In a meeting in July 2022, Kuaishou CEO Cheng Yixiao asked the executive in charge of commercialization: Why is it that Kuaishou's user base is half that of Douyin, yet its commercial revenue is only 1/5 of it?
Cheng Yixiao's question reflects a common predicament faced by many second- and third-tier internet companies in China, with Kuaishou, which has a daily active user (DAU) count exceeding 300 million, being among the least troubled.
According to statistics from the data analysis firm QuestMobile, by the end of 2022, the combined DAU of content platforms such as Bilibili, Zhihu, Xiaohongshu, and Weibo reached 80% of Douyin's, but their total revenue was less than 1/8 of ByteDance's revenue in China.
Currently, there are 57 internet products in China with a DAU exceeding 10 million (excluding games), most of which belong to major internet companies like ByteDance, Alibaba, Tencent, Pinduoduo, and Meituan, with only 21 independent products. Among the non-major company products, Weibo and Kuaishou have DAUs exceeding 200 million and 300 million, respectively, while the rest are of medium scale, ranging from 10 million to 100 million, including Xiaohongshu and Bilibili with around 100 million DAU each.
In the past year, the stock prices of Chinese internet companies listed overseas have generally declined, but the decline for companies operating products with a DAU in the 10 million to 100 million range is generally higher than that of the most representative Chinese internet index fund (KWEB).
The agglomeration effect is becoming increasingly evident. The profitability of these medium-sized companies has generally decreased over the past three years. Insufficient profits mean less money available for growth, and more than half of these products have seen their DAU stagnate or decline. In contrast, except for QQ, the DAU of the top 12 internet products has all increased by more than double-digit percentages compared to pre-pandemic levels, with Douyin and Pinduoduo even doubling.
The internet industry once believed that network effects were sufficient to build a moat, meaning the more people using a product, the higher its value. In 2004, when the entire Tencent Group's products had only a few tens of millions of DAU, they had a net profit of over 300 million yuan.
But today in China, network effects have higher demands. A company reaching tens of millions or even 100 million people can no longer guarantee profitability. By the first three quarters of 2023, Bilibili, which has more daily active users than Tencent's "16-day active users" 20 years ago, had a loss of 3.5 billion yuan.
Bilibili is just a sample of a medium-sized platform with monetization difficulties. These companies have all experienced a period of rapid expansion, using checks from investors to increase their user base to eight digits, which made them less eager to establish a positive commercial cycle.Starting from the second half of 2021, the stock and venture capital markets successively contracted. After the external funding was cut off, they found themselves in a dilemma, much like falling into a trap: to increase commercial revenue and maintain company operations, they needed to further expand their user base, but the existing funds were insufficient to support this; yet, actively downsizing faced numerous obstacles.
"During the previous rounds of financing, investors did not anticipate a 'small but beautiful' outcome, and employees did not join for the sake of a 'small but beautiful' company," said the founder of an internet platform company.
If they cannot accept the prospect of shrinking, small and medium-sized platforms can only become more aggressively commercialized, but this process may harm the user experience and lose their uniqueness compared to larger platforms.
When climbers ascend the southeast ridge of Mount Everest, the last obstacle they need to overcome is a 12-meter high, almost vertical rock face known as the "Hillary Step." At the peak of commercial mountaineering in 2019, about 200 people could be stuck near the "Hillary Step" in a single day, struggling to move up or down in the extremely cold and oxygen-deficient environment.
An increasing number of small and medium-sized companies are reaching a similar place in the business world.
The Shadow Cast by Giants
For internet platforms, advertising is the most natural source of revenue. Although giants have always existed, there used to be ample space for medium-sized companies to survive under the shadow of these giants.
In 2016, Baidu, which dominated the search advertising market, recorded 64.5 billion yuan in advertising revenue, while Toutiao (Today's Headlines) was rapidly rising. However, these two platforms each had only 120 million and 60 million daily active users, respectively, covering only a small fraction of China's population. Advertisers had to allocate their budgets across multiple platforms. Smaller-scale companies like Sohu and Weibo received 7.2 billion yuan and 4 billion yuan in advertiser budgets that year, respectively, with the combined total being 17% of Baidu's. Even Douban, known for its resistance to commercialization, was confident at the time, preparing to expand its mobile applications and sell more advertising.
