Why is the number of views always low? Five data indicators that short video operators need to pay attention to

Data analysis is a very necessary part in short video operations . Today, Xiaoyu will introduce to you some basic but very important data concepts, which I believe will be helpful to you.

1. What conclusions can we draw from the most basic playback volume?

First of all, there are several basic key data: number of views, number of comments, number of likes, number of reposts, and number of collections.

The number of views is a basic quantity among the basic quantities, and it is also one of the important criteria for judging the quality of a video. I once wrote a post about some tips for writing headlines on Toutiao . The whole article sorts out the characteristics of the top 100 video titles with the highest number of views of Isese Shen’s skills.

For example, how long should a qualified title be? The number of characters in the title that can be fully displayed on mobile phones and web pages is 26. Among the top five videos of Isshiki's skills with the most cumulative views, four have titles with around 20 words, and one has 30 words. Then we averaged the word count of the top 50 and top 100 titles, which are both 20 . At this point we proceed to the second step of statistics. Among the top 50, there are 7 titles with more than 25 words, and among the 51st to 100th titles, there are 6 titles with more than 25 words. This proves that a title with more than 26 characters does not necessarily mean a low number of views, but we should try to shorten the video title to 26 characters.

The above paragraph is actually the simplest process of helping operations colleagues make decisions by analyzing some existing data . Here we are only analyzing the data through the number of views, and using only this one data we can draw some regular conclusions.

We have four other quantities, and depending on the channel , the key data that should be analyzed is also different . For example, for large channels like Youku and iQiyi , the most important thing is the number of views . It can also be seen from the page that there are almost no entrances to display the number of comments, likes and so on. In fact, we can also see from our statistical tools that there is no such thing as forwarding and collection on these classic video playback channels. Apart from the number of views and very few comments, the number of likes and dislikes can also be ignored. Then when analyzing the data of these channels, it is enough to just look at the number of views.

However, on channels like Meipai and Miaopai , we may need to look at the number of likes in addition to the number of views, and on Weibo we need to look at the number of reposts and comments . In short, it is necessary to understand the characteristics of the channel and find key data for analysis, otherwise the conclusions drawn will be useless.

2. Several ratios obtained by processing key data

In addition to the amount of data, we can actually process the data before analyzing it. Here I would like to add a few more concepts, that is, based on the number of views, we need to do some processing on the other four data quantities, which are the four ratios that Meimiao Academy often emphasizes: comment rate, like rate, forwarding rate, and collection rate . To put it simply, it is the ratio obtained by dividing the number of comments, likes, reposts and favorites by the number of views.

What is the meaning of ratio? Why do we have to remove it? The number of views of videos posted by many video accounts can differ by dozens of times. Some videos may have millions of views in a flash, while others may only have 10,000 or 20,000. The amount of data may vary , but the ratio obtained by division is basically stable . We use this division to find the ratio, so that videos with many times different playback volumes can be comparable. Therefore, in addition to the number of views and the evaluation of the video based on the viewers' aesthetic taste, these four ratios are also very important data indicators.

Next, let’s take Toutiao as an example and talk about how to analyze the ratio. The data provided by Toutiao is the most comprehensive channel. As we mentioned above, platforms like Youku and iQiyi do not have data on forwarding and collections, but Toutiao has all of them. The more types of data there are, the easier it is to draw conclusions with higher reliability . Here I picked out two IPs as cases for analysis.

Example Analysis

These are two IPs that focus on food making tutorials, one is Summer Chef SK and the other is Poverty Life Cooking.

1. The significance of collection rate

Let’s take a look at the data for the ten videos they posted on Toutiao.

The first picture is from Xiachu SK. It can be seen that their video playback volume is almost stable at around 30,000, with a higher number of 50,000 and a lower number of 10,000 or several thousand . Let’s ignore the videos with only a few hundred views. The grey words under the last column of collections are the collection rate, which is generally high. The lowest is just over 3%, and the highest is over 10% or even up to 14%.

The second picture is from Poverty Life Cuisine. This picture was taken a little early, but we only use it for data analysis, so it doesn’t have much impact. The number of views of "Poverty Life Dishes" can reach as high as 190,000, and the normal number is around 20,000. But you can see that the video with the highest number of views also has a collection rate of 6% in the last column. This is where the significance of finding this ratio comes into play. Even though the number of views varies greatly, the ratios are comparable.

So, why did I put two food-related IPs here and compare their collection rates? If you have watched the videos of these two IPs, you will know that they are both food tutorial videos. The formal editing techniques are slightly different, but there are differences in the data.

