How to play with Tik Tok’s recommendation rules?

How to play with Tik Tok’s recommendation rules?

How to create a hit on Douyin? What are the six models of Douyin conversion?

1. How to play with Tik Tok’s recommendation rules?

The latest official data from Douyin shows that last month, the number of daily active users of Douyin exceeded 320 million. What is the concept? In the Q1 2019 financial report, Weibo's daily active users were 200 million, which means that Douyin now has nearly 100 million more daily active users than Weibo.

In addition, in terms of user usage time, Douyin users use the app for 78 minutes, which is also the longest among all domestic apps.

It took Douyin less than three years from the launch of version 1.0 in September 2016 to its current 320 million daily active users and 500 million monthly active users. Among them, the overseas version of Douyin currently covers 150 countries and is the most successful Internet product that has gone overseas in China. When it was first launched, the 1.0 version of Douyin was just a tool for shooting and making videos, similar to the photo editing tool Meitu XiuXiu.

By the 2.0 era, it has evolved into a community for young people, with the slogan being "a music short video community focusing on the young new generation."

In 2018, TikTok entered the 3.0 era and officially transformed into a mainstream mass community, with its slogan adjusted to "record a beautiful life".

In 2019, in order to make up for its shortcomings in social networking, Douyin connected to Duoshan’s social relationship chain and officially became a community and social Internet product.

This community plus social product form and positioning directly affects the traffic distribution and recommendation logic of this platform.

At present, domestic traffic is mainly divided into two types: one is private domain traffic, and the other is public domain traffic.

Only when we follow this account can we receive the information released by this account. This is private domain traffic.

WeChat is a typical private traffic pool. For example, if my WeChat public account has 1 million followers, then the content I publish will be seen by 1 million people. Weibo is also a private traffic pool mainly based on social networking.

Douyin mainly relies on public domain traffic, while also taking into account a small amount of private domain traffic. This is in line with Toutiao, and the distribution of public domain traffic is completely determined by algorithms.

The traffic distribution logic of Douyin is: when a video is released, a first-level traffic pool is allocated first, also called the initial traffic pool - the size of the initial traffic depends on the weight of the account. The higher the weight, the larger the initial traffic pool will be. The algorithm will determine whether the video is popular with users based on the data feedback from the initial traffic pool. If the values ​​are higher than the average, the background algorithm will continue to distribute it to the secondary and tertiary traffic pools.

This distribution logic completely abandons the traditional layout method in which journalists and editors decide the quality of content or its placement.

On TikTok, we often see that the accounts behind popular videos often do not have many followers, maybe only a few thousand or ten thousand! This is determined by the algorithm recommendation model.

So, how does the background algorithm screen videos and how does it measure the quality of videos?

There are nearly a hundred dimensions here, the main ones include (but not limited to) completion rate, like rate, comment rate, and forwarding rate. If they are all higher than the average, the algorithm will automatically determine that it is a high-quality video and obtain more traffic distribution.

In this way, we can judge the quality of the content ourselves, mainly based on these two dimensions.

The first-level data dimensions include the number of views, likes, comments, and reposts.

What do these indicators represent? The number of views represents the excitement of the content, the number of likes represents the user's recognition, and the number of comments and reposts represents the value of the video's so-called social currency.

The secondary data dimensions mainly include attention rate, like rate, comment rate, and forwarding rate.

The follower rate is the number of followers a video converts divided by the number of views a video receives.

For example: a video received 1 million views and brought in 10,000 fans, which is a relatively high conversion rate. The same video may have 1 million views or even 10 million views, but its fan conversion rate may only be a few thousand people, which is an extremely low fan conversion rate.

Why does this phenomenon occur?

The video has a high number of views, but users are not willing to follow us after seeing this video. The reason is that the account lacks its own positioning and personality, and lacks an overall content planning and packaging, so users have no expectations of us after watching our video.

It can be seen that we must not only create popular content, but also persevere in adhering to character settings and content planning to allow users to have continuous expectations, so that we can gain more fans through content. Otherwise, the result may be that the video becomes popular but the account does not.

Behind the like rate is the content quality of the video, behind the comment rate is the interaction index, and behind the forwarding rate is the value index of social currency.

After an account uploads a video, it will be distributed in the initial traffic pool, and the size of the traffic pool is determined by the weight of the account.

What is the concept of account weight? It is similar to the credit index of an account and is a standard used by the algorithm to measure the quality of the account. High-quality accounts can obtain relatively high initial traffic.

In the general sense, there are several types of account weights: zombie accounts, low-weight accounts, accounts that have been downgraded midway, accounts to be recommended, and accounts to be popular. 1. Zombie Account

A zombie account refers to an account that posts videos at a low frequency for a week and the number of video views is less than 100. The algorithm will classify it as a zombie account and it will receive the lowest initial traffic.