Six years have passed, and Douyin's (TikTok's Chinese version) daily active users have exceeded 700 million, already covering more than half of China's population, consuming users' free time for more than two hours a day. In 2022, the combined advertising revenue of Xiaohongshu (Little Red Book), Bilibili, Zhihu, Sohu, and Weibo was less than 50 billion yuan, accounting for only 15% of ByteDance's advertising revenue in China. This figure only accounts for advertising revenue. If e-commerce and live streaming revenues are included, the combined income of these medium-sized companies is an order of magnitude lower than that of Douyin.
The most significant factor affecting the industry's distribution pattern is Douyin. It has not only taken away the budgets that older internet products used to receive but has also impacted the upward potential of advertising businesses for new generation platform companies.In 2022, an employee of Weibo traveled south for a business trip to visit the company's advertising clients and introduce a new advertising product that the company had been working on for a long time. After a presentation, the client raised a doubt: "Can its delivery effect be as good as similar products on TikTok?" At that time, Weibo already had more than 200 million daily active users (DAU) according to QuestMobile's statistics.
"They asked the right question; we indeed can't catch up with TikTok," the employee admitted later. He ultimately failed to persuade the client to place an order.
Compared to other platforms, TikTok's advantage starts first with the number of users. From the perspective of advertisers, the value of media lies in helping them reach the widest potential consumer base at the lowest cost. Including the Lite version, TikTok has more than 700 million DAU. The user overlap between Xiaohongshu, Bilibili, and TikTok is between 70%-80%. Advertisers prioritize TikTok to cover most users of these medium-sized products.
"For example, to find young people, they might go to Bilibili; to find female users, they might go to Xiaohongshu, and more often, it is to address the needs of reaching specific groups," said a person in charge of a new media e-commerce team.
The scale advantage is also a cost advantage to some extent. For advertisers, allocating several teams to advertise simultaneously on multiple small and medium platforms can achieve the same population coverage as TikTok. However, considering labor costs, the overall cost may be higher than having one team optimize ad effects on TikTok.
Beyond user numbers, TikTok's "single-column video stream" product form is naturally also suitable for advertising. Liu Peng, the author of the internet advertising textbook "Computational Advertising" (@ Beimingchenghaisheng), believes that Toutiao and TikTok with "single-column + video information stream" are "advertising products of a different era." The advantage of the latter is primarily its strong immersion: "Compared to the information stream of the graphic era, short videos are full-screen and also have sound, which is a strong impact."
More importantly, the single-column information stream can feed more ads to users. Assuming a penetration rate of 5% from seeing the ad cover to clicking to watch the video, then with the same number of video views, the number of ads that a single-column product can display is 20 times that of a dual-column product.
After leveraging its product form and user scale to create a gap with other small and medium platforms, ByteDance can further build its moat with advertising revenue.In the first half of 2020, after research, ByteDance's commercialization department set a goal to transition from an "advertising business platform" to a "business operation platform," aiming to enable merchants to conduct business directly on Douyin. In June of the same year, Douyin E-commerce was upgraded to a first-level department and began to build infrastructure such as payment and logistics.
Building an e-commerce platform means being able to obtain more user behavior data, which improves the efficiency of advertising algorithms; it can also capture some of the traditional e-commerce platform's advertising revenue that originally belonged to platforms like Taobao and JD.com. However, for small and medium-sized platforms, e-commerce is a scale business, and without sufficient financial resources, it is not sustainable.
When it comes to the advertising revenue brought by self-operated e-commerce platforms, a senior B station executive said the company is "very envious," but the company's head of advertising business and COO, Li Ni, once said that B station does not plan to try. "If we haven't done a good job with the membership purchase system, what kind of closed-loop e-commerce do we want to do?"
B station's approach to e-commerce advertising is to cooperate with e-commerce platforms such as Taobao and JD.com, allowing users to jump to e-commerce platforms to purchase through internal links, and to improve the accuracy of advertising matching through the data returned by the latter, but the effect is limited.
