Today’s topic is still the video account. I don’t know email list how many people use the video account every day. However, I am a heavy user of the video account and a fan of the video account. From the video account, I was the first batch of users. I also wrote an article before the video account announced that the number of users reached 200 million seat. On the WeChat night in 2021, Zhang Xiaolong made it public for the first time: in June 2020, the proportion of video accounts recommended by machines is still very low. . Zhang Xiaolong once predicted that in the video playback, the ratio of attention, friends likes, and machine recommendation should be 1:2:10.
Because I am a heavy user, this distribution is a bit email list exaggerated. Through chatting with friends around me, I found that the video account did not rob them of much time from Douyin, and the general feedback was that the content of the video account was not good enough. However, they also admit that they occasionally take the initiative to click in to see the content of the video account, and more are actually the content of the video account shared by friends in the circle of friends or sent to group messages. I divide the video account user groups into three categories, namely: For these three types of users, the proportion of the number of videos watched on the Follow, Friends, and Recommended tabs generally shows the following trends: Remarks:
The content of the video number received through the circle of friends or group messages is counted as the amount of friend recommendation (not necessarily in the strict sense, if the friend does not like it, but just shares it with you, this type of video playback should be excluded from the calculation scope), and then continue to scroll down to view of videos count as recommended videos. The final distribution of the calculated data is about 1:4:40. That is to say, the amount of recommended videos viewed is 10 times that of friends recommended, and 40 times that of videos posted by followers. Due to the limited data set and the fact that reasoning is mostly based on individual subjective inferences, it is itself a predictive value.