2. Low weight number

A low-weight account means that the average number of video views for an account for seven consecutive days is around 100 to 200. In this way, the traffic distributed to it by the algorithm will become less and less!

3. Downgraded account

Some accounts are operating well, but suddenly there will be a decline in data in all aspects. At this time, three situations are likely to have occurred: first, he moved the video; second, he posted an advertising video; third, he brushed the data in the background.

What is moving video? For example, if your video has been published on Youku and iQiyi, and you copy it to Douyin, the algorithm will downgrade this account and think that you are promoting other platforms.

Both hard advertising and data manipulation are easy to understand.

4. Wait for the recommended account

An account updates frequently within a week, and the average number of video views exceeds one thousand, which is the minimum limit. The algorithm will then regard it as an account with a decent weight.

5. Wait for the hot number

Accounts whose average video views are over 10,000 within a week are considered to be of relatively high quality.

2. How to create popular content?

After having a Tik Tok account, how do you make users like you?

The core content is three points: 1. Arouse cognition; 2. Arouse emotions; 3. Summon desire.

Through a lot of practice, we have found four effective methods: support, criticism, advocacy and encouragement.

1. Support

Support means telling your users that they are right and they are always right.

The first step is to find the scenarios in the user’s memory that are denied, doubted, and opposed.

The second step is to rationalize his/her behavior with our words and videos.

For example, if our user is a young person in a rebellious period, we can rationalize his impulse and rebelliousness as a statement of self-pursuit, and rationalize the scenario in which he is doubted and opposed.

For example, young people in the workplace may not go to get off work or leave work on time. We can describe this behavior as a love of life.

This is the thinking logic of "support". If you support your users, they will have a strong resonance with you and then like you.

2. Criticism

"Criticism" means helping our users criticize the unreasonable phenomena in their lives.

The first step is to find out what unreasonable phenomena users often face in their lives.

The second step is to help users expose and criticize this unreasonableness.

For example, queue-jumping is an unreasonable phenomenon in life, and feminism is also an unreasonable social phenomenon. Therefore, we can gain the support of users by criticizing queue-jumping, criticizing feminism, criticizing this kind of gender discrimination, etc.

3. Endorsement

To put it in one sentence, it is telling users: "I'll do it for you." The key is to find the things in the user scenario that they want to do but dare not do or are not convenient to do.

For example, a friend borrows money and doesn't pay it back, a roommate is too noisy, a driver is impolite, a boss is too harsh, etc. We can find these phenomena and use our videos to help users fight against them.

4. Encouragement

In other words, it tells users: "You can do it!"

The first step is to find the factors in the user scenario that prevent them from getting better. For example, procrastination, sloppiness, and lack of persistence.

The second step is to think about how to help them improve their self-confidence and overcome difficulties.

For example, if the users are college graduates who have just graduated and are entering the workplace, and have suffered setbacks and become depressed, we can use videos to show the inspirational process of a young person taking the IELTS or TOEFL exams, so that the users can feel that their lives can become better.

You can use "chicken soup article" as an example to understand this concept.

3. Six models of Douyin conversion

After completing the above four steps, users will already like our account very much. The next step is to convert them and let them act according to the direction we provide.

Now we come to the third part of today: six models of Douyin conversion.

In the first half of this year, especially in the second quarter, Douyin connected to the backend data of e-commerce platforms such as JD.com, Taobao, Tmall, NetEase Kaola, and Vipshop, perfectly realizing a closed-loop payment chain.

During this process, users will not experience link jumps, nor do they need to fill in their Alipay/Taobao/JD.com/Kaola account or delivery address. It is very easy to use.

Therefore, in the first half of this year, Douyin has officially realized the user's "what you see is what you get" payment closed loop. It only takes one minute to arouse consumption desire and only ten seconds to realize consumption impulse.

Through a lot of practice and experience in helping some fast-moving consumer goods brands and 3C brands to transform, we currently have six relatively mature sales models that can achieve a win-win situation of integrating product and effect.

The six modes are: factory short video live broadcast, grass-planting video, giant short video, evaluation short video and celebrity short video.

1. Factory short video

This is a type of content that has grown relatively quickly on Douyin this year. The core logic is direct sales from manufacturers and genuine materials. The core purchasing power is that there are no middlemen to make a profit from the price difference, making users feel that it is authentic and reliable.

2. Video promotion on Douyin live streaming

There is no need to explain more about this. Everyone has watched the live broadcasts of Kuaishou and Taobao, where people buy while watching the live broadcast.

3. Videos that will inspire you to do something interesting

By using some top KOLs’ persuasion on product functions, fans can be motivated to yearn for the product, thus achieving impulse purchases.