"The data returned to various media by e-commerce customers is not rich. Those media with self-operated e-commerce platforms have more advantages in user data dimensions, and the algorithm models are also more accurate," said an advertising industry insider.
In addition to e-commerce business, with the support of advertising revenue, ByteDance also has the ability to maintain a product matrix including products such as Xigua Video and Car Emperor. It further amplifies its user scale and data advantages.
A former advertising sales employee of Xiaohongshu recalled that two years ago, he once found a marketing person in charge of a top home appliance brand, hoping to get more advertising investment for that year's Double 11. "But he frankly told me that most of the company's budget would be inclined to ByteDance."
The reason for the customer is that ByteDance can unify the traffic of Douyin, Toutiao, Xigua Video, and Car Emperor to achieve its marketing goals. "What else does Xiaohongshu have besides Xiaohongshu?"
Having a product matrix also means a better understanding of users. A former employee of Weibo reflected on the gap between the company and the advertising efficiency of Douyin, and believed that the lack of rich user behavior data is one of the factors that restrict the company's advertising business: "Weibo has 200 million DAU, and the number of users is not low, but most of them browse entertainment or political content on Weibo, making it difficult for us to know what the user's commercial preferences are. ByteDance can integrate user behavior data from different applications under its data platform to get a more complete user portrait, and the degree of advertising matching is naturally high."More users and better matching capabilities make advertisers willing to flock to TikTok for ad placements. Since most internet advertisements are sold through auction, the relationship between ad revenue and the number of advertisers is not a simple linear one.
"Compared to increasing the advertising budget of major advertisers, increasing the number of advertisers is the most important way for a platform to increase ad revenue, as this will intensify the competition in the bidding. If a platform has twice the number of advertisers as another platform, its ad revenue could be more than three times that of the other platform," said Liu Peng.
ByteDance also realized the importance of the number of advertisers early on. In 2016, when ByteDance was still in the Toutiao era and its ad revenue mainly came from major clients, it began to recruit small and medium-sized advertisers.
In order to expand the market in the shortest possible time, ByteDance chose to operate its own business. Within four years, the number of employees in ByteDance's commercialization team serving small and medium-sized clients swelled from 0 to nearly 20,000.
A former Zhihu employee attributed part of the rapid growth of ByteDance's ad revenue to its self-operated small and medium-sized sales team: "These new employees are under the performance pressure of developing new users, and they will approach any potential advertisers of any city level and any industry. Zhihu is relatively slow in this regard, and it still hopes to hold onto the first and second-tier advertisers."
After mastering advantages in user, product, data, and advertiser resources, ByteDance has entered a virtuous cycle of "profitability - ad efficiency" that mutually enhances, making it almost impossible for small and medium-sized platforms to catch up.
An employee of a platform company with nearly a hundred million daily active users said that in their own sub-industry, the company's ad conversion efficiency was once twice that of TikTok, but starting from the fourth quarter of last year, TikTok quickly completed a reversal.
"ByteDance sent several engineers to specifically help clients adjust algorithms and polish products, making TikTok's efficiency three or four times ours. Our company may not have a similar team even when several industries are added together."
How the gap was widened
In 2015, Silicon Valley investor and LinkedIn founder Reid Hoffman taught a course at Stanford University called "Technology-Enabled Blitzscaling." The term "Blitzscaling" comes from the Blitzkrieg of Nazi Germany: "You only carry the essentials, and you can move very, very quickly, surprising the enemy... Once you decide to move forward, you must go all out. You either win big or lose big."The logic of blitzscaling is similar—prioritizing growth as the company's top priority in a short period, taking on higher risks while rapidly capturing the market.
This is widely accepted as the growth path for internet companies, especially for platform products that can divide users into "creators/audiences" or "merchants/consumers," because an increase in users on one side of the platform can enhance the utility of users on the other side. When facing competition from peers, scale itself is the best weapon.
In 2016, Toutiao was being squeezed by Mobile Baidu, UC Headlines, and Daily Express, which were supported by Baidu, Alibaba, and Tencent, respectively. At a company-wide meeting in the middle of the year, an employee asked what the company's "secret weapon" was to defeat the enemy.
