"Our passing CPI is at the level of 0.20-0.25$": Flime about the development of hyper-casual games

Flime is a young company developing instant and hyper-casual games from Belarus. In the two years since its foundation, together with Gismart, it has managed to mark five successful releases. Two of them have managed to get a total of 45 million downloads since August alone. We talked about their creation and internal processes with the co-founder of the company Artem Legonkov.

App2Top.ru : Hello. Let’s start with the formal reason for our conversation. One of your last projects — Foil Turning 3D — in December passed the mark of 30 million installations. At the same time, your other new project — VIP Guard — also performed well, reaching 15 million. Congratulations! Were the projects difficult in terms of production and history?

Artem Legonkov

Artem: Hi. Thanks! In terms of production, these products were completely different. If we talk about VIP Guard, then the very first retention measurement showed that the game is promising. R1 is more than 30%, in addition, the low CPI is $0.14 on iOS. In general, from the first iteration it became clear that this is a potential hit and that it needs to be finished. All we had to do was a little polishing, which, however, dragged on for 3-4 weeks. We have embedded rewards (rewarded videos), a small meta, and we have finished.

But the game Foil Turning 3D from the point of view of product iterations was very difficult. Initially, its gameplay was limited to the following mechanics: the player had to simply tap a foil ball.

The metrics were not good enough, retention was kept at 27%. We did several iterations, tried different chips, but it didn’t give noticeable results. At one point, we were even on the verge of shutting down the project. As a result, the gameplay was improved. They made it so that after tapping, polishing the foil and other satisfying actions, the ball had to be thrown into the vase. This iteration gave a crazy boost, retention reached 40%, which allowed us to make a global launch. But even at the final stage of the launch, when we saw the difference between CPI and LTV, this project did not look very promising. In the end, largely thanks to the gameplay, which is interesting to both female and male audiences, the project shot.

About the beginning

App2Top.ru : From the very foundation of the company, it was decided to engage in hyper-casual projects?

Artem: No. Initially, we did not plan to engage in hyper-casual games in Flime: we wanted to apply our expertise in the HTML5 games market (before the launch of Flime, we worked in another startup that prepared projects for Instant Games).

With this desire, we came to Gismart as a partner who worked closely with Facebook, Snap and others.

The first project we launched in Flime was a game for Facebook Instant Games — Word Search Together. The game has been downloaded more than 20 million times. It still makes a profit today.

Word Search Together

After that, we started looking for new directions. Thanks to Gismart and the partnerships it has established, we were among the first (along with such a giant as Zynga) to go to Snap Games. A year later, we became their key partner. Our Color Galaxy game is one of the flagship products of the platform.

Color Galaxy

Today we are developing several new games that will be available on the platform in the coming months. Snap-direction brings good money, but the most important thing is that here we are at the forefront of the industry: Snap Games runs in all new features primarily on our product, together with us discusses what works better, what is worse. For example, monetization is currently being tested on Color Galaxy. We were the first to have access to snap tokens. This is a beta feature with which you can make purchases in the game.

Also, a little over a year ago, we launched the development of native hyper-casual games. We had no experience working on them. After a huge amount of trial and error, we managed to develop certain rules and guidelines that allow us to launch successful projects in this niche.

App2Top.ru Q: Where did you start and how fast did you grow?

Artem: The three of us started out, but we quickly started to grow. The first money we earned, we just invested in recruiting a team for hyper-casual games. They didn’t take a penny for themselves, because we were interested and are interested in completely different scales.

Now we continue to actively invest in the development of the studio. We have already employed more than 70 people. There are two directions of development in the studio: instant and hyper-casual games.

Grow (in terms of personnel) we started immediately after the first successful project. I think we were in a hurry then. We had little experience in hypercage, we didn’t understand how to make products. They recruited a lot of people and did not know how to deal with them.

However, the more mistakes we made, the faster we learned.

When we launched the direction of hyper—casual games, we focused more on traditional approaches to project management – SCRUM, Kanban and other methodologies. We built the company based on these principles, built clear processes, hired people for these processes. This led to the fact that we simply did not have time to launch products.

In hypercage, speed and flexibility are important, and we built processes for the sake of processes, believing that this is more important. We couldn’t even launch projects with a low CPI. Besides, we were cloned: other studios worked faster.

Complex processes suitable for large companies that develop complex products are not needed by hypercage at all.

Under our inflexible processes, a team suitable in spirit and thinking was not always assembled. As a result, really cool guys left us, with whom we were lucky enough to work at that time. They understood the full scale of our mistakes. At first they endured, and then they left. Perhaps they still think today that we still have the same processes, resulting in a waste of time.

It was the departure of top employees that made us realize that we were doing something wrong. Gradually we came to what we have today.

About processes

App2Top.ru QUESTIONER: And what is there today? Tell us about the current development processes in Flime.

