As one of the world’s largest short video platforms, Douyin has a large amount of content and users. However, many people may not give full play to the functions of Douyin, and there may even be some misunderstandings in their usage habits. This article will explore the correct use of Douyin from multiple perspectives, including how to optimize content recommendations through the “not interested” feature, the hidden advantages of landscape viewing, the potential of Douyin as an auditory content platform, and the logic behind algorithmic recommendations.
I know that many people think that short videos are pig food, and Douyin is nipple music.
At least the large Internet companies that have become famous in China all have the label of original sin. These stereotypical prejudices, in my opinion, are almost impossible to eliminate. After all, denouncing capital, traffic, and business is always naturally standing on the moral high ground, and against the trend, there is a sense of coercion that everyone is drunk and sober.
In this article, I just want to talk about my experience. Whether you have received money or not, you can judge for yourself.
01
I have never shied away from saying that I spend at least an hour a day on Douyin. If there are no mobile games that I am addicted to, this is the second most screen time app on my phone for a long time.
Some people denounce entertainment to death, in addition to secretly despising this person for not understanding what entertainment means to death at all, I also secretly admire that this person can actually cut off entertainment so much, and he is still alive after living so tired.
Douyin does have a lot of entertainment content, but this is also related to how individuals use it.
I rarely like, bookmark, retweet, and am unlikely to comment on it. But I like to use “not interested”. Douyin does not have an account blocking mechanism, but it is basically not interested, which is equivalent to the account being blocked. There are some well-known media accounts that are on my list of not interests, so they never disgust me.
If you are not interested, you can not only get infinitely close to the account to block it, but you can also do this kind of content, don’t recommend it to me again. For example, in order to eat the melon of the former gymnast member, I was looking for relevant information to read, but it was pushed violently for a while. And one year I knew that a Chinese died in Russia and Ukraine, and I also searched for it, and I was also pushed a similar video, but I was not interested in it for a day or two, and now it’s gone.
To achieve these three challenges, product managers will only continue to appreciate
Good product managers are very scarce, and product managers who understand users, business, and data are still in demand when they go out of the Internet. On the contrary, if you only do simple communication, inefficient execution, and shallow thinking, I am afraid that you will not be able to go through the torrent of the next 3-5 years.
View details >
The reason why disinterest is important is that you want to tell the algorithm that you are really not interested in (or not at all when you are interested). As for what interests you, you can actually open it a little.
02
The second tip: look at the screen as horizontally as possible.
After the horizontal screen, the fragmented real short video disappears in your information flow. Instead, it includes film and television drama slices and a large amount of intellectual content. If you think slicing is annoying and infringes on people’s copyrights, it’s still the same sentence: just don’t be interested.
Another advantage of the horizontal screen is that it will automatically play the next one, instead of the vertical screen that loops the current one. This makes Douyin my cooking artifact, and I don’t have to touch it with my hands, which is completely in line with the scene of eating crayfish.
It is said that Bilibili is educational software, but under the horizontal screen, Douyin has become a treasure trove of cold knowledge that I am useless – for example, I recently learned how to fill in the nationality on the ID card if foreign friends want to naturalize as Chinese through the horizontal screen of the meal.
The recommendation of Bilibili is still almost interesting, so I use the horizontal screen Douyin as a hundred thousand why. In my personal experience, Station B is a game strategy software.
03
The third skill, I learned from others, I don’t have any needs myself: Douyin can be used to listen.
An executive of Douyin mentioned such a set of data at an industry forum:
In 2024, a long video of “450 Minutes of In-depth Interpretation of Dream of Red Mansions” was posted on Douyin, which finally received more than 300 million views and 7 million collections.
It’s nothing. Because I also saw it, I watched it for half a minute, and then clicked on the collection once, always thinking that I would look at it, although I never read it again.
The next data the executive disclosed was beyond my knowledge: there were more than 1 million complete broadcasts.
450 minutes! Seven and a half hours! 1 million completed? As far as I know, watching half a minute and then pulling it directly to the end will not be regarded as a complete broadcast statistics.
When talking to friends about this, people proposed: Douyin can be listened to! Therefore, many of the broadcasts should actually be broadcast while hypnotized and sleeping, at least the system thinks that the broadcast is over.
That’s reasonable. Seven and a half hours, wouldn’t it be over when I woke up in the morning? But if you want to listen to something, why not go… Not… For example, a small universe?
Another friend said that his wife has been listening to Douyin for two years.
I thought about it carefully, indeed, there are many horizontal screen knowledge videos, and it doesn’t hurt to listen to them. The video footage, to be honest, is very low, and most of the time, it doesn’t matter at all – as if I have always thought that the header picture of the official account is purely a fart by taking off your pants.
04
Next, I want to talk about a more important thing, which is also related to how to use it.
The core word of algorithm recommendation is actually “marking”, which includes labeling users, labeling content (advertising is also content), and labeling content sources, that is, accounts. The finer the granularity of the tag, the higher the matching hit rate is likely to be.
This is at least my understanding of algorithmic recommendations for a long time.
On March 31 this year, at the 2025 China Online Media Forum, Douyin President Han Shangyou delivered a speech, he said:
Nowadays, with the development of machine learning technology, Douyin’s recommendation system almost does not rely on labeling content or users, but directly estimates the target behavior of each user for each content through a series of neural network calculations, such as the probability of liking, following, sharing, and commenting, and selects the most probable part of the content to recommend to users. This means that algorithms can directly predict user behavior without understanding content type or semantics. ”
On the official website of the “Douyin Security and Trust Center”, the core of the official summary recommendation algorithm is: learning human behavior with mathematical calculations. Compared with human recommendation, recommendation algorithms achieve an important paradigm breakthrough: it abstracts users’ specific preferences for content (such as clicks, ratings, and other explicit behaviors) into mathematical mapping relationships in high-dimensional space.
The above content can be seen publicly in media reports, and there are some things that I have obtained, but it is said that they have not been made public, so I will not show them.
I did some study, to be honest, I really didn’t understand the details that were too detailed, and the mathematical mapping relationship in high-dimensional space was beyond the framework. But I can basically understand it this way: algorithmic recommendations are identifying your “intent”.
In the current popular AI chatbot, you can see the AI pondering your intentions in the deep thinking mode, such as this:
Those who know me well know that I don’t care much about the relationship between algorithmic recommendations and information cocoons. Because algorithm recommendations focus on your interests, but interests are not positions, opinions, and attitudes, and cocoons are related to the latter. My interest in the U.S. election is not the same thing as whether I stand for Trump or Harris. In the case of non-extreme stations, the algorithm is difficult to know. On the contrary, the content of various tendencies is pushed to me, which can break the cocoon.
Maybe, this idea needs to be corrected. Therefore, it is actually very important to like, collect, retweet, and comment on the algorithm platform as little as possible. You just stare at it if you are not interested.
Of course, my real advice is to focus on objective statements and avoid touching things with strong tendencies. This kind of thing has no other value other than the cheap emotional value that resonates strongly with you.
In the final analysis, media literacy is self-training. The poop is pulled by yourself, whether it is thin, lumpy or stripped, it is your business, not the toilet.