Zhang Yiming replied: "The best secret weapon is to run forward... The core is still faster, faster, and even faster." At that time, the first OKR he publicly announced to all employees was user growth: Toutiao's daily active user count needed to reach 66 million by October.
CEOs generally believe in growth. In 2018, when Xiaohongshu was wavering between e-commerce and community, and the company's morale was low, the remedy founder Mao Wenchao offered was "Keep growing and fuck everything else." In 2019, before going public, Bilibili CEO Chen Rui explained the company's desire to grow by saying, "A small country with few people is happy, but even if you are a paradise, you will be destroyed by powerful ships and artillery."
The difference among the companies is that almost at the same time as Toutiao's launch, ByteDance began using the same set of algorithms to push news on one side and display advertisements on the other, while Bilibili and Xiaohongshu's main sources of revenue in the early days were games and e-commerce, respectively. Their core product—community—was protected from being disturbed by advertisements.
Ten years ago, this was a common choice for platform companies and had its rationality: why be aggressive in monetization and disturb user experience when there are investment institutions to provide blood?
Xiaohongshu, Zhihu, Ximalaya, Bilibili, Kuaishou... These platform companies in the fields of graphic, audio, and video were established around 2010 and then encountered the big boom of China's venture capital industry. PitchBook data shows that in 2009, only $3.2 billion went to start-ups, and by the peak in 2018, this number had increased 41 times to $131.2 billion.
At the beginning of 2016, Bilibili, which originally only wanted to raise $50 million, was stuffed with $100 million by Zhengxin Valley Capital. In the same year, Zhihu also completed a $100 million Series D financing round. Afterward, Zhihu CEO Zhou Yuan recalled in a letter to employees: "The deal was settled before I had time to write a PPT, and the money arrived quickly."
According to the exchange rate in 2016, a round of $100 million in financing was higher than the average size of an A-share IPO. In an environment where financing is not difficult, the advertising business can be slow, and then slower.A person close to the company's senior management said that Bilibili did not bring in executives from outside until 2018, when it "truly began to value advertising"; Xiaohongshu opened up advertising solicitation a year later. Zhihu and Kuaishou, which started a bit faster, began their advertising business in 2016 and 2017, respectively, but this was also after their products had been online for 5 and 6 years.
Delaying monetization has been exchanged for user growth, but these platforms are still not running fast enough. It took Xiaohongshu and Bilibili 10 years to reach 100 million daily active users (DAU), Kuaishou's main site took 7 years, while Douyin only needed less than 2 years. According to QuestMobile, by the second half of 2023, Douyin Group's de-duplicated users had already reached 1 billion, covering almost all mobile internet users.
"Users have to use many applications every day. As an advertiser, after placing ads in one super app daily, the remaining apps may not need to be advertised anymore, because it is not necessarily required to repeatedly cover the same group of users," said an advertising salesperson from a platform company with a DAU of over 20 million. This means that small and medium-sized platforms want to compete for advertisers' budgets, they need to spend more effort and tell their own unique stories to compete for advertising placements.
Another more covert threat is that large platforms with healthy advertising revenues can use a free model to impact other small and medium-sized platforms that rely on user payments for monetization.
The online literature giant China Literature has long relied on users paying per thousand characters, with the platform then sharing revenue with authors. However, starting from 2018, free novel platforms such as Qimao and Tomato Novels have emerged one after another, eroding China Literature's user base. In mid-2021, Tomato Novels' DAU reached 30 million, more than the sum of China Literature's paid reading platforms such as Qidian, QQ Reading, and Chuangshi China.
A senior executive at China Literature described the impact of free products on the company in this way: "The old enemy, even if I let him use one hand and one foot, he couldn't catch up. Now, even if the enemy lies at home for half a year, we can't catch up." In 2019, at the most pessimistic time for investors, China Literature's stock price once fell by 80% compared to its listing price.
Similar things also happened to Ximalaya. Since its launch in 2013, Ximalaya has experienced a stable expansion period of 7 years and has taken membership subscription fees as the main source of income. However, when Byte's free audiobook product Tomato Changting was launched in 2020, Ximalaya found that some of its users were diverted by Tomato Changting. "Some users will not come back after going to Tomato. Although our content quality is high, it can't resist being free," said an Ximalaya employee.
