The article "How the Flywheel of TikTok Was Started" written by Tie Shun, the founder of Blur, is extremely valuable. He learned a lot from the success of TikTok. You can fully see the shadow of this article in Blur's operation and strategy. It is worth reading for every web3 builder.
Original text (please forgive the inaccuracies in the translation)
How does TikTok guide the flywheel?
Douyin’s success is attributed to Bytedance’s personalization engine, but the personalization engine alone doesn’t explain how Douyin got off the ground. When Douyin first launched, the functionality was so poor that early employees were embarrassed to ask influencers to post in the app. It didn’t make sense to me that the algorithm alone could explain their success, especially before Douyin had much organic content to recommend. After getting caught up in this curiosity, I learned that while Douyin’s much-lauded personalization engine contributed to their success, getting the flywheel going was largely due to the efforts of their operations team.
start
When ByteDance launched Douyin, the short video market was already hot; Douyin had been around for 2.5 years, and China's leading short video app Meipai had 160mm DAU and was becoming celebrities from their popular creators. The leader of the pack is Kuaishou, which has targeted rural consumers from third-tier cities and below.
In order to figure out how to make people like their product in such a crowded market, the entire Douyin team conducted extensive competition research. Team members tried almost all short video products from China and abroad - about 100 applications. Everyone in the team uses short video products every day to find out the strengths and weaknesses of existing players.
According to Kelly Zhang, CEO of ByteDance China, the culmination of this research was the TikTok team realizing that none of these products impressed them. They decided to focus on four dimensions to differentiate: full screen and HD video, music, filters, and personalized recommendations. In effect, TikTok decided to launch a Musically clone. This *should* have failed since Musically had already attempted to launch in China a few years ago, and as mentioned, the app launched embarrassingly badly. Additionally, Musically was founded by a Chinese team, so TikTok didn’t have the local cultural advantage either. The difference that led to TikTok’s success was the operational effort their team put in to launch the app.
Guide culture
Douyin was positioned as a trendy app for the fashionable urban elite. To ensure that early adopters had a culture that reflected this positioning, Douyin recruited early adopters by hand in large numbers. They mobilized their entire staff to attract creators from competitors, scouting for talent on other social networks and short video apps (including DMing music creators). To get fashionable urban youth to create content, Douyin sent team members to art schools across the country to find beautiful students to become its users. Overall, the team persuaded hundreds of people to join by promising to help them become famous online.
After recruiting creators on the app, TikTok treated them like royalty to keep them. Team members chatted with creators every day, listened to their ideas, and made them feel involved in the early growth of the platform and shaping its direction. They sent custom swag to creators and even sent celebratory videos on employees’ birthdays. The best video creators received gifts like cameras, celebrity merchandise, and snacks to make them feel special. One staff member even bought a creator a Christmas tree. These efforts helped retain creators until user growth was enough to retain them on its own.
The operations team then began manipulating the visibility of videos to encourage the type of trendy content they wished to cultivate. Videos that did not fit the tone and values of the community would struggle to gain visibility. This move was particularly important given Douyin’s remix feature. Like Musically, Douyin uses theme-based challenges to encourage content creation and encourages users to remix each other’s creations to build on shared memes. By highlighting challenges and content they wanted to remix, Douyin catalyzed the creation of more content and reinforced the culture they wanted. Users can submit their own challenges, of course, and the operations team often collaborates with users on new challenges.
Creating Lightning
Douyin’s operations team scoured the app for content that could go viral and amplified it. They set up accounts on other platforms to post watermarked content — including WeChat, which was a big channel for them before WeChat banned them — and go beyond virality. For example, when a video parodying a famous Chinese comedian surfaced on Douyin, the operations team repeatedly played the comedian on social media until he shared the watermarked video, resulting in millions of views.
TikTok’s ultimate source of power is owning consumers’ attention, but intuitively the way they achieve that power is by leveraging distribution through existing social channels. This reminds me of Instagram being used to post more beautiful photos on Facebook, and Airbnb automatically posting listings to Craigslist in its early days.
Why Tik Tok Succeeded Where Music Failed
When Musically launched their Mindie competitor, they had already spent 90% of their seed funding on a failed education app, with less than $30k remaining. To increase their odds of success, Musically launched globally, resulting in them gaining traction in the US but bombing in China. As TikTok proved, it was possible for Musically to win the Chinese market, but doing so would require an extensive operational effort that Musically could not afford. Given the information and funding they had at the time, it made more sense for Musically to focus on the US market.
Musically eventually re-entered the Chinese market a few years later, but by then TikTok had already gained traction in China. There was little that Musically could differentiate, as TikTok was already a clone of Musically, had a better recommendation engine, and had a bigger wart than Bytedance.
The importance of personalization
In TikTok and Sorting Hat, Eugene Wei insightfully argues that Bytedance’s personalization engine contributed to Douyin’s insanely high engagement rates and allowed Bytedance Music (later renamed Douyin) to break out of the lip syncing niche it was stuck in. However, the success of Kuaishou, and even the success of Bytedance Music before the Bytedance acquisition, shows that one doesn’t need an advanced personalization engine to get off the ground. Music grew to millions of users, using a less sophisticated personalization engine, and Kuaishou succeeded without such a sophisticated algorithm (listing at a $200B+ MKT cap) — long term, Kuaishou users consume up to 50% of their content from in-app visitors they follow, while TikTok users consume 80-90% of their content from new accounts that become popular among other users. Douyin’s operational efforts were a more important factor in getting off the ground than their more commonly cited personalization engine.
Other ideas
TikTok’s operating playbook is similar to Andy Johns’ technique of “white hot coals” for building a small but engaged community before scaling. Building culture requires a lot of manual effort, and it’s not a problem that can be solved by just throwing money at it. In fact, having lots of money and a large audience can even be an inhibitor, because it creates noise that makes it difficult for a particular culture to form.
Bootstrapping culture is very heavy on the action side, and is a departure from the playbook of previous social networks, which relied on great products and recommendations alone to succeed. If I were starting a new content community from scratch, here’s the playbook I’d use:
Conduct extensive, globally competitive product research before designing your app. Copy the best features
Choose a niche community to focus on first
Recruit content creators from existing competitors and other social/physical networks
Treat them like royalty so they will create content until the consumer side itself is big enough
Manually highlight and reward good content that you want to see more of
Use remixes and share memes to increase the creation of that content
Leverage existing social networks to distribute content and look for opportunities to create virality
Then, feed the flywheel by improving personalization and scaling vertically. Andy Johns has a great framework for how to do this