Gamedev and the crisis

Probably many have already noticed this alarming trend ... falling profit, loss of paying players, decrease in purchases and average check. And having paid attention, they began to read metrics on the subject - โ€œWhat did we do wrong? Why are players leaving? Maybe the logic has become a curve or the transition button is not pressed, maybe we have offended the players in some way? " After debriefing the picture does not become clearer, 10 years small fluctuations but stable numbers and then suddenly ... bam and a steep dive!



You will have to start a little from afar, otherwise the context of the disaster will be incomprehensible. Here you need to clarify for yourself one very important concept "Model".



What is the first thing that pops up in our memory at this word? Model cars or spinning balls on a stand? This is it! It is a reduced copy of some original object or process with varying degrees of reliability. "Why do we need models at all?" - you ask. It's all about our brain and its limitations. Take a car for example, it has weight, width, height, engine power, body material and countless more parameters. Every time you want to think about a thought related to a car, you have to load all the information on the object into a tiny part of the brain, and there is little memory, neurons quickly get tired (in fact, they don't get tired, they just run out of mediators, but the meaning is the same). If the object is simple, then there is enough memory, but if the object has thousands of parameters, what about the poor brain? Evolution has found a way out!You can give weights to the parameters of the object and load, depending on the available resources, the parameters with the largest of them. As a result, we have a simplified (incomplete) description of an object or process - the Model. The model can have a different degree of reliability - the correspondence of the described parameters and reality. The model has a degree of detail - these are the same features with assigned weights, the more there are, the more accurate and detailed the model. Now the object or process represented by the model can be loaded into the brain and understood (and forgiven) and realized, perform various operations with it, reduce, expand, provide input information and receive outgoing information. Yes, of course it will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).depending on the resources available, the parameters with the largest ones. As a result, we have a simplified (incomplete) description of an object or process - the Model. The model can have a different degree of reliability - the correspondence of the described parameters and reality. The model has a degree of detail - these are the very features with assigned weights, the more there are, the more accurate and detailed the model. Now the object or process represented by the model can be loaded into the brain and understood (and forgiven) and realized, perform various operations with it, reduce, expand, provide input information and receive outgoing information. Yes, of course this will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).depending on the resources available, the parameters with the largest ones. As a result, we have a simplified (incomplete) description of an object or process - the Model. The model can have a different degree of reliability - the correspondence of the described parameters and reality. The model has a degree of detail - these are the same features with assigned weights, the more there are, the more accurate and detailed the model. Now the object or process represented by the model can be loaded into the brain and understood (and forgiven) and realized, perform various operations with it, reduce, expand, give input information and receive outgoing information. Yes, of course it will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).As a result, we have a simplified (incomplete) description of an object or process - the Model. The model can have a different degree of reliability - the correspondence of the described parameters and reality. The model has a degree of detail - these are the same features with assigned weights, the more there are, the more accurate and detailed the model. Now the object or process represented by the model can be loaded into the brain and understood (and forgiven) and realized, perform various operations with it, reduce, expand, give input information and receive outgoing information. Yes, of course it will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).As a result, we have a simplified (incomplete) description of an object or process - the Model. The model can have a different degree of reliability - the correspondence of the described parameters and reality. The model has a degree of detail - these are the same features with assigned weights, the more there are, the more accurate and detailed the model. Now the object or process represented by the model can be loaded into the brain and understood (and forgiven) and realized, perform various operations with it, reduce, expand, provide input information and receive outgoing information. Yes, of course this will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).The model can have a different degree of reliability - the correspondence of the described parameters and reality. The model has a degree of detail - these are the very features with assigned weights, the more there are, the more accurate and detailed the model. Now the object or process represented by the model can be loaded into the brain and understood (and forgiven) and realized, perform various operations with it, reduce, expand, give input information and receive outgoing information. Yes, of course it will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).The model can have a different degree of reliability - the correspondence of the described parameters and reality. The model has a degree of detail - these are the very features with assigned weights, the more there are, the more accurate and detailed the model. Now the object or process represented by the model can be loaded into the brain and understood (and forgiven) and realized, perform various operations with it, reduce, expand, provide input information and receive outgoing information. Yes, of course it will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).perform various operations with it, reduce, expand, give introductory information and receive outgoing information. Yes, of course it will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).perform various operations with it, reduce, expand, give introductory information and receive outgoing information. Yes, of course it will not be a real process, and its imitation and the results will be only with a certain percentage of reliability (more detail - more percentage of reliability).



Now we will describe the current player model and selling him something for money.



What and how much we sell according to statistics:



  • Applications 5-7%
  • Subscriptions 10-15%
  • Inapy 40-45%
  • Clicks on commercials 20-25%
  • Views of commercials 15-20%
  • Sale of "cosmetic" skins that do not affect the balance. 5-10%
  • Selling battle pass - opening quest branches. 5-10%
  • Selling user data (if you are told that it is โ€œfreeโ€, then the product is you)
  • Promising sale of contextual advertising in in-game widgets. 0.1%


Our player has free time to play, at least an hour a day. Has money (at least 10 bucks a week).