Difficult Self-Help
As the pressure to monetize grows, the way out for small and medium-sized companies is not complicated: cut off branches unrelated to the main business, reduce the size of the company, and then use the most direct and effective way to increase revenue.
When iQIYI went public in March 2018, it mainly made money by selling memberships and collecting advertising fees, with the two sources of income being roughly equal, accounting for more than 86% of the revenue. However, the company does not intend to stop there.In May of that year, at the iQIYI World Conference, iQIYI CEO Gong Yu stood in front of a slide depicting eight apple trees and announced that iQIYI's business would cover eight areas including the mall, games, live broadcasting, and comics, with each apple tree representing one of them, and the closest fruit was within reach.
Based on these business scenarios, iQIYI planned nine monetization methods, including advertising and paid memberships, which were "already very complete." The following year, iQIYI had 21 products under its umbrella.
As the business expanded, by the end of 2019, iQIYI's net loss reached 10.3 billion yuan, nearly three times the size compared to the year before its listing. At the same time, changes in the broader environment led to a contraction in one of the company's revenue pillars, advertising.
"By 2019, it was already evident that advertisers were shifting their budgets from brand advertising to performance advertising," recalled an advertising industry professional. Long video platforms excel in pre-roll and embedded advertisements, which are forms of brand advertising, while performance advertisements scattered in short video information streams can track and optimize user behaviors such as clicks, favorites, and cart additions, thus becoming increasingly popular with advertisers.
QuestMobile data shows that in 2018, the online video industry, where iQIYI is located, could still secure 7.4% of the advertising market share, but by 2022, this figure had dropped to 3.2%, while the market share of short video advertisements increased by 2.2 percentage points. "Starting in the second half of 2019, the company gradually became aware of the pressure from short videos. Advertisers' budgets are limited; some invested in short videos and stopped investing in long videos," said an iQIYI employee.
By the end of 2021, iQIYI's available funds reached the lowest point since its listing, and the pressure to repay debts increased sharply. At this time, it began to take self-help measures. The company significantly reduced staff in its content business group, new consumer business group, and strategy department, the research institute and game center were halted, and funding for the short video product "Sui Ke" was also cut off. By the end of that year, the number of iQIYI employees was 24% less than the previous year.
At the same time, iQIYI increased membership fees and reduced discounts. A few quarters later, the growth in membership fee revenue offset the decline in advertising and other innovative business revenue. By the third quarter of 2023, more than 60% of iQIYI's revenue came from membership fees, and the company has been profitable for four consecutive quarters. The cost was that iQIYI is no longer seen as a "technology company," but more like a traditional TV station, with a market value that is still lower than that of Mango Super Media, a subsidiary of Hunan Broadcasting Group.
Long video platforms generally shifted from advertising to membership fees to turn losses around and support development. However, more companies have failed to explore new monetization methods. Zhihu, which struggled to find a suitable monetization path, once had to rely on serving more advertisements to users, which also impacted its positioning as a "professional knowledge community."
The launch of the "Zhi +" system in 2020 marked the beginning of Zhihu's aggressive commercialization. Through Zhi +, advertisers could contact authors within the community to write marketing content and purchase traffic for that content. Marketing content pushed to users' fingertips by Zhi + helped the company's advertising revenue triple in two years, but it also eroded users' trust in community content.
A Zhihu employee said, "In 2017, what a Zhihu big V said might be taken as the truth, and with the title of a Zhihu big V, one could even give lectures and conduct training. Now, when people mention Zhihu, they might think it's all advertisements."Products similar to Zhi + can also be found on Xiaohongshu and Bilibili, but they have a smaller impact on the platform ecosystem. "The content on Xiaohongshu is sharing-oriented, and creators are not required to be very professional or authoritative," said a Zhihu employee. Bilibili, which relies more on the trust relationship between the audience and the UP main, is more cautious about advertising. When Zhihu's calculated advertising load rate was nearly 15% in 2021, Bilibili's advertising load rate was only 5%.
