Douyin, Kuaishou, Xiaohongshu, and Bilibili, which one has a lower advertising ceiling?

In October, the article "Watching Douyin and Using Kuaishou" demonstrated the core differences between Douyin and Kuaishou: the single-column and double-column product designs lead to differences in fault tolerance, and the differences in the core optimization indicators of the recommendation algorithms ultimately shape the attribute differences between Douyin and Kuaishou, one being a strong media type and the other a strong community type.

The difference in fault tolerance can be further explained by the data differences caused by differences in user behavior.

In the feed stream of Douyin, users can scroll down infinitely and choose to stay or not, but cannot choose whether to watch a certain video. In the double-column feed stream of Kuaishou, users select the video they are interested in and click to watch it. After browsing/liking/commenting, they exit to the double-column feed to continue browsing. This means that the data collected by the two products are basically the same: video completion rate, like rate, comment rate... and Kuaishou has one more link and one more unique data indicator: the CTR from the waterfall flow to the click-through content viewing.

Clicking on the cover to enter the details page is an action that expresses strong subjective intention . This highly expressive data enables the algorithm model to accurately capture user intentions. In contrast, user behaviors such as how long a user stays on a video and how often they refresh the video are not necessarily clear, and require the algorithm system to spend more time calculating and identifying them. Users who use both Douyin and Kuaishou often feel that Kuaishou iterates faster and can more accurately reflect users' immediate interests. To some extent, this is a difference caused by the different dimensions of input data.

Single column + relatively weak user behavior expression vs double columns + strong user behavior expression allows Kuaishou to tolerate showing more diverse content to users, while requiring Douyin to focus more on top content, resulting in a difference in tolerance. Combining the differences in core optimization indicators of Kuaishou's "submission rate > average VV per person > average number of followers per person" and Douyin's "average VV per person > submission rate > average number of followers per person", we ultimately see a series of differences between the two in terms of community atmosphere and content consumption efficiency.

There is one more link in the product funnel. In today's world where advertising is content, this model can be further extrapolated to explore the advertising monetization potential of the platform.

The impact of single-row and double-row product differences on advertising

From the perspective of the user's content consumption chain, the double column has an additional layer of click jump compared to the single column. Therefore, in the funnel model of advertising commercial products, the double column should add an additional layer of CTR.

For products in a single-column model, the advertising revenue formula is easy to understand, which is DAU (daily number of users) multiplied by average VV (the amount of videos watched per person per day) multiplied by Ad Load (ad load) multiplied by CTR (ad click-through rate) and then multiplied by CPC (ad click price). In the end, you can get the total revenue scale of the product.

For double-column products, the formula needs to be modified to further break down the per capita VV indicator into total exposure multiplied by the CTR from the cover to the content. Per capita VV = exposure × CTR (according to the data from Kuaishou's "2019 Kuaishou Creator Report", Kuaishou's CTR is about 20%). In the mature stage of the product, when the user time is similar, the per capita VV of single-column and double-column products is basically similar.

For example, suppose two products are single-column and double-column short video products, with an average daily duration of 60 minutes, an average video duration of 30 seconds, and an AdLoad of 15%, that is, the average VV per person of both products is 120. For single-column products, the number of ads displayed is equal to average VV per person × Ad load = 18.

The content exposure click-through rate CTR0 of the double-column product was 20%, 600 videos were exposed to users, and 120 VVs were obtained. The number of ads displayed = total exposure × Ad Load × CTR1, where CTR1 is the click-through rate of the ad.

Ideally, ads and content are completely equivalent, and the click-through rates of ads and content are exactly the same, which is 20%. In this case, the number of ads displayed by the double-column products is also 18. However, in reality, the click-through rates of ads and content for users cannot be the same, and the ad click-through rate CTR1 may be lower than the content CTR0.

According to interviews with actual practitioners, the gap between CTR1 and CTR0 is larger than expected, even reaching 5-10 times. That is, if the CTR of the content is 20%, then the CTR of the advertisement may be only 2%-5%, which means that the advertising inventory of double-column products is much lower than that of single-column products. However, the good news is that active clicks mean a clear representation of user intent, and double columns can theoretically charge a higher price for each click, similar to the unit price of search ads being much higher than display ads.

