What happened in the four months before the establishment of the “Canaltowns” project?

This article was originally writen by the Chinese media Game Grape (游戏葡萄) and translated by PandaYoo.com after it was licensed. Copyright belongs to the original author and it is forbidden to reprint without permission.

“Canaltowns” may be the first popular style in the mobile game market in the second half of 2020.

Canaltowns, owned by Coconut Island, has never dropped out of the Free App list Top 3 since it launched on App Store on July 2. On TapTap, a local mobile game ranking platform, the game scored 8.1 out of 24,000 reviews.

But what you may not know is that “Canaltowns” has a less “independent game” project establishment process.

After three months of pre-research, they identified the scene and architectural style, and then quickly did a suction test in Facebook Hong Kong, Macao and Taiwan-there was no game installation package even on the landing page at this stage. Four months after the pre-research, the team opened the Soft Launch observation and retention in Taiwan, and the team decided to set up the project only after it was found to be effective.

But if the data determine the project of the game, then in CEO Wesley’s view, the success of “Canaltowns” can not be separated from the uncertain side, that is, the artist’s creative part. Because the data can never tell you how popular a simulated operation of an ancient town in the south of the Yangtze River will be, and how its art should be done.

On July 29th, at the TapTap developer Salon, Wesley shared the specific project establishment process of “Canaltowns” to the outside world.

The following is a compiled transcript of the speech:

Hello, everyone. I am Wesley, CEO of Coconut Island. I am mainly responsible for the product and part of the operation work. Today, let’s talk about how “Canaltowns” set up the project.

A brief introduction to “Canaltowns”, it is an ancient town theme simulation business game, we refer to a lot of things in the process, such as “Along the River During the Qingming Festival“, “Prosperous Suzhou“, “Portrait of the prosperous Society of Southern Capital” and so on. The details in the ancient paintings depict the way of life of the ancient people in the south of the Yangtze River, give us a lot of reference and help us fill in a lot of content.

Of course, first of all, we have to admit that there is no great innovation in the mechanism and way of playing the game, we can only say sorry: the team has spent a lot of time on the content, and there are many cities in the game, such as Nanjing, Suzhou and so on. It is hoped that in the future we can make games with both new content and new ways of playing.

Portrait is also a small feature of the game, the more important point is that we do not want to talk about history from the perspective of the emperor, but to tell how the small potatoes of that era lived, so we added a lot of people’s stories. The creative starting point of this mechanism is “Rimworld”, and there are also many small people in the surrounding world, who have birth, old age, sickness and death.

Today’s topic revolves around certainty and uncertainty. I’ll start with certainty. In the TapTap developer salon, everyone talks about creation and feelings, but first I want to talk about buying volume and business.

Because of the data, there is a “Canaltowns” project

This game is a low-risk project in Coconut Island, and it has been eliminating risks in the process, and it is becoming more and more certain. How do we know that? After two months of pre-research, we got the following picture:

The whole picture is hand-drawn, and at this point, we think it is more reliable and we can move on, so based on this picture, we cut open the house and put it in the engine to make it feel playable. Of course, at that time, the program team was already working on the game itself, and there was already a prototype that could be placed to play. This sketch is closer to what the final game looks like.

The second is the characters. What should the characters look like? Our first print looks like this:

At that time, I thought this version was a little awkward, and the rightmost character was holding the grass and had no recognition. In fact, he was Li Shizhen. The art of painting this set of characters is the art of “Naughty Kitties”, and his painting style is more lovely.

But it doesn’t matter. Let’s do the test first. I have always stressed that we should do the art style test as soon as possible, because in the end, there is a high probability that the art content will be used to attract users, even if it is not used as a promotional picture, whether it is recommended by Apple or Steam, there will be the question of click-through rate, and we can see how many people like it and how many people don’t like it.

At that time, the game did not come out, and there was no game installation package on the landing page, but we still used the scenario diagram to buy advertisements in Facebook in Hong Kong, Macao and Taiwan to do click testing. For many commercial companies, this test method is very simple, in fact, independent game companies can also consider this mechanism. After our test, we found that the click-through rate was 11.87%, which was very high.

The click-through rate itself does not matter, and the benchmark of each platform is different. For example, it is more accurate to use Douyin when sending messages to the domestic market, but before “Haywire Hospital” and “Monster Chef” are tested on Facebook, the process is familiar, and 11.87% is twice as much as these games. So we’re sure we can go in this direction.

Then do the click test with the portrait, and in the end, the click rate of 4.82% is not low, but we hope it will be better. I don’t come from an art background, and I don’t know how to communicate with art staff. But when the data come out, they will think for themselves if something is wrong. Of course, before the test, you should talk to the team about how to screen the crowd and area, test it several times, and the team should approve the result. Because we have run and approved this process, everyone thinks there is something wrong with this style.

So we tried a lot of new things. For example, all of them are painted with animals, which is very interesting, and the click rate has indeed increased, but unlike the idea of the planner, the planner still wants to show people’s life in the city, so the idea was not approved.

We are also thinking that the reason for making the cute style is to avoid the beautiful style of antique ACG. First, it is not characteristic enough, and most of the games made by Coconut Island want to be distinctive. Second, its production cost is very high, drawing a picture is very complex and expensive, while the game in the first 1/2 stage only two art, we can not afford so much production.

We thought, how can we make something cheap and distinctive, so we said to the art team whether we could draw lovely things as good-looking as possible and make them into handsome guys, and finally found the style relatively quickly, and found that the data of this edition was very high. We’re pretty sure we can do this.

So after 2 months of pre-research, we determined the architectural style; in the third month, we identified the character portraits, and the test data were all very good. Later, it took another month to polish the scene and got this style, and the process was smooth.

