揭秘抖音算法邏輯背后的密碼，記得收藏！

要在抖音平臺(tái)上嶄露頭角，深入理解其內(nèi)在規(guī)則至關(guān)重要。這里我們將詳述抖音獨(dú)特的推薦機(jī)制，它與今日頭條的去中心化推薦算法原理有著異曲同工之妙，宛如一個(gè)精密的漏斗篩選過程。

In order to emerge on the Tiktok platform, it is very important to understand its internal rules. Here we will detail the unique recommendation mechanism of Tiktok, which is similar to the principle of today's headlines' decentralized recommendation algorithm, just like a precise funnel screening process.

1、冷啟動(dòng)階段：設(shè)想每日有百萬級(jí)別的創(chuàng)作者在抖音上傳短視頻作品，抖音會(huì)運(yùn)用一套公平機(jī)制，為每一條通過審核發(fā)布的短視頻分配一個(gè)初始平均曝光流量池，例如，每個(gè)短視頻初步可獲得大約1000次的曝光機(jī)會(huì)。




1. Cold start stage: It is envisaged that millions of creators upload short video works in Tiktok every day. Tiktok will use a fair mechanism to allocate an initial average exposure flow pool for each short video that has passed the review and release. For example, each short video can initially obtain about 1000 exposure opportunities.

2、下一輪流量池篩選：在這一階段，抖音系統(tǒng)將對(duì)這個(gè)百萬級(jí)短視頻的初次1000次曝光進(jìn)行深度數(shù)據(jù)挖掘和分析，涵蓋點(diǎn)贊數(shù)、關(guān)注數(shù)、評(píng)論互動(dòng)以及轉(zhuǎn)發(fā)次數(shù)等多個(gè)核心維度。從中甄選出各項(xiàng)指標(biāo)表現(xiàn)超過平均水平10%的視頻，再賦予每條視頻額外的10萬次曝光量。緊接著，系統(tǒng)繼續(xù)監(jiān)測這些視頻在新一輪曝光中的點(diǎn)贊、關(guān)注、轉(zhuǎn)發(fā)及評(píng)論等數(shù)據(jù)的增長情況，以此作為評(píng)判標(biāo)準(zhǔn)，推送至更大規(guī)模的流量池中進(jìn)行展示。

2. The next round of traffic pool screening: In this stage, the Tiktok system will conduct in-depth data mining and analysis on the first 1000 exposures of this million level short video, covering many core dimensions such as likes, concerns, comment interactions, and forwarding times. Select videos with performance indicators exceeding the average level by 10%, and assign an additional 100000 exposures to each video. Subsequently, the system continues to monitor the growth of data such as likes, followers, reposts, and comments for these videos in the new round of exposure, using this as a criterion for evaluation and pushing them to larger traffic pools for display.

3、精品推薦池晉級(jí)：經(jīng)過多輪嚴(yán)苛的數(shù)據(jù)驗(yàn)證與篩選，只有那些點(diǎn)贊率、播放完成率、評(píng)論互動(dòng)率等關(guān)鍵指標(biāo)均表現(xiàn)出色的短視頻，才有望進(jìn)入抖音的精品推薦池。用戶在打開抖音時(shí)所看到的那些動(dòng)輒幾十上百萬點(diǎn)贊量的爆款視頻，正是由此誕生。

3. Promotion of boutique recommendation pool: After several rounds of rigorous data verification and screening, only those short videos with excellent ratings, playback completion rate, review interaction rate and other key indicators are expected to enter the boutique recommendation pool of Tiktok. It is from this that the popular videos that users see when they open Tiktok are often tens of millions of likes.

因此，通過科學(xué)提升點(diǎn)贊量、關(guān)注量、評(píng)論量以及轉(zhuǎn)發(fā)率等關(guān)鍵指標(biāo)，創(chuàng)作者能夠贏得抖音官方更為精準(zhǔn)的大流量推薦，從而實(shí)現(xiàn)自身作品更大的曝光度和影響力。

Therefore, by scientifically improving key indicators such as likes, attention, comments and forwarding rate, creators can win more accurate mass traffic recommendations from Tiktok officials, thus achieving greater exposure and influence of their works.