At the same time as the launch of Zhi +, Zhihu began to try a single-column video stream. An employee who has been involved in the videoization of Zhihu since 2021 said, "At that time, the background was that everyone recognized the higher commercial value of video. The same traffic can be sold at a higher price, which is very tempting for the platform."
The problem is that videos are not compatible with Zhihu's original content structure. For users who come to find answers with questions, the efficiency of presenting information in videos is far less than that of text and pictures.
To support video content, the Zhihu management once set the "number of 6-second video views (VV)" as the core assessment indicator, and the team controlling the homepage recommendation information flow also reported to the person in charge of the video business department. The department with performance pressure has the right to distribute traffic, and naturally, there is more video content, but this interferes with the user experience.
At the end of 2021, in the article where Zhihu's official account explained the 8.0 revision, the first comment in the comment area was "Can you add a 'close video' function? That is, there are no videos in all sections?"
QuestMobile data shows that the year-on-year growth rate of Zhihu's DAU began to slow down in the second half of 2021. In December, the company experienced the first year-on-year decrease in DAU since its listing. In the middle of 2022, the video strategy was terminated, but the company's advertising revenue has fallen into a year-on-year contraction situation and continues to this day. Currently, Zhihu's largest revenue-generating business is the membership mainly used for reading web text, and the company has not yet achieved profitability.
Compared with Zhihu, which is based on Q&A and emphasizes neutrality, the advertising environment of Xiaohongshu and Bilibili is relatively friendly, and they are also trying to further expand their advertising revenue. The most important way to achieve this is to continue to expand the platform's volume. This may be inspired by Kuaishou.
In 2019, after witnessing the rapid rise of Douyin, the management of Kuaishou launched the "K3" campaign to impact 300 million DAU, and launched a single-column information flow product similar to Douyin, "Kuaishou Speed Edition," with the full strength of the company, expanded the scale of the operation team, and stimulated users to settle in with red envelopes. These are all extraordinary measures that Kuaishou had not taken before. "The company has entered a life-and-death battle," said a Kuaishou employee.
After the Spring Festival of the following year, Kuaishou achieved the user growth target and laid the foundation for the expansion of the advertising business in the future. In the first three quarters of 2023, Kuaishou's advertising revenue was 42.1 billion yuan, and it has achieved profitability.
At present, Xiaohongshu's latest DAU target is also set at 300 million, and Bilibili proposed last year to double the DAU of the main mobile terminal. This process will not be easy.An employee of Xiaohongshu who participated in urban research at the beginning of last year stated that the company assessed that Xiaohongshu's penetration rate in first- and second-tier cities is approaching its ceiling, but further expansion into lower-tier cities is challenging.
"Users in lower-tier cities may prefer content that is directly pushed to their faces, but Xiaohongshu wants to maintain a dual-column layout on the homepage to differentiate from Douyin and Kuaishou. When supplementing content supply, the strategy team also indicated not to use AI-voiced second creations in the style of Douyin and Kuaishou, even though research results show that users actually like it."
Bilibili considers user-generated videos (UGV) as one of the key drivers for user growth, while professionally created videos by UPs (PUGV) are the main reason users open Bilibili every day. If the recommendation algorithm's capabilities do not improve in sync with the expansion of UGV supply, users annoyed by low-quality UGVs may leave.
The market's patience is rapidly diminishing. According to private equity data provider Altive, compared to the peak in 2021, Xiaohongshu's valuation in the primary market has nearly halved this year, and the stock prices of Bilibili and Zhihu have also dropped by more than ninety percent. Such a decline always reminds investors of the scene at the beginning of 2000 when the internet bubble burst.
The sense of disillusionment behind the two major drops is similar: a company can buy growth with investment funds and tell a story; but it does not have the ability to maintain its market value or even user base.
A former employee of a content platform with over 20 million daily active users (DAU) said that he joined the company "with a passion" during its rapid user growth, but then the company's several hasty attempts at commercialization impacted the platform's original positioning. After witnessing the departure of core users and the company's inability to make a profit, he eventually chose to leave the place where he had worked for 5 years.
"It will just exist," he replied somewhat pessimistically when asked about the company's future. This may also be the ultimate fate of most medium-sized platforms.
Comments