In addition to testing the number of clicks, we also test the retention rate. This team has been simulated for a long time, and there is not much innovation in the way the game is played, so I will speak faster.

After four months of pre-research, the team started Soft Launch, with hundreds of users a day in Taiwan and measured it for one to two weeks. This version is mainly retained the next day, and the seven-day retention and payment are only reference indicators, but our two-week retention is between 9.5% and 13%, which is very good. As for the payment is not very good, it is because a lot of things have not been done, it is understandable, so we think this project is reliable.

It was only then that we decided to formally launch the project-it only took four months. In Coconut Island, a project that is not going well may take six to nine months, a year may be possible, and will even be cut off in the end if it does not work well. Because the final launch of the game will be measured by these indicators, why not think about it in the first place? So I said, this is a risk-free project.

But data does not dominate the creation of “Canaltowns”

But if I say today, “Canaltowns” is made in this way, this method can make a very good game, many manufacturers can do this, then I am definitely fooling you.

Many people in this industry pay too much attention to certainty: buying advertisements, channel referrals. Basically follow this set of logic, that is, the so-called certainty. But the question is, why do we play games like Canaltowns? Why can you make a scene like that after 2 months of pre-research?

Our certainty is very different from that of large companies: the practice of big companies is to look at all categories, let A do this category, and let B do that category. They studied certainty very thoroughly. If a company of the size of Coconut Island only plays that game, it will definitely fail.

There is not so much certainty in real life, and there are many uncertain and random idea, such as the creation of artists.

Usually, I would think that if the producer and the main strategy are the deep users of this category at the time of setting up the project, it will be a great bonus; if the team has the production experience of this category, it will also be a bonus. We made a Demo in 2 to 3 weeks. In the construction part, many of the core mechanisms you see were identified three weeks before R & D, and there are no entangled points. Then we chose a traditional model, which is not quite the same as the resource transfer chain of “cartoon farm”.

But there is another big problem: art style. We are going to do simulation games, but there are too many simulation games. In order to highlight the differences, the main strategy said at the beginning: I want to depict an ancient Jiangnan town. Because he has a group of simulation operators, he asks in it: if someone does historical simulation operations such as “Emperor: Rise of the Middle Kingdom” and “Caesar.III” on mobile phones, do you want to play? People say they want to, so we think we can try this concept.

But what does the simulated operation of the ancient town in the south of the Yangtze River look like? We are not people who love history. We don’t know anything about it. We just have this idea. Our art director is not the one who studies ancient paintings. How to create?

Most of the games made by Coconut Island have the following tonality and needs in art:

1. Not vulgar. Coconut Island to do the game art style is generally unique, or stylized, we also hope to do a less common sense of antiquity.
2. It’s easy for the public to accept. Coconut Island’s vision has always been to hope that the game has a cultural influence, so that more people can understand the beauty of the game. This feels in conflict with the first rule, as if I told art that we want a “colorful black”, which is really difficult.
3. The production cost is low. This kind of game experience is based on content consumption, which inevitably requires the production of a large number of art materials. Due to the characteristics of the painting style, it is relatively difficult to replenish manpower. We need to be prepared for a chronic shortage of manpower. Therefore, the production cost of individual art resources must be kept low.

In determining the theme of the town in the south of the Yangtze River, our first step is to organize all the art designers of the company to brainstorm and freely try all kinds of possibilities along different directions. Among the experimental results, some of the results were inspired by the “Chicken eating Rice Picture” in “Flirting scholar”; some imitated the artistic style of the Ming Dynasty; some were realistic in proportion; some were drawn with cute cartoons; and some even tried to change their perspective and draw them as horizontal boards.

Among the many attempts, a designer stumbled upon an old album, the Prosperous Suzhou, while visiting the bookstore in his spare time. After flipping through it, he decided to follow the direction of traditional Chinese painting. It wasn’t long before he drew a concept map that imitated the effect of ink painting. At this time, the embryonic form of the direction of art has emerged, but at that time we felt that the pure black and white picture was limited in public acceptance, which violated the second principle mentioned earlier.

[sociallocker id=”5614″]

So he tried a color version. The inspiration for color and painting comes from the works of Wu Guanzhong, a famous artist in China. This time, although the color is rich enough, there are problems in other aspects such as details.

Can those two things be combined? It took us one to two weeks to get this result after two months of pre-research. It was scribbled, but we thought it was in the right direction and slowly evolved to where it is now.

In the process of refinement, not only the style, but also a lot of content filling. Fortunately, the ancient paintings we found provide enough content reference, and they have much more details than we do. The reason why this picture looks good is not its artistic style, but that it has so many detailed contents that it can reproduce the feeling of that era.

To sum up: games are always influenced by data and business. To consider sales and ranking, the final result must be certain. Then since it is inevitable, we should test it as soon as possible and pursue more certainty. When we tested, the click-through rate of a portrait of Wen Zhengming was not very high; later, when we changed the name to Tang Bohu, its click-through rate doubled, so there were a lot of things that could be tested.

But data should not dominate creation, because data can never tell you that you should make a Jiangnan-style game. I have also discussed with the art staff whether we should take out the games of Top 50 downloaded by App Store in the past 3 years and find some rules. But such a thing is already being done by big companies, and if we do it again, it may not be better, and there is nothing new. What we should do more is the part of uncertainty. Of course, we also encourage creators to read the feedback. All the top game producers I see like to listen to players’ feedback, but how to change the game is the part of personal creation.

Most people in the Internet industry are correct in trying to rule out uncertainty, but Coconut Island is willing to retain the uncertainty that belongs to artists. It doesn’t necessarily guarantee the success of the next game, but it’s a part of our lives.

[/sociallocker]