for the most powerful advertizing on TV and on the Internet. Thus they completely covered sector of the early majority. And in 2014 marketing expense were much lower at similar rate of a gain of audience. It began to be connected the late majority. In 2014 developers earned 1,8 billion dollars. In 2015 will earn much, most likely more than 1,5 billion dollars too. At all this you should not forget also about such factor as "it was stupidly lucky". :)

Other example — the Supercell company with the game Clash of Clans. They were guided by early followers. Clash of Clans created almost new genre which did not exist. They integrated genres of city builder and tower defense in asynchronous fight, and the new genre — Clash of Clans was born. Now all games of this genre are called clones of Clash of Clans. Early followers made game popular, and thanks to huge efforts of the company also the early majority began to be tightened. The late majority does not accept Clash of Clans yet. According to the statistics for 2015, Clash of Clans still one of the most profitable mobile games in the world.

In other words, the correct use of a curve of acceptance of innovations allows to optimize marketing.


There is a method of segmentation of audience on the amount of effected payments.

  • First of all we select those who do not pay. Their majority. In a mobile segment of their more than 90%, according to some information even 97%.
  • The second group — accidental payments, their about 8%. All regular payers in a mobile segment make 1,5–2%. In social spheres a similar situation. In browser games pay about 5% of players. But there are exceptions: for example, "Legend" and many other browser toys had more than 15% of the paying audience. In client games normal digit of 7-10%. In some taytla, for example in "Alodiums", the share of the paying players is even higher.
  • The third segment which we already mentioned — regular payers. They are divided, in turn, into "gudgeons", "dolphins" and "whales". "Gudgeons" pay on slightly, literally couple of dollars. They make the main amount of the payers bringing the companies of 15-30% of revenue. "Dolphins" bring more, and so-called "whales" create the main share of revenue. Them is insignificant a little, but each of them gives such big money that it pays back practically everything.

In China, for example, all monetization is based only on "whales". There consider: let in our game of only 0,5% paying but each of them bring 1000 dollars monthly. Monetization is not so strongly started in Russia on "whales", but we consider them too.

In Europe, on the contrary, everything is shifted even above on a scale towards "gudgeons", the share of "whales" is slightly lower there, as well as the amount of their payments. For example, in FarmVille, one of the first successful social farms, average ARPPU of "whale" — 90 dollars from one person a month. And in Heroes Charge — 240 dollars.

Russian realities

The Russian game market has a number of features. It both rather low virality, and small average check, and small percent of paying. ARPPU in the market of Russia and the CIS is much lower, than in the USA and China.

From the point of view of audience psychotypes in our market slightly big, than on the world, a share of killers. Many think that killers are an evil audience, and sotsialshchik — kind. Anything similar. For example, trolls from Dota 2 — it is normal sotsialshchik. In general the player's psychotype does not characterize its personal moral qualities in any way. He speaks only about a set of activities and the purposes which are preferred by the player. If in the world on average 10% of audience belong to group of killers, then we have 15–20%. The same high rates are observed in Turkey and in South Korea. It is interesting that in Europe it is less killers, than in Asia.

In Russia Asian mechanics of monetization, so-called gacha well work (trunks, tape measures, lotteries). Some of them are even prohibited in Korea. Our legislation prohibits only one so far: to sideline real money, otherwise it falls under "the law on a casino".

Despite domination of Free-to-play of model, in our market there are also games with subscription model — Pay-to-play. On a subscription the only a few number of games, such, for example, as World of Warcraft works for us. The so-called Freemium, that is Free-to-play plus the premium account is popular. In this Freemium a third of revenue can fall on a bonus subscription.

In Russia rather negative attitude to alien cultures. We did research on game settings and found out that the most hated settings for the Russian players are superheroes. Only 4% of the Russian audience want to play them. But if to select from all audience of players till 18 years, then the share of persons interested to play superheroes sharply will rise to 12%. The younger generation more is positive to superheroic settings.

Westerns, Asian genres, Samurais, anime — all this do not love ours too. But besides the young audience perceives it more positively. There are settings which nobody, for example loves in general "Stone Age".

And the last — so-called universal values. The audience in the USA most favourably perceives the appeal to democracy and freedom, ideas of justice, "the brother for the brother", "on concepts" are close to the Russian players. Each person has values.

If you have questions, will be glad to answer in comments.

This article is a translation of the original post at
If you have any questions regarding the material covered in the article above, please, contact the original author of the post.
If you have any complaints about this article or you want this article to be deleted, please, drop an email here:

We believe that the knowledge, which is available at the most popular Russian IT blog, should be accessed by everyone, even though it is poorly translated.
Shared knowledge makes the world better.
Best wishes.

comments powered by Disqus