Artem: Today, within the framework of the hypercausal direction, we have four product teams working. It would even be more correct to call them internal studios. We treat them exactly as studios. Each of them is autonomous to the maximum extent: it determines what processes it needs, which people to assemble, which products to make and how to scale.

At the head of each micro studio there is a Product Owner (PO) or Producer who independently manage the processes within the team. Therefore, now the processes look different in each team. It all depends on how PO built them. That’s the whole magic.

It’s very cool that each PO has their own opinion, they argue among themselves, everyone goes their own way. And, as practice shows, you can really go to the hit in different ways. You can test the product immediately on LTV, spending a little more time on creating a prototype, you can quickly make prototypes with basic functionality, testing them on CPI, you can look for balance, you can concentrate on solyanka products, where there are several mechanics at once, and you can, on the contrary, concentrate on finding one cool core mechanics.

At the same time, there are 3 pillars common to all teams of the principle on which we build our work.

The first is the independence of the teams.

The second is the responsibility for the product. All team members should feel it. All employees take maximum part in the development of the game, in creativity, in inventing features. We welcome disputes with PO. No matter how experienced he is, personal opinion is always subjective. Sometimes it’s worth listening to, for example, a Unity developer. This allows PO to get an alternative view on the issue, it can make you think and help correct the original opinion.

This can be achieved only if you limit the team to resources, so that each of them relies solely on their own strength. When each team member understands that no one else will complete the task, that’s why he feels responsibility better.

The third is product orientation. It is impossible to be responsible for something without understanding the essence of the product. For this reason, each of the team members should understand what is really important in the product and what is not. Everyone, regardless of their profile, should be a bit of a game designer.

When switching to such a system of work, there was also a side effect. People who initially came under clear processes could not adapt to the new format of work. I had to part with some of them. On the other hand, those who used to feel violated, who wanted a more flexible approach to work, have now been able to open up great.

And for all this, of course, now we have a bonus system set up. For teams not including PO, this is a certain percentage of the net-profit of the product. The team that made Foil Turning 3D, for example, received a bonus to the existing salary of up to $ 10 thousand.

Services and PO receive interest separately.

App2Top.ru : From numerous conversations with developers, it often seems that success in the hyper—casual games market (at least in terms of finding a successful concept) is always luck. Is this the case and, if so, how do you try to maximize it?

Artem: The whole essence of modern hypercage is the race for CPI. At the same time, the market is gradually becoming more mature, finding a concept with a low CPI is no longer enough to make a hit out of it. Now the focus is shifting more to being able to hold the product up to good product metrics.

We have a know-how that allows us to make products without relying on a random CPI. Its essence is in well—established marketing research. We have moved away from coming up with random ideas. Now all the products that we take into work already have a good forecast in terms of CPI at the start. Their chances of success are correspondingly higher. This allows you to move away from the endless race for CPI that has developed in the current market.

App2Top.ru : How many projects are being prepared at the same time, how many are usually spinning in tests, and how many are already going into full-fledged production?

Artem: One team, as a rule, works on one or two products, but, of course, there are exceptions.

Each team has 2-3 Unity developers and 1-2 artists. If we are working on a product that we are testing for CPI, then one developer is allocated for this. Accordingly, two products that are tested for CPI can be worked out simultaneously in the team. Or it can be a mix of a product in the CPI test stage and a product that is already at the grinding stage. Now the whole Flime can work on 4-5 products at a time.

But there is a nuance. The resources of each team individually are not infinite. Because of this, their PO are not always ready to take risks, to do something completely new and untested that can fire.

To push PO to experiment, we created a Fast Team. It contains a large number of developers who are able to quickly make prototypes, focusing exclusively on CPI. Any PO who feels sorry for the resources of his own team throws all his ideas into the Fast Team. If any idea gets a good CPI, he takes the product to his work. This is a kind of additional service that does not affect the loading of the main commands. PO actively use it.

Now we are aiming to increase the number of internal studio teams to 7 pieces. However, we are ready to launch them only with top-end PO who already have successful experience. We are ready to invest in such people, give them freedom, help assemble and build a team, provide all our tools, Fast Team. In return, we are waiting for good hits on a regular basis. Naturally, there is no question of working for wages here.

App2Top.ru: To what extent is production an internal process? Or does the publisher in the person of Gismart pull up with expertise and marketing at each stage?

Artem: No one understands their product better than the product team, so now marketing and the product team work together from the very beginning. This is not publishing, this is a super-dense cooperation from the start between the product producer Flime and the creative producer Gismart.

App2Top.ru Q: How much time does it take on average for a project that eventually goes into production?

Artem: It really depends on the product. There are products-solyanka with a lot of mechanics. They take longer. There are those in which you only need to tighten the metrics. Somewhere we have to spend more time searching for working creatives. There are cases where everything goes smoothly right away.

We strive to ensure that 4-6 weeks pass from the start of product development to its entry into the global market. It doesn’t always work out, but this deadline is a kind of guideline for teams.

