Though Neihan Duanzi, which had 200 million registered users, was shut down permanently in April 2018 for inappropriate content, Zhang became a very respected figure in Chinese tech circles for her product acumen, which is evident in her 2018 speech.
In her capacity as CEO of ByteDance China, reporting directly to company founder Zhang Yiming, Zhang is “responsible for the company’s products in the China market, leading product management and operation, marketing, and partnerships for Toutiao, Douyin, Xigua, Toutiao Search, and other products and services,” according to a profile.
Given that six of the top 10 most popular apps in China were owned by ByteDance as of the beginning of 2020, and that substantially, all of its alleged US$28 billion in target revenue this year will probably be from China, that’s an impressive amount of responsibility. Zhang is also still in her 30s. Born in the early 1980s, there’s very little public information about her, other than that she’s from Xi’an.
Here is the transcript of Zhang’s speech, a rare firsthand account of the rise of Douyin, annotated for the readers’ benefit. Credit goes to Pandaily editor Lu Zhao for her translation work, more of which you can learn about here.
Full transcript, part one
Kelly Zhang: Douyin has been around for more than a year; we started it in 2016. Today, we are going to talk about how [we created] a short-video product that has over 1 billion daily views. In fact, it is far more than 1 billion now, thanks [to] your love for Douyin.
Annotation: Douyin was officially launched in China in September 2016. At the time of this presentation, it was just under 18 months old. I estimate that development and beta testing took about six months because past interviews reveal that the idea to make a short-video product was finalized during the company’s management retreat in Japan around the Chinese New Year of 2016, which would have been around February.
KZ: Let’s go back to the second half of 2016 when we decided to enter the short-video market. You probably know that 2016 was the heyday of the short-video industry, and the competition was very intense. But we still decided to join the battle because we saw that there was great potential here. We thought it was worth trying.
At that time, we also made a bold decision. Some of you may already know this, but we made two short-video products at the same time, one called Douyin and the other called Huoshan.
Annotation: If you’re not familiar with Huoshan (which has since been folded into the Douyin brand), we’ve got you covered.
TL;DR: Huoshan was targeted at rural China and can be considered as ByteDance’s answer to Kuaishou, the leading short-video app at the time and still a very strong competitor to Douyin in China.
Launched in April 2016, Huoshan actually preceded Douyin by a few months. It was renamed Douyin Huoshan in January 2020 and has effectively become a sub-brand to Douyin. By the end of 2019, Huoshan had about 50 million daily active users.
By the way, ByteDance’s nickname in Chinese is “super-app factory,” so having two apps in development and then launching them at around the same time (though targeting different demographics with different value propositions) were totally normal for the company. And as the succeeding years have shown, the two different demographics were sufficiently different to need two different products.
As for the competitiveness of the market, see the slide below that shows how anybody who is somebody in China has had a horse in the race.
Just some of the leading players in the short-video app market by 2017. Bucketed by “families” of ownership or affiliation, they are (from top left to bottom right) ByteDance, Sina, Baidu, Netease, Tencent, and Qihoo360.
KZ: As crowded as the market already was at the time, we started thinking about how to find our user base, define our own characteristics, and win the market. We struggled a lot to figure out how to make people like our products. We felt that only new and good experiences could win the hearts of users.
But what is a new and good experience? We spent a very long time doing research and data analysis. We tried almost all the short-video products available – both the domestic and international ones. We downloaded roughly 100 products into our phones. Everyone on our team used these short-video products every day.
As the saying [from Sun Tzu’s Art of War] goes, know yourself and know your enemy, and you will never be defeated. After we had tried all the short-video products, we realized none of them impressed us. So we thought, this may be an opportunity for us. If we wanted to [create a short-video app], we had to do better than those products, we had to impress our users.
I don’t know how you felt when you first used Douyin. Did you feel pleasantly surprised? We made it based on a lot of data and user feedback. Next, we’ll talk about our pursuit of the ultimate user experience. How did we do it?
That’s a big topic. I was even discussing what “achieving the ultimate” means with a few friends last night. You may have heard your boss say, “You’re not pursuing perfection,” or “You are not doing work in a perfect way.” I think “achieving the ultimate” is actually an attitude. If you have the attitude, you’ll have your own method of improvement.
If we measure perfection on a scale from zero to 100, with 100 being perfect, I think everyone actually has a different understanding of what the ultimate is. Some people may think 60 is already the best, some may say 80, some may think 100 or even higher. Therefore, I think the ultimate state is subjective.
We will use Douyin as an example to talk about how we pursue perfection. I’ll show you several cases later.