This June, we promoted an electric toothbrush from a fast-moving consumer goods brand, priced at 299 yuan. In two days, 100,000 people came into the store and paid about 430,000 yuan. In one month, the payment exceeded 1 million. Now the effect of Douyin's promotion is very good.

4. Drama short videos

It stimulates users' pleasure and pain points through plots, and uses empathy to stimulate users' desire to buy. It is suitable for some low-priced, high-volume, fast-moving consumer goods and clothing products.

5. Review short videos

Recommend products through the perspective of expert reviews/trials, and help users try out the products from the perspective of a real and sincere user, so as to achieve the promotion effect and realize the effect of planting grass.

6. Celebrity short videos

The volume of celebrity short videos grew rapidly in the second quarter of this year. Celebrities like Wong Cho-lam and Chen He have already started to sell products on a large scale, using their authority and endorsement to recommend products to fans, with good results.

4. In terms of content, choose the right topic and avoid making mistakes

Whether we are operating a company’s Douyin account or a personal Douyin account, the first thing we need to think about is the topic selection: What should we shoot this week? What are you shooting next week?

So, where do the topics come from?

1. Three ways to find topics

First, prepare in advance.

For some holidays, we can make preparations in advance. For example, during the Spring Festival, there will definitely be some topics where relatives and friends will be concerned about young people’s marriage, love, salary, etc. By preparing such topics in advance, you can create good playback and interaction volumes.

Second, benchmark against competitors.

Generally, we need to identify 3 to 5 competitors, and then look at how the competitors’ Douyin videos are selected and produced. By imitating angles, reverse thinking, extended thinking and other different methods, we can make a better video based on the competitors.

Third, follow hot spots.

Follow hot topics, including platform hot topics, topics hotly discussed by netizens on the platform, emerging topics, etc. Content creation must remain sensitive to hot topics.

After the video is produced, it is generally recommended that you publish it between 12:00 and 13:00 and between 20:00 and 23:00 in the evening, because these two time periods are the peak periods of Tik Tok activity.

2. Four reasons for traffic decline

Some topics may not be good. If you suddenly find that your traffic has dropped, the main reason is probably that you posted a video that the platform does not recommend.

There are four types of videos that the platform generally does not recommend. First, the video quality is poor and blurry.

Second, moving videos, as I just mentioned.

Third, the values ​​​​do not match the tone of the platform, such as videos involving non-mainstream values ​​such as pornography, gambling, and drugs.

Fourth, the video content involves child abuse, gambling, feudal superstition, etc.

These will result in the video being taken offline or even the account being demoted.

There are also some banned industries on Douyin, such as lottery, adult products, medical beauty institutions, pharmaceuticals, insurance, finance, etc. The official will not accept hard advertising in these industries.

At the same time, in the process of operating corporate brand accounts, we also find that the traffic fluctuates greatly, or declines, or some videos are blocked. The main reason is that they do not comply with content marketing standards and cross the red line of content marketing standards.

For example, WeChat accounts, QR codes, link addresses, dates and some guiding words are not allowed in the video; promotions, product price information, discount information, coupons and other information are not allowed; exaggerated descriptions of goods, services and products are not allowed; brand freeze frames are not allowed at the beginning or end of the video, that is, a brand poster is not allowed where the brand logo is; it is not allowed to show the product and logo you want to promote for a long time in the video; it is not allowed to lure users through raffles, red envelopes and other behaviors.

From its launch in September 2016 to today, Douyin's entire product system and gameplay system have become quite large, and it is difficult to explain it clearly in a short period of time. If we want to make our content, brand, and products popular on Douyin, remember one thing: on Douyin, whether we are doing marketing or creating content, it is 50% certainty of science + 50% possibility of art .

The so-called deterministic science is the algorithm, logic and other structures of Tik Tok. The 50% art of possibility depends on the level to which our content can be achieved.

Author: Zhao Enze

Source: Noteman

<<:  The highest realm of activity operation: product activity & activity productization!

>>:  Top 10 rules for short video operations to become a hit!

Recommend

Inventory of customer acquisition channels for Internet advertising!

Whether it is a service or a physical product, on...

ASO optimization, some things you have to know when doing ASO!

1. The world is in dispute, and CPs discuss ranki...

How much does it cost to be an agent for a decoration mini program in Xuchang?

How much does it cost to be an agent for a decora...

A brief history of underwear advertising

Bras, like sex, are topics that Chinese women are...

PR Event Planning Guide

It is not difficult for PR professionals to carry...

2019 Internet Marketing Promotion Tips!

With the gradual improvement of current Internet ...

3 strategies for native advertising!

Only by using fewer "tricks" can you tr...

Community operation: How to obtain accurate user groups?

To do community operation, you must learn how to ...