Analyze the algorithm mechanism behind Tik Tok!

Understanding the platform rules is the key to playing TikTok. This article will analyze the algorithm mechanism behind TikTok one by one.

01: Douyin review mechanism

When a user uploads a video, the video will be initially screened through machine review. Mainly review the video title, screen title and keywords. If the machine determines that there is a suspected violation, the video will enter the manual review stage. Once a violation is manually confirmed, the video will be deleted or restricted. If the video does not violate any rules, deduplication will begin, and videos with high duplication rates will be included in low-traffic recommendations.

▲Image source: Douyin Community Self-Discipline Convention

▲Image source: Douyin Community Self-Discipline Convention

02

Account homepage construction

Tik Tok will refer to the weight of the account when allocating traffic, so it is necessary to improve the official certification, account name, introduction, avatar, and background.

① TikTok official certification

Douyin's official certification is the platform's confirmation of the authenticity of the user's identity. Identity authentication is helpful in gaining more recommended positions for the video.

② Complete personal information

Taking "Hua Feifei's Beauty Treasures" as an example, the account has a distinct personal style. The account name is concise and consistent with the overall tone of the account.

▲Image source: Douyin celebrity "Hua Feifei's beauty treasures"

03

Tik Tok’s recommendation mechanism

Douyin gives every video creator the right to compete with big accounts. It is well known that Douyin's traffic distribution is decentralized. Even if an account with no fans publishes a video, the Douyin platform will allocate a small amount of traffic to you. As long as the content is high-quality enough, it can also become a hit.

① Stepped flow pool

Douyin’s popular recommendation mechanism is divided into three stages: primary recommendation, multi-level recommendation and popular recommendation.

② Hotspot weighting algorithm

Douyin will give recommended positions to current hot news and hot topics to increase exposure.

③ Account weight algorithm

For a new account, the quality of the first five videos it publishes determines the weight of the account.

Works published with less than 100 views for a week are considered zombie accounts; accounts with 100-200 views are considered low-weight accounts, and those that continue for half a month to a month are considered zombie accounts; accounts with 1,000-3,000 views are accounts waiting to be recommended, and at this time, you should pay attention to maintaining the likes and comments data to strive for recommendation; accounts with more than 10,000 views are accounts waiting to become popular.

④ Continuous maintenance of high-quality videos

Users who frequently use Douyin may find that some videos have been released for some time, but they still appear in popular recommendations. Many people also suggest that "the video will become popular if it is released a second time". What is the principle behind this?

First, Tik Tok will continue to discover high-quality content and bring it secondary exposure. Secondly, it is possible that the first wave was not pushed to the right audience. As the content published by the account becomes more vertical, the system distributes users more accurately.

Therefore, even if a video is not recommended, it still has the potential to become popular as long as it continues to receive likes and comments.

04

Identification of user portraits

The premise for Tik Tok to recommend personalized videos to users is to judge the user portrait and understand the user's interest preferences. The judgment is mainly based on the following five aspects:

05

Account content minefield

When creating videos, try to avoid the following types of content:

① Picture carousel

② Soul-stirring videos

③ No-mouth unboxing video

④ Street interviews, selling unrelated products

⑤ Recommend good products

⑥ Small theater performance

For short video practitioners, the most important thing is to understand the platform rules, and on this basis use content to fit the algorithm to obtain maximum exposure.

Author: Kas Data

Source: Kas Data