In terms of this ultimate experience, Douyin focused on four key [ideas]. The first one is “full screen and high definition.” You might have thought Douyin is sort of different when you first used it. Why is the video on such a big screen? Well, I don’t know what you think “full screen and high definition” is, but we made an experiment with many short videos and phones. We uploaded videos to different apps. Some videos were displayed in a rectangular format, some in square, but at least on the domestic short-video apps, they were rarely in full-screen mode.
But when we uploaded the same short video and viewed it in different sizes on different mobile phones, it was clear that the full-screen video impressed us the most. It was the most visually impactful and immersive. We know this because we did a lot of blind testing with many people.
As for [creating] high-definition [videos], we all know it costs money. Doing high-definition short videos means we needed to pay more for content delivery network/bandwidth costs. It was a hard decision to make, but in retrospect, we think it’s 100% correct.
Annotation: Most of you probably know that the full-screen functionality that Douyin “copied” was actually from another Chinese app, Musical.ly. Even though it seems inconceivable to have it any other way these days, at the time, it was truly pretty novel.
To get a good sense of what the competition looked like at the time, consider these 2016 screenshots of US-based Dubsmash and China-based Miaopai and Meipai. All of these were very successful at the time, and all were known for surfacing viral content — usually of people dancing or singing — that made its creators into legitimate stars, much like the ones we see on TikTok today.
But in the Chinese market, there were no other full-screen competitors who were mainstream, probably because these firms were already doing well and so they didn’t think of “fixing” their platforms.
Despite its nifty design and success overseas, Musical.ly, which Douyin was modeled after, failed to gain much traction in the Chinese market. It was also late to the game, though not that much later than Douyin.
Musical.ly entered China – what should have been its home market – in May 2017, eight months after Douyin’s launch, under the name “Muse.” There have been many reasons cited for its failure over the years: its clunky registration process, poor choice of marketing partners, weak recommendation engine, lack of dimensionality in its content, and even faults in its generally admired user interface that did not appeal to Chinese users.
If you ask me, it was all of those things, yes, but the main thing really was that they weren’t sufficiently localized. It’s true that the Musical.ly team was Chinese, but that does not mean they understood the Gen Z urban Chinese youth that well. That’s something the Douyin team had to go and discover for themselves as well.
A screenshot of a leading Chinese short-video app at the time, Meipai, from 2016. By this time, it had 160 million monthly active users and was making celebrities out of their most popular creators – proof that Douyin had plenty of stiff competition.
KZ: I don’t know if you’ve noticed, but videos shot via the phone’s built-in camera are in full screen. So now that we’re making a short-video mobile app, why don’t we showcase the videos in their original [format]? We thought that may be the most suitable option, although the previous standard feed system couldn’t present full-screen very well; we had to make a new user interface design.
Then we came to the music. Our slogan used to be “Douyin, a short-music video app.” People may think we were just trying to differentiate our product from others, but we [weren’t]. If you take a look at short videos both with music and without music, you’ll feel totally different [with each one]. In fact, music is more like a filter in short videos. It helps them become more expressive.
We made two apps [Douyin and Huoshan] at the time, and we wanted to make one more special and personalized [while] the other is more general. So the original idea [for] Douyin was for it to be personalized and expressive so that it could be more attractive to our target users. We also conducted in-depth research on our users. For example, we found that many young people, especially students, walk in the streets with their headphones on.
Their lives are actually quite simple: going back and forth between home and the school. Music keeps them company the most. Everyone has a need for music, and it also goes very well with short videos. So we decided that full-screen high-definition videos with music were what we wanted to achieve.
Annotation: Douyin did indeed position itself as a “music short-video community for young people” for the first two years. Before it officially changed its slogan in March 2018 to “Record your beautiful life,” it was roughly translated to “Let [the] adulation begin here, a music short-video community focused on the new generation.”
Pretty specific, right? It’s very different from the competing products at the time, whose positioning was much more ambiguous and generic. You could say this was really just copying from Musical.ly, but at the time, this would have also been risky. Very few consumer products then worked both domestically and abroad.
You could argue that Instagram might have, had it not been banned, but very close (and well-funded) clones such as NiceApp also failed to make it [in China], so I don’t know if that supposition would necessarily be correct.
In fact, before TikTok, it’s pretty hard to imagine a product that could bridge the China-and-the-rest-of-world divide, so differently do Chinese consumers seem to behave. Thus, it was crazy that this seemingly obvious combination – young people, short videos, and music – that had achieved some success abroad wasn’t that self-evident in China.
Sure, since then, I think we’ve all discovered that Gen Z globally may have more in common with each other than they do with us older folk. But five years ago? Not everyone could see that clearly. So for Zhang and her team, I can understand that this was still a “discovery” of sorts.