In general, single-column products are more suitable for advertising commercial monetization, and the advertising ceiling of double-column products will be lower than that of single-column products.

Based on the actual data, we selected representatives of domestic and foreign single-column products: Facebook, Twitter, Weibo, Toutiao, and Douyin, as well as representatives of double-column products: Kuaishou, Bilibili, Xiaohongshu, and Pinterest, and normalized the company's monetization capabilities according to the user's advertising monetization capabilities per hour of viewing.

The advertising monetization capabilities of double-column products are weaker than those of single-column products. However, considering that several companies with double-column characteristics are generally in the early to middle stages of commercialization, there is still room for growth in the future.

There are some interesting numbers here that are worth discussing.

From the internal comparison of single-column products. If Facebook is used as the monetization ceiling for a single product, we can see that the biggest difference in Toutiao's monetization efficiency comes from the exchange rate difference. After excluding the exchange rate factor, the difference in monetization efficiency is around 20%, which is already at the same level. After excluding exchange rate factors, the monetization efficiency gap between Weibo and Twitter is about 30%. These two Weibo companies are considered to have weak monetization capabilities both at home and abroad. Compared with Facebook and Toutiao, the gap is indeed obvious.

In order to snipe Kuaishou, Douyin has cut the time spent on live streaming this year, and its advertising monetization ability has been affected to a certain extent. Otherwise, the data would be much higher than the current data.

The normalized monetization capability of Xiaohongshu is almost the same as that of Douyin, and there is still room for growth in the future, reflecting the powerful monetization capability of women's vertical communities. Judging from Xiaohongshu's products and community atmosphere, Xiaohongshu's content and advertising content are highly unified, users are highly receptive to advertising, and a large number of soft ads are distributed in the content, so that the gap between the CTR0 of the content and the CTR1 of the advertisement will not be exaggerated, but will be closer. The precise user base and clear click intent allow Xiaohongshu to charge higher prices for advertising.

Xiaohongshu plans to double its advertising revenue next year. While focusing on user growth, it also strives to expand from beauty products to fashion and beauty categories to increase the source of advertisers.

Pinterest's unit monetization capability is higher than Twitter's. The logic of this double-column waterfall flow pioneer is similar to that of Xiaohongshu. Users browse content in order to find shopping, home furnishing or design inspiration, and are naturally highly compatible with advertising content. In terms of commercialization, we continue to acquire technology companies to strengthen our ability to make precise recommendations. At the same time, we launch various advertising formats and develop new commercial products such as shopping tags and e-commerce shopping guides.

Bilibili's advertising monetization capability is indeed too poor. It is limited by the advertising inventory in the double-column mode. The length of small videos is longer than that of short videos, which makes the average VV per person far lower than that of short video products. In addition, Bilibili's own relatively shallow pool of advertisers causes the same ads to appear repeatedly, ultimately leading to extremely low advertising monetization capabilities. Judging from Bilibili’s financial reports in recent quarters, advertising growth is mainly driven by brand advertising, and user growth remains the core driving factor.

Currently, Bilibili is testing a new, larger cross-waterfall advertising format, but it is still very cautious about the form of advertising within videos. For Bilibili, maintaining the community atmosphere remains its top priority.

Kuaishou does not have as many burdens as Bilibili. In order to break through the advertising inventory limitations brought by the double columns, Kuaishou has launched in-video pop-up ads and post-roll ads this year, and has tried to open up new monetization models through various methods such as challenge competitions. Kuaishou Express Edition, which has grown rapidly recently, adopts a single-column product format and may become a breakthrough for Kuaishou's advertising monetization in the future.

In conclusion, the funnel model path of single-column products is shorter and more suitable for advertising monetization. The ceiling of advertising monetization of double-column products will be lower than that of single-column products, but they have more competitive advantages in terms of user retention and duration. It can consider directing users to other monetization methods, such as Kuaishou's live broadcast and sales, and Bilibili's live broadcast and games.

Author: Pan Luan

Public account: Luanbooks (ID: luanbooks)