App2Top.ru: Do you pull out those concepts that you subjectively like, but which do not show the necessary KPIs on the first tests? For example, recently in the Telegram group HYPER CASUAL there was an interesting case when one team had a very interesting concept (a mix of shooter and Snake vs Blocks), but they noted that they did not develop it because the indicators turned out to be so-so.

Artem: If the indicators turned out to be so-so and you don’t see a way to rethink the product and get a different result, then this concept won’t be interesting for the market either. Subjectively, the game may seem good, but the market dictates its own rules.

As I said before, here we rely on our know-how in the form of marketing research. They help predict whether there will be demand for the product. In most cases, we make a decision based on this forecast.

In our work, we take only those concepts that suit us and according to marketability [with the Cambridge Dictionary, the term is understood as “ease of sale”, attractiveness to the audience. — Editor’s note.], and by retension. For example, a concept with good marketability, but low retention will eventually give a small ROI and, as a result, there will not be much profit there.

At the same time, of course, everything is situational. This is proved by the experience of Foil Turning 3D, where the KPIs from the start were not very high, but in the end the game was released and was able to gain a large audience.

App2Top.ru: By the way, since we are talking about KPIs, what are your “passing scores”?

Artem: Our “passing” CPI is at the level of 0.20-0.25 $. But again, it all depends on the correlation of CPI and LTV. If LTV is promising, then we can give up a little bit regarding CPI. Although now we have learned how to use our know-how not only in the search for concepts, but also to drastically reduce CPI within existing products. Therefore, the CPI is even at the level of $ 1, although not the most pleasant, but not always a verdict.

A little sofa philosophy

App2Top.ru : Isn’t there a feeling that now everyone is obsessed with metrics and there is not a creative process of inventing mechanics, but simply going through them?

Artem: Great question. Usually, when communicating with a potential PO, when a candidate says that hypercage boils down exclusively to a race for a low CPI, for me this is an indicator that he is most likely not deeply immersed in hypercage and perhaps he does not have enough experience. My opinion: creativity in hypercage is not just to take an original idea from somewhere, but to bring it to a finished product with good intra-product metrics. And this is what defines a professional in hypercage.

Sometimes we get together with all the Product Owners in order to help one of the teams rethink the product. And we hang out for hours. This is, of course, real creativity, with a flow of ideas and discussions. Real space!

App2Top.ru : In this context, can we say that the authors of hyper—casual games are already a completely separate profession, the experience of which, after several years of work on hyper-casual games, is almost not applicable to games from other niches?

Artem: I absolutely disagree with this. On the contrary, I believe that the work of PO in hypercage gives the sharpest understanding of the core, which is not present in the context of developing more complex games. There are often PO rely on meta, and it’s cool, it’s high LTV, revenue. But I am sure that in the future there will be a new niche, there will be games in which everything will be built not only on meta, but also on a keen understanding of the core. And such specialists will be able to move into more complex niches that are just beginning to emerge. But this is solely my subjective opinion.

App2Top.ru: In the CIS, there is a very strong bias towards hyper-casual development. How healthy is this story?

Artem: It depends on what kind of distortion we are talking about. If we are talking about revenue, then there is no bias towards hypercage. If we talk about HYPE, then there is certainly a bias. And there is one reason for this — a colossally low threshold for entering the market. Although in 2020 everything already works differently. Two years ago, it was possible to assemble a conceptual build, get together with friends, get a low CPI, heal without fear of cloning, and share the rhubarb. For one-time earnings, it worked. Today, as I have already said, the grocery approach is important. There is no longer a low entry threshold. There are enough professionals on the market who know how to make products many times cooler than amateurs, no matter what cool idea was invented. Today, for indie developers, hypercage is a meat grinder, where it is difficult to compete and even more difficult to earn.

App2Top.ru : Is there any fatigue from working on hyper-casual games?

Artem: Regarding fatigue, the question is to the point. To be honest, I can’t imagine how we would work now without having our know-how in the form of deep market research. I myself had burnouts and situations when I tried my best to earn money, and in the end realized that I had spent too much while I was going to the desired result. It’s really difficult mentally.

But our know-how has breathed new life into the processes. Now we constantly live with the feeling that a new release is somewhere nearby. It motivates. In this state, it is already more difficult to burn out.

App2Top.ru : Aren’t you afraid that in the case of IDFA, the hyper-casual games market will collapse?

Artem: No, I’m not afraid. Now there will be a pure sofa philosophy: the more people think about this, the less they do and the less exhaust they have. I see no reason to be afraid. A solution can always be found. The main thing is to move and develop.

As for the IDFA, it will probably affect the CPI. On the other hand, restrictions often lead to new revolutionary solutions. Perhaps such measures were just not enough for hyper-casual games. I do not exclude that the IDFA ban will give a kind of kick to the market, force the hypercage to evolve, form a new niche with even more money. It will be interesting to observe and participate.