How players are divided:



  • The rich (vip, whales) buy constantly. five%
  • Medium - rarely buy. 30 - 45 or teenagers with parents' money. 35%
  • Poor people hardly ever buy. 60%


Except for isolated cases of original monetization, the main sales model is as follows:



  • Selling inaps - 40-45% to middle and wealthy
  • Advertising to the poor


Selling advertising itself has many nuances, here the game should keep the player and advertising should not be intrusive appearing in the midst of a battle. Advertising is one of the most powerful demotivators for the player. Make the game less interesting and all the player is lost, and 10 cents per 1000 clicks is too little against the background of the resources spent on development. The only reliable source is inaps. Igrodely certainly learned marketing from drug dealers. A well-built balance and the cherished conversion into a paying one happened, first a little, then a little more, and now we have a new conventional poppy.



Why doesn't the old model work anymore?



Here we come to the worst moment ... the economic crisis. It began 10 years ago, but it is going into a steep peak today and, according to various estimates, the drop in real GDP for the states and China will be from 40 to 50%, in Europe 35%, we have about 5% (we have already fallen to the bottom) ... What is the threat? Those โ€œaverageโ€ ones will move one step lower to โ€œpoorโ€, which means that only 5% of the rich can sell inapas (and there won't be enough of them for everyone). What follows the breakdown of the old sales model? Change of monetization mechanics! And plots, logic, balance are tied to them. Wow, sailed! Yes, it essentially needs to create a new game, and not just as usual figak-figak in production, but on a floating target group, which promises the game to be bad. Add to this 30% of the proceeds of tears will be taken away from you app stores. Their profits will fall, but they have forgotten how to live modestly. There is little hope for a reasonable 5%here we will see the rise of any huavei stores with a rapid jack. But 30% is just the edge of the problem, shops prohibit what you can earn at least something on sex, violence, gambling, jokes about blacks.



But there is also a plus, the demand for games will increase. People will want to escape from problems, someone will smoke, someone will drink, and someone will play - it's cheaper.



So the demand will be even better than the previous one, the truth is low-paying.



Knowing such gloomy layouts, let's describe a new sales model and a new user model.



Consumer model:



  • Rich 5%
  • Poor 95%


Sales model:



  • Selling ad clicks
  • Selling views of commercials
  • Built-in contextual advertising


The new user will not tolerate injustice in the form of buying inap and an advantage in the game (and for the rich this is the main incentive). Advertising will be difficult to sell, as a new wave of games will enter the market and other monetization may replace it for competitors. Selling user data to marketing companies? - Yes! But for this you need to have a famous game with high attendance, and this is rarely the case. Built-in contextual advertising only in the first prototypes. A good option is not to take the user's money, but to create a plot or a situation in which it will be possible to freely enter, withdraw, exchange, etc., and take a tiny percentage for input and output. Transparency is the guarantee of trust - if the user understands the rules and sees that he is not being robbed, then he will be motivated and tell others, and for the new "dark ages" this is a lot.



If earlier, in order to increase the installation, it was enough to bring several million greens to the store in order to raise the game higher, to Facebook - to direct them with advertising, now there will be problems with the greens. The size of the skid is not predictable at all. So what can you do? Go to the ground! It is necessary to get people who will search for any communities on the network and gently, unobtrusively push them to create their own game (and only studio N makes really good games)! Good for those who have a "universe"!



A ready-made fan group that will buy and forgive a lot, and if it doesn't exist? The community has to come up with it. Accountants - non-payment quests. Bandits and deputies - about robbery and violence against ordinary citizens. Ordinary citizens - about violence against MPs and bandits. In general, each group has interests, this is how they differ from each other, to find them is our task! Further, a group or community forms proposals by voting on the concept, genre, logic, graphics ... Yes, yes, you can shift the work of GDshnikov and scriptwriter to users! Users will be in the thick of it, they will not be cut off from the result, the game itself will be the consistency of what they dreamed of:

- โ€œDo you want to rob cows? Take it! "

And then they should object to me:

- โ€œAre you a medex in your mind at all? Do you even know how much development costs? What salaries do you need to pay to developers, civil servants and qashniki? How much does a communal apartment cost for an office, how much does Unity cost? โ€.



Unfortunately I can. And here I, as if with the last bit of strength, pull out a card from my pocket, and it says - โ€œ2 levels on Godotโ€. Three words and one number, but the meaning for a good candidate. Using Godot we get a unified development environment for artists, animators, technicians, developers and testers. We get IDEs with source code under the MIT license, with backing on the pros and development of logic in python-like GDScript with dynamic typing and replacing the garbage collector with smartpointers, with animation and logic on blueprints, and as a highlight - you can write logic in C #. What does a python script give us? Reducing the cost of developers with about the same level. There are also disadvantages, poor integration with fashionable metrics services, but this case is already full on github. What do we have as a result - if in the logic, the managers of the binding (quests, sales, push notifications, store,game server, metrics, competitions, etc.) then only the logic of the world and the game scene will remain - for 2 levels it is 2 months with the starting graphics purchased in ready-made assets or on pinterest. Moreover, if the genre is popular, then the genre engine can also be bought inexpensively ready-made! And then interested users (interest or game currency) test the levels ... save on QA.



And now the MVP is ready, users got what they wanted, went to downloads and deposit money. All the special forces have done the work and further simple laborers will release the pumped levels until the second coming.



Something like this I see the near future, whether it will come true time will tell, but for now a la ger com a la ger gentlemen!



All Articles