內容標簽匹配

Content label matching

一篇筆記發(fā)布以后，會(huì )被系統拆分成N個(gè)標簽，然后去推送給--近有同樣內容喜好的用戶(hù)

After a note is published, it will be split into N tags by the system and pushed to users who have similar content preferences recently

內容側: 一篇筆記不管是圖片還是文案，會(huì )被系統按照標簽去區別和理解

Content side: A note, whether it is an image or a copy, will be distinguished and understood by the system according to its labels

用戶(hù)側: 用戶(hù)--近的閱讀喜歡也是會(huì )被系統按照標簽進(jìn)行分類(lèi)和匹配。

User side: Users' recent reading preferences will also be classified and matched by the system based on tags.

本質(zhì)上，是一種內容標簽對用戶(hù)標簽的匹配

Essentially, it is a match between content tags and user tags

筆記推送給用戶(hù)后，根據用戶(hù)的一系列交互行為，給筆記打上內部分數，以決定是否要繼續推給其他更多用戶(hù)。這個(gè)評分體系在小紅書(shū)內部稱(chēng)為CES

After the note is pushed to the user, an internal score is given to the note based on the user's series of interactive behaviors to determine whether to continue pushing to more other users. This rating system is called CES internally in Xiaohongshu

小紅書(shū)流量分發(fā)機制及算法

Xiaohongshu Traffic Distribution Mechanism and Algorithm

CES計算方法

CES calculation method

點(diǎn)贊1分 + 收藏1分 + 轉發(fā)4分 + 評論4分 + 注8分

1 point for likes, 1 point for favorites, 4 points for reposts, 4 points for comments, and 8 points for annotations

時(shí)至今日，算法的判斷指標相較以前更加豐富和多元化，綜合在小紅書(shū)內的投放實(shí)踐經(jīng)驗，如今平臺更側重以下幾項:

Today, the judgment indicators of algorithms are more diverse and diverse compared to before. Based on practical experience in advertising in Xiaohongshu, the platform now focuses more on the following:

點(diǎn)擊率、互動(dòng)率、完播率、互動(dòng)值

Click through rate, interaction rate, completion rate, interaction value

小紅書(shū)流量層級

Xiaohongshu Traffic Hierarchy

1級流量層級一筆記瀏覽量約0一200

Level 1 traffic level 1 note views approximately 0-200

只要筆記不涉及違規情況，不管你的筆記內容質(zhì)量如何，都能獲得200左右的閱讀

As long as the notes do not involve any violations, regardless of the quality of your notes, you can get around 200 readings

2級流量層級一筆記瀏覽量約200-500

Level 2 traffic level 1 note views approximately 200-500

屬于正常流量，大多數的賬號都能達到2級流量層級，比較容易，說(shuō)明賬號無(wú)違規情況。

Belonging to normal traffic, most accounts can reach level 2 traffic level, which is relatively easy, indicating that the account has no violations.

3級流量層級一筆記瀏覽量約500-2000

Level 3 traffic level 1 note views approximately 500-2000

說(shuō)明賬號狀態(tài)正常，筆記內容質(zhì)量還可以，但是互動(dòng)率/點(diǎn)擊率相比大盤(pán)平均數據略低，只能進(jìn)入到第3流量層級，到達這一層級已經(jīng)超越80%的小紅書(shū)賬號，內容質(zhì)量上請繼續努力。

The account status is normal, and the quality of note content is decent. However, the interaction rate/click through rate is slightly lower than the average data of the large market, and can only enter the third traffic level. Xiaohongshu accounts that have exceeded 80% of this level have reached this level. Please continue to strive for content quality.

4級流量層級一筆記瀏覽量約2000一2w

Level 4 traffic level 1 note views approximately 2000-2w

說(shuō)明內容已經(jīng)獲得了不錯數據，具有小爆款的潛質(zhì)數據在持續爬升。

The content has obtained good data, and the potential data for small blockbusters continues to climb.

5級流量層級一筆記瀏覽量約2w-10w

Level 5 traffic level 1 note views approximately 2w-10w

5級流量池是自然流量的--后一關(guān)，依然是用戶(hù)互動(dòng)數據決定是否進(jìn)入下一流量層級

Level 5 traffic pool is the last level of natural traffic, and it is still the user interaction data that determines whether to enter the next traffic level

6級流量層級一筆記瀏覽量約10W-100w

Level 6 traffic level 1 note views approximately 10W-100w

已經(jīng)進(jìn)入熱門(mén)筆記門(mén)檻，達到這個(gè)階段的筆記已經(jīng)具備一定的稀缺性，用戶(hù)互動(dòng)數據比較好，也是從這一層級開(kāi)始測試更多用戶(hù)后端數據

Having entered the threshold of popular notes, notes at this stage are already somewhat scarce, and user interaction data is relatively good. Starting from this level, we will test more user backend data

7級流量層級一筆記瀏覽量約10W-100w

Level 7 traffic level 1 note views approximately 10W-100w

筆記進(jìn)入這一層級就說(shuō)明已經(jīng)成為爆文，不管是用戶(hù)互動(dòng)數據還是用戶(hù)后端數據，都有不錯的表現也是從這一層級開(kāi)始就有人工層面的干預了，比如價(jià)值觀(guān)是否符合社區?加權推薦后是否存在輿情風(fēng)險?是否存在版權風(fēng)險?用戶(hù)觀(guān)感是否優(yōu)質(zhì)?

Entering this level of note taking indicates that it has become a hot topic. Both user interaction data and user backend data have performed well, and there has been manual intervention since this level, such as whether values align with the community? Is there public opinion risk after weighted recommendation? Is there a copyright risk? Is the user's appearance high-quality?

8級流量層級一筆記瀏覽量約500w+

Level 8 traffic level 1 note views approximately 500w+

難得一見(jiàn)的大流量爆文，用戶(hù)互動(dòng)數據和用戶(hù)后端數據表現都非常出色，遠高于大盤(pán)平均數據，且獲得了平臺的加權推薦，算是平臺的香餑餑，這個(gè)階段的內容一般是熱點(diǎn)新聞/時(shí)事之類(lèi)的居多，頭部博主想要達到這個(gè)數據的流量，也非常困難

A rare high traffic explosive article, with excellent performance in user interaction data and user backend data, far higher than the average market data, and receiving weighted recommendations from the platform, it is considered a hot topic of the platform. At this stage, the content is usually hot news/current affairs, and it is also very difficult for top bloggers to reach the traffic of this data

以上就是山東短視頻運營(yíng)小編為您講解的問(wèn)題內容了，您對此有怎樣的疑惑或者需求就來(lái)我們網(wǎng)站http://www.sg8020.com做咨詢(xún)吧！

The above is the question content explained by the Shandong Short Video Operations editor for you. If you have any doubts or needs about this, please come to our website http://www.sg8020.com Do consulting!