A screenshot of Douyin’s webpage with its original slogan “Let [the] adulation begin here.” The description says, “Douyin Short Video: a music short-video community focused on the new generation.”
It is quite different with foreign users. When we did research on overseas products, we found that they did not pay attention to these details at all. They didn’t think about adding some beauty effect to make people look prettier.
I do think there are differences between countries and cultures. When we decided to serve young users in China, we found that they really cared about special effects and beauty filters. Who doesn’t like [seeing their] prettier self? So we also put this as one of the key elements of Douyin.
Annotation: What’s funny is that the concept of a filter was originally popularized by Instagram, although of course, Chinese app beauty filters do far more than change a photo’s colors and contrast levels. They change shapes and render whole faces completely unrecognizable – er, I mean, more beautiful.
Anyway, as we all know by now, even the very aggressive filters and effects did eventually make it overseas. It’s true that China still takes it one step farther though, and for better or for worse, how much the filters can beautify a photo still remains a main criterion for users who are choosing apps to download.
KZ: Lastly, personalized recommendation. I think everyone can understand this without my talking about it. Yes, personalized recommendation is indeed what Toutiao knows best. And as a user-generated content community, it’s hard to distribute short videos uploaded by so many users without having a personalized recommendation system.
After we decided on these four key [ideas], we started to think about how to do our best in each. We’ll talk more in detail shortly.
The next step is extremely crucial: to give the product a good name.
I don’t know what you think about naming, but it’s true that many companies don’t pay attention to it; they may just come up with a few names and choose one casually. But we did massive tests on names. I remember our entire team and even the rest of the company coming up with hundreds of them.
Using these options, we did a lot of research and tests on users. We found that a name is very important for a product. When a user doesn’t know your product [yet], they judge and understand it based only on what you call it. So the name was very important to us.
And if you are in the internet industry, you will surely know that names are very important for the conversion rate as well. We spent a lot of money on testing our name options. I think this may be why we ended up with Douyin. In fact, Douyin was called “A.me” at first, which is also a great name, but it didn’t work very well in China when we promoted the product and spoke [about it] with users. After all, we’re a Chinese company making a product for Chinese people.
Douyin was designed to serve the domestic market at first, so we thought it’d be better to have a Chinese name. Although A.me sounds cool, because it means “Awesome Me” in English, it’s not intuitive enough. So we changed the name again after we had launched the product for several months.
The name Douyin is actually quite interesting. I didn’t come up with this idea; one of our project managers on the team did. I don’t know how you understand these two words but “dou” is actually a verb [that means “to shake”]. It then gives you a vivid feeling when you add the word “yin” [which means “sound,” and together, the word “douyin” is a noun that means “vibrato”].
During the era of A.me, we found that some users added special techniques like cool camera movements in their videos. We also realized that our users spent a lot of time – sometimes hours – to produce a 15-second video and make it cool. Users’ love for camera movements and special effects inspired our product management team.
Moreover, Douyin’s short videos have music, so that’s also crucial. I don’t know if you feel this, but when you hear a song you particularly like, you can’t help moving your legs to the rhythm because you really enjoy the music, you’ve been conquered by the song. We thought the word “dou” expressed the love and acknowledgement of the product, and so decided to go with that.
Some companies named their products with verbs like “pat,” but we thought that wasn’t good enough, as it wasn’t special. After we came up with Douyin, we [sent out] a lot of questionnaires and did user research. We found that – as we had assumed – the name increased people’s acceptance of and curiosity about the product. We decided it was a good name and have been using it since.
The name issue is particularly important: It is a critical step for you to reach the user, so be sure you pay attention when naming your product.
Finally, why do we call it “Douyin Short Video?” Because not only should you have a name but also a category [that explains it]. Otherwise, people will still have no idea what your product does. You don’t know what Douyin does if you just look at the name, but if you say Douyin Short Video, it’s clear to users that this is a product that shoots short videos, and you can play around with it. It’s a very unique name.
Annotation: When Douyin was launched in September 2016, it was initially called A.me. It wasn’t until December that it changed its name to Douyin.
It’s actually a bit surprising that a more rigorous approach to naming wasn’t instituted from the very beginning in beta testing. ByteDance’s first product, Jinri Toutiao – widely considered an “uncreative” and “too direct” a name since it literally means “today’s headlines” – was named via pure A/B testing.
In fact, as Zhang Yiming has alluded to in interviews, the story of Jinri Toutiao’s name is so well known that a lot of people think ByteDance A/B tests everything – even when there’s still a good amount of human judgment and intuition that go into its decision-making.
Stay tuned for part 2.
This is an edited and condensed version of an article first published on Tech Buzz China.