Research of YouTube
Introduction
One commonsense we must know that is more ads means more income; more audiences and YouTubers will bring more videos and watch time. One thing someone may ignore that is traditional video works such as a movie or a TV series may only have several ads per one episode leading to high bid per advertisement. However, every YouTube video has at least one advertisement at the start; as the video becomes longer, the ads will be more. If we let the numbers of ads divided by the total videos on YouTube, it may be a tremendous number. In other words, YouTube wins by the volume of advertisements, instead of the bid per advertisement. Therefore, the main revenue of YouTube is ads, which will share with YouTubers inspiring more interesting works.
KPIs
How does YouTube help YouTubers gain more subscribers and watch time? To further explore how YouTube works, I tried to upload a video on YouTube, which means I became a YouTuber. In comparison to another streaming platform named Bilibili, YouTube has a much more convenient experience. You do not have to register for a YouTuber; any YouTube account has the right to upload a video or post. The only thing you need to do is to click the first button on the top of the website – Create a video or post.
After uploading a video, YouTubers could view the detail of the channel when you click your profile and select YouTube Studio. The most important three things in this research are Monetization, Videos, and Channel.
Monetization
Account -> YouTube Studio -> Monetization
According to the YouTube Policy: “To get into the YouTube Partner Program, your channel needs 4,000 public watch hours in the last 12 months, and 1,000 subscribers. Your channel will also get reviewed to make sure it follows” and Intro to Making Money on YouTube
Prerequisite for earning money through YouTube:
- 1,000 subscribers
- 4,000 public watch hours in the last 12 months
- Join the YouTube Partner Program (YPP)
- Be in good standing with YouTube
Revenue from YouTube:
- Ads income from YouTube
- Channel membership
- Chat replay (a kind of donation)
- Ads income from sponsors
“If you like this video, remember to hit like, subscribe, and turn on the notification bell…” This may be the most common sentence that appears at the end of every video on YouTube, Which means YouTube may set a huge weight on the functions of like and the notification. Try to recall how many YouTubers you have subscribed to and how many guys just never appear on your YouTube home page or called the YouTube recommended videos page. I think the reason why they disappear is you do not watch their videos anymore though you still subscribe to them. That is why the notification bell is important for YouTubers – the appearance rate.
Video Analytics
Account -> YouTube Studio -> Videos -> Select one video -> Analytics
Therefore, YouTube should help the YouTubers focus on the subscribers, the likes and dislikes, and the public watch hours. And there are several indicators are useful to judge the quality of the channels or one video:
- Views, Watch Time, Subscribers, Like (vs. dislikes): They are crucial indicators showing the result of your video. If you make a great video, you will gain a huge number of subscribers and views which gives the YouTubers a good sample and lets them keep the style; if it does not work well, YouTubers may try to improve the quality of next videos.
- Audience Retention: It is the best index for YouTubers to troubleshoot the videos. The exit time could help you know when the audiences close the videos. Then the video makers can find out which part of the video is bad. For example, Is the content too boring, or is the opening too long?
- Traffic Source: How do the audiences know this video?
Overview Page
Take the video – “Beijing Housing Price Analysis” I uploaded as one example to explore. After clicking the Analytics, we can see the detail of a video. The overview page shows the Views, Watch Time, and Subscribers at last 24 hours. Of course, click SEE MORE you can change the timeline.
Reach Page
Traffic source is another important index for YouTubers, which can provide information about how do the audiences know this video. For example, the most proportion of the Traffic source types of my video is Direct or unknown since I just shared the link of my video to my group members for testing. YouTubers could set a proper label for their video resulting in more appearance rates of their videos in similar videos.
Channel Analytics
Account -> YouTube Studio -> Analytics
Channel Analytics has a similar interface as Video Analytics. But it shows overall information about your videos. It can help you track your channel situation.
- Total Amount of Video Views: It is a useful index to show which video is more popular in the overall view. YouTubers could analyze the reasons and make their next videos better.
- Number of Subscribers: It could help YouTubers track the number of their followers.
- Daily Active Users (DAU): YouTubers could know their “true” subscribers by analyzing the DAU. For instance, they can create more videos to attract users with high DAU status, instead of the low one.
Predictive Analytics
One more question, how does YouTube distribute the limit resource, whatever attracting more new users or keeping the current users online. YouTube does not have unlimited money. Additionally, the space on everyone’s home page is limited. It is impossible to give every YouTuber the same resource and suggest thousands of videos on your home page. More detailed questions are: Suggesting more related videos from different YouTubers or the same YouTuber to keep the users? Helping the new YouTubers or old but famous YouTubers to increase subscribers? The total watch time, total subscribers, total views, average views, and average retention of each YouTuber may help YouTube make a decision. The decision tree is a great model for preliminary research for making a decision, or furthermore, a machine learning. Since watch time, views, and subscribers are three dependent variables that are interactive. More subscribers will cause more watch time and views; also, more views may attract more subscribers. That is why the decision tree a great model for this research.
The decision tree can help the department reasonably allocate resources. For example, organizing the YouTubers’ party to attract the audience. Also, it is not necessary to show a private channel that is only shared with one’s family or friends on everyone’s home page. Instead, YouTube should show the videos you may interest in on the home page. A decision tree can also help YouTube build a unique model showing the potential interesting videos for every user.
Additionally, everyone could have a customized decision tree to predict the next series of videos he/she wants to watch. A machine learning model including label of YouTuber and video, duration of a video, language of the video, likes and dislikes, views, even the gender of YouTuber as the parameters of the recommendation system are helpful to keep the users on YouTube.
Remember, the core of profit is watch time and subscribers, which means YouTube needs to provide the most convenient service for YouTubers in order to maximize the watch time and subscribers. And make sure everyone can watch the videos he/she preferred.