Xia Chu prefers the most popular food tutorial video format, where one person makes the video step by step, shoots the production process very beautifully, edits it very beautifully, and adds cute little pets to complete a video. One advantage of such videos is that the production process is very clear and easy to understand. It is nothing more than changing the previous graphic tutorial into a video tutorial. Because it is a clear and easy-to-understand tutorial, people will save it if they find it useful after reading it once, and read it again when needed. Therefore, the collection rate is high. Here we can draw a criterion for judgment : whether this video is a good tutorial depends on whether it has a high collection rate.

So, why do some videos about cooking recipes for poor life have a low collection rate but still have a decent number of views? Students can search for the video on Toutiao and watch it. It is a creative video that turns the story of Rou Cage into "Gangs of New York". First of all, this number of views is a very good performance for a new account; secondly, it is not a tutorial video, so the collection rate is definitely low; thirdly, this IP has an interesting data phenomenon, that is, the forwarding rate is quite high. Students can compare it with Xia Chu above.

Although the collection rate of Poverty Life Recipes is not as high as 14%, they have a forwarding rate of 4.6% . They posted a total of 7 videos. Excluding the first two which had very low playback rates, 3 of the remaining 5 videos had very high forwarding rates . You may want to watch these videos and feel what part of them makes you want to share them.

2. The significance of forwarding rate

A forwarding rate is mentioned here. In fact, the forwarding rate represents a sharing behavior. This data is not that meaningful when viewed on Toutiao. It is more meaningful on channels like Weibo that focus on sharing and interaction. However, Weibo is a special channel, as the video is also posted as a Weibo post. At this time, there is no need to look at the ratio, because everyone has a very unified concept of the number of reposts. For example, over ten thousand or one thousand reposts are considered a very high number of reposts, so this channel looks at the quantity rather than the ratio.

At the same time, Weibo is a channel that emphasizes fan operations . To put it bluntly, the more people follow you, the more people will share your Weibo posts, which will bring you more new fans and the cycle will continue. So how do we analyze and optimize a video in a Weibo post? In addition to actually looking at the content of the comments and feedback, two key data are how many followers a new video gains after it is released, and how many times the video is forwarded.

It is impossible to accurately tell how many fans a video can bring from external data, but we can estimate it. For example, this video was released for three days, and on the first day it gained 10,000 followers. On the second day, nothing new was released, but it still gained 1,000 followers. On the third day, a new video was released, and it gained 5,000 followers. Then we can estimate that the first video gained more than 11,000 followers. At the same time, videos with a high number of reposts are almost always those with a large number of followers. If there is no condition to observe the daily increase in followers, then add the number of reposts as a criterion. This is a data analysis and operation method that is basically only applicable to Weibo.

In fact, there is another data on Weibo called the number of likes. Ever since Weibo introduced the like function, many people choose to click the like button to indicate that they have seen it, and are too lazy to share or forward it. A high number of likes can only determine how many people have seen the video to a certain extent. It can only be said that this channel is not suitable for analyzing this data. So in which channel does the like rate make sense?

3. The significance of like rate

I guess many people can guess that Meipai and Miaopai attach great importance to the like rate on these two channels. Meipai’s like rate is more suitable for comparing your own videos with each other, rather than horizontal comparison. I have observed a lot of accounts, and generally speaking, the like rate of this channel is relatively stable. The average for large sizes is around 3-4 percent, while for small sizes it may be as low as 1 percent. However, the quality of your own videos fluctuates, so this ratio data can be used as one of the criteria for judging the quality of each video on Meipai .

at last

Xiaoyu has only briefly introduced a few data to you above. There are still many points to analyze in data analysis, and each type of data can be used in more than one direction. You can look for more patterns like the collection rate. So, let me briefly summarize it for you:

• First, the basic data includes the number of views, comments, likes, reposts and collections . Based on these data, in order to make videos with very different numbers of views comparable, we divided these numbers by the number of views to get the ratio.
• Secondly, based on the comparison of playback volume, we can draw some preliminary and simple conclusions through statistics . For example, questions like title length, or even which topic direction is better, can all be selected based on the playback volume.
• Thirdly, the ratio can be used to compare video data between accounts , and the conclusions drawn from this comparison can mostly be used to optimize video content. By finding the data patterns of excellent videos and comparing them with the data of your own videos, you can know which aspects can be improved and optimized.

Later, we introduced an analysis point of collection rate through the data of two food IPs. We also mentioned that Weibo, as a channel, mainly uses two data points: daily increase in fans and forwarding volume. Meipai can analyze and compare its own videos through the like rate and optimize them.

In the operation of short videos, data analysis is very important. We need to observe the phenomena behind the data, which will help us adjust the video content and provide guidance for optimization.