Tuesday, October 23, 2018
Cases - Audit YouTube channel using YouTube Analytics Eugene Zaremba 7-9 ...
Cases - Audit YouTube channel using YouTube Analytics Eugene Zaremba 7-9 ...
YouTube channel audit | using YouTube Analytics
When you take someone else's channel into promotion or you want to change something on your own, then first of all you need to conduct a competent audit. In this article, I will share some case studies from my own practice that are completely based on YouTube Analytics . I hope that this will help to find some technical flaws and raise the views in the least cheap way. If you are interested in promotion, it is better to read the article about promotion on YouTube .
- Studying spikes on graphs of main metrics
The first step is to conduct a basic analysis. Usually for this, I study the charts of the main metrics (views, subscribers, likes) and I look for there if there are any sharp rises and falls. Studying spikes makes it clear:
- Was cheat used on the channel
- What videos bring views (regular or some specific)
- Is there a correlation between different graphs
Let's look at an example. Here are the graphics of two channels that release videos with the same frequency.
Chart 1.
Chart 1.
Chart 2
What conclusion can be made? The second graph shows a sharp increase in views on certain days. Most likely, the channel has a permanent audience, which follows the release of commercials. When the video comes out, they watch it.
On the first chart there are no sharp rises and falls. One of the initial hypotheses is that the channel has several videos that generate organic traffic, while the audience does not follow the rest of the content. Most likely, the channel has several videos with a huge number of views, and the rest of the video are gaining a couple of thousand. Those. surely the channel has a large variance in views from the video.
Example 2
Here we see a very sharp rise in views on a particular day. If these were real people, then the schedule of the fall would be more gentle. Advertising and crops also always have a smoother drop. Right there, views the day before and the day after have almost the same value. My hypothesis was - cheat or deception service crops. But of course we must check.
2. We study playback devices
Studying the devices from which people watch video content helps me improve image pictures and increase the amount of recommended traffic in the future.
Consider an example. Below are the devices of two different channels.
Channel 1.
Consider an example. Below are the devices of two different channels.
Channel 1.
Channel 2.
On the first channel, the number of views from mobile phones is 46%, on the second channel - 24%.
If the number of views from the phones is quite large (for me it is now 35% +), then rather large text should be made on miniatures. If there is a lot of desktop traffic, then the emphasis should be placed on the image (although the text should also be)
If the number of views from the phones is quite large (for me it is now 35% +), then rather large text should be made on miniatures. If there is a lot of desktop traffic, then the emphasis should be placed on the image (although the text should also be)
3. Check the channel for cheat.
If you undertake a channel on a commercial basis, then you should definitely check it for cheating. Twisted channel can move much worse. He may not fall into the recommendations, as well as with great difficulty to move to the top.
Cheat, I advise you to look for the following algorithm:
- See if there are jumps in the graphs of metrics up and a sharp drop on the initial value (as in the example of the first item).
- We study external sources of traffic and try to find “left” sites.
- We study the sources for deviations (both up and down) from the retention of the channel or roller.
To make it easier, let's take an example. We found a graph with a metric which has a sharp jump and the same sharp fall.
Next, go to the analytics of this particular day and analyze the traffic.
On that day, when I found the jump, there was no great hold on external sources.
On that day, when I found the jump, there was no great hold on external sources.
After examining external sources, we see that the traffic comes with the Safari app and the Apple Messages app.
Having studied the information on the Internet, it becomes clear that the cheat was used.
This should be immediately discussed with the client. There is a likelihood that the channel is already under pessimization, and that this could happen in the future.
I had such a case: they started growing the channel to the client, everything went well, but one day YouTube wrote off a certain number of subscribers who were fakes. Before me, the channel was wound up, and, in fact, the final number of subscribers after 2 months of work turned out to be less than it was at the beginning. In my case, there was an adequate client, but otherwise everything could turn out to be worse. Therefore, always be sure to check the channel for cheating!
This should be immediately discussed with the client. There is a likelihood that the channel is already under pessimization, and that this could happen in the future.
I had such a case: they started growing the channel to the client, everything went well, but one day YouTube wrote off a certain number of subscribers who were fakes. Before me, the channel was wound up, and, in fact, the final number of subscribers after 2 months of work turned out to be less than it was at the beginning. In my case, there was an adequate client, but otherwise everything could turn out to be worse. Therefore, always be sure to check the channel for cheating!
4. Check the number of subscribers who watch your channel.
If more than 10% of views are from subscribers, then you can increase the views after a while by optimizing the time and day of the videos. According to my calculations, every 10% of views from subscribers can increase the total views by 4-5%. For example, if you have not optimized the time and day of the commercials, and subscribers account for 30% of all views, then you can raise the total views to 15% more.
You can find out the number of views from subscribers in the “Views” tab and then “Status”
You can find out the number of views from subscribers in the “Views” tab and then “Status”
In this example, it makes sense to spend time optimizing the output of commercials.
In this example, it will not do anything.
I will write a separate material in the near future about optimizing the release time, but in general it is based on two components:
- Report “Real-time data”
- Your own observations
Take a test: release the same type of video on different days of the week and track user behavior in the analytics in the “Real-time data” report. The main task is to find the day and time when users are most active and the video is gaining more views.
5. We are looking for a correlation between different types of traffic.
One of the most difficult points, but at the same time the most effective. Without practice, it is rather difficult to find something here, but when you “put your hand on”, such things will strike the eye that, for example, the channel has a wrong correlation between some indicators. After analyzing the lower value, you can find some problems that will help increase traffic.
I once worked with a channel that has almost all traffic from a YouTube search.
But there was almost no traffic from google
The picture when organic from YouTube is more than organic from Google is quite normal. But here the ratio was 141: 1. This is a lot and take a look at organics from Google for sure.
In my case, the problem was the following: translation of descriptions and titles into other languages was used on the channel. The English version of the vidos got into the Russian issue, but there was no Russian-speaking version at all.
I checked different sites for this problem, but at that time (and even now too) there was no solution. Seeing that there is no English-speaking Google traffic either, I simply deleted the English version.
Traffic in a few days began to grow and grows until now. Indicators of course miserable, but for commercial subjects not so bad.
If this type of material is interesting, then a little later I can write a material on a more complex audit of YouTube channels. Ask questions in the comments, it's always interesting to discuss some