As an operator and trader, the first step is to have a thorough understanding of user portraits, content features and recommendation mechanisms. Otherwise, if you start blindly, the biggest result may be a hasty end.

1.

What is the algorithm?

To put it simply and popularly, an algorithm is a set of evaluation mechanisms. This mechanism is effective for all users of the platform, whether they are content producers (people who shoot videos) or content consumers (people who watch videos). Many times we are both producers and consumers.

Every action we take on the platform is like a clear instruction, and the platform judges our nature based on these instructions.

We will be divided into high-quality users, silent users, lost users, recoverable users, etc.; we will also be judged whether we are a marketing account and whether there are any illegal operations.

If it is a marketing account that does not meet the tone of TikTok, the platform will lock our account in a dark room; on the contrary, if it is judged that we are a high-quality user, the platform will give us certain support.

►What are algorithms used for?

The greatest use of algorithms for the platform is to manage user data on its own platform, and improve platform functions based on a series of user feedback behaviors, improve user experience, so that the platform attracts and retains more users, and ultimately forms a virtuous and recyclable ecosystem for the platform.

►What are the benefits of algorithms?

The benefits of algorithms to content producers: Since we want to attract fans for ourselves in other people’s territory, we must understand other people’s rules.

Just like chasing a girl, you have to understand the girl's likes and dislikes, and only then can you have a chance to get into her heart. What's more, it's much easier to understand the platform than to understand girls!

As long as we are aware of the platform's recommendation mechanism, we can consciously design our own behavior and guide the platform to determine that we are high-quality users and assign us more and more accurate flows and higher permissions.

The benefits of the algorithm to content consumers: If you browse Taobao, Douyin or Toutiao a lot, will you feel that a lot of the recommended content you like and are interested in watching when you open the APP next time?

In fact, the platform aims to improve user experience and retain you. It will analyze your interests based on your behavior, then tag you and recommend content from content producers with similar tags to you. You are in the same pool.

The process of Douyin’s recommendation algorithm can basically be summarized by the following figure:

The first step: review

After uploading the video, the first step will be machine review.

Then the review includes the video screen, title keywords, and video copy; for example, whether there are advertisements, whether there is a watermark or LOGO, whether the content is nudity, indecent, bloody, etc. If there is content prohibited by the platform, our video will be bounced or blocked (only you can see the content you posted).

If there are violations, it may enter the manual reviewer stage in advance.

Step Two: Intelligent Distribution

If there are no keyword violations or screen problems, the system will combine the keywords to match around 200 to 300 users, which is what we call the initial pool.

After the video is released, the system will tag your video according to the video content (such as travel, beauty, food, Chongqing, Xi'an, islands...), and then the machine will recommend it to people who may be interested in your video tags on a small scale, and calculate the number of comments, likes and shares from the audience within a unit of time.

The specific formula is: popularity = A × video completion rate + B × number of comments + C × number of likes + D × number of shares. The coefficients A, B, C, and D will be fine-tuned in real time according to the overall algorithm. Generally speaking: play volume (completion rate) > likes > comments > forwarding. This is the first recommendation.

Step Three: Expand Recommendations

If your video gets better audience feedback after the first recommendation, then your video will be recommended to more potential viewers. This is called expanded recommendation.

The mechanism is the same as the first recommendation, and the number of viewers reached this time is about 1,000-5,000.

The platform with better feedback on the second recommendation will recommend it a third time, and the third time will be tens of thousands or hundreds of thousands of traffic, and so on. If the feedback is still good, the platform will use a big data algorithm combined with a manual review mechanism to measure whether your content is acceptable.

According to the practical experience of DouShang Commune, generally within an hour after a video is released, if the video views reach more than 5,000 times, the number of likes is higher than 100, and the number of comments is higher than 10, it will basically be successful.

Therefore, please remember the following numbers: 1-5000-100-10;

What does it mean? It means that the video you publish should have more than 5,000 views within an hour, more than 100 likes, and more than 10 comments. Then, the chance of being recommended by the system is much greater, and it is basically not far from being popular.

2. What kind of videos have the potential to become hits?

I believe that when everyone operates an account, they often feel a sense of expectation and even anxiety. That is, after the video is released, I will check the playback volume from time to time to see if this video will become popular.

According to the actual operation of DouShang Commune, we actually pay more attention to the first hour after publishing the work. Under normal circumstances, if the traffic provided by the system in the previous hour exceeds 5,000, the likes exceed 50, the play-to-like ratio is 100:1, and there is a 90% probability that it will enter the next larger traffic pool.

If in the next traffic pool, the user interaction exceeds 100:1; it may continue to explode, and generally it will be a bottleneck when 10 million playbacks occur (after the video becomes popular, the human will review it again, and if the human feels that the content has any edge, the traffic will be withdrawn directly).

This is why many people are confused, why the system suddenly stops recommending even though the traffic is running well.

In fact, sometimes even if the content is well done, it may not be liked by the system. The decision lies with the platform. In particular, sensitive factors such as politics, privacy, exposure, and mutual fans may be suspended or even banned directly.

But if you want to enter the million pool and make a small splash, you must maintain a like rate of around 10% in the first hour. Otherwise, no chance. Entering the 10 million pool requires a higher like rate.

3. Why does a 1 million views require a like rate of 10%?

According to our practice and feedback from members of DouShang Commune, 3% of the likes rate is the average rate of Douyin, so if your work does not get 3% of the likes rate in the first hour, there is basically no possibility of becoming a hit.

For a work to impress your fans, it needs a like rate of at least 3%.

When a work is pushed to the mobile phones of strangers with tens of thousands of views, if there is no early 10% like rate, it will be very difficult to cope with how many likes strangers can give you after tens of thousands of views.

A work with a 10% like rate in the early stage will drop sharply after it enters the big pool, because your work is vertical, fans will like it, but strangers will not like it so much.

After giving you tens of thousands of views, the like rate will drop to about 7%. This is still a very good playback rate and can continue to support your work to roll into the 100,000-level pool and reach the 1,000,000-level pool.

After entering the million pool, a bigger wave of people will come, so your like rate will further drop to around 5%, or even around 3.5%. This is why your work feels like it's stalling. Because it has reached the average of Douyin.

Therefore, after a work is released, you can predict whether it will become popular within one or two hours, or about 30 minutes. Examples of overnight fire or future fire belong to another category - that is, what we have said is secondary recommendation and grave digging; the applicable algorithms are different.

What is a secondary recommendation?

The so-called secondary recommendation is commonly known as "grave digging". In fact, it comes out through the closed-loop effect.

For example, after your work was released, Douyin actively pushed 2,000 views and received 0 likes. It must have shut down. But if there is a closed-loop effect, if others come in through other videos and like this work once, if 200 people can come in and all like this work, they will get 200 likes. At this time, this work has reached a 10% like rate.

Douyin is scanning at any time and will restart the push of this work. This is called grave digging. It's just an algorithm behind it. Second recommendation and grave digging all have this meaning, and this is the principle behind it.