I think I figured out YouTube's secret

[FamilyToyReview]

I am not there yet. My channel only has 1,000 subs in the kid toys category. I've wondered why other channels have been more successful even though we have similar content.

Youtube tells you to do multiple things like have good quality video, improve retention rate, total minutes, good titles, engage with viewers, etc etc. And I've been doing most of the stuff but have neglected some. And that was my problem, I think you need to follow an "all of the above" strategy because all of the suggested guidelines are important for specific reasons. I realized I've made a very critical error in neglecting one area.

I bet YouTube uses click to rates to help sort their "suggested video" section. Even if I have a great video with highly relevant content and very good retention rate, it is just wasting real estate space in the #1 or 2 spot on the "suggested video" section if the click to rate is very poor. Why would Youtube even keep my video in that spot if the thumbnail is very poor and barely anyone is clicking on it. They would be better off displaying another video with a higher click to rate, which would help extend the overall session view time.

I haven't found any source directly telling us that impressions and click to rates are used as part of YouTube's algorithm. But those are basic standard metrics for online advertising or displaying products. So of course click to rates must have some value. They don't tell us directly, maybe to prevent abuse and excessive click baiting. But they do tell us indirectly by recommending good titles, descriptions, and thumbnails which are basically key to improving your video's click to rate.

I could be wrong. YouTube doesn't provide us impressions information in our reports. But they must be using that data.

 

[dreambridgelists]

Yea, it just seems like getting clicks and viewers can be completely random at times. Even if your content is perfect and amazing with the best hook it just seems like somethings stick and some just fall short :/

 

[xingcat]

So now it's time to test your hypothesis, make some changes, and report back on the results!

 

[kevsup]

It may be possible that clicks have nothing to do with the algorithm- rather, the more people click the more watch time you'll get, which of course factors into the algorithm. This would make sense in that if you get a lot of clicks but no watch time, you might have a click bait title and thumbnail. Whereas if you get fewer clicks but people watch the video beyond the first few seconds, you might get more watch time than a click baiter. This may be what YouTube is looking for

 

[Shendijiro]

Moved to Youtube Chat, Gossip and Help

 

[babyteeth4]

Interesting question, I have no idea what the answer is, but it would make sense that the algorithm would favor more clickable stuff. At any rate, it makes sense to always be making the best thumbnails & titles that you can, without being misleading.

 

[FamilyToyReview]

It may be possible that clicks have nothing to do with the algorithm- rather, the more people click the more watch time you'll get, which of course factors into the algorithm. This would make sense in that if you get a lot of clicks but no watch time, you might have a click bait title and thumbnail. Whereas if you get fewer clicks but people watch the video beyond the first few seconds, you might get more watch time than a click baiter. This may be what YouTube is looking for

What you say is true about producing quality content. I think total minutes watched is not only directly related to retention rate but its more important. But I have some videos with great retention rate for my category 60 to 80% that get no time in any suggested video section. One of the largest youtubers (3 mil+ subs) in my category once mentioned their avg retention rate was less than 30%.

literally clicks/uniqueimpressions is a basic standard metric for almost anything else online. It must have some impact in the sort. And that would partly explain why click bait Facts Verse is only 6 months old but has half million subscribers. Their retention rate must be pretty bad.

theyre all important. Retention rate. Minutes watched. Youtube doesnt directly tell you they look at clicks/impressions but it must be a factor.

lets look at extreme examples. Lets say the algorithm for suggested video ranking is choosing video in suggested section between video A and B. Video A has great content with 70% retention rate but title and thumbnail are not appealing so the video is clicked 1 time out of 1000 impressions. Video B has not quite as good content with 10% retention rate but is clicked on 1 out of every 100 impressions. If both videos are 10 minutes long then video A has accrued a total 7 minutes of total watch time and 1 view after youtube has offered it 1000 times to viewers. Meanwhile video B has accrued 10 minutes and 10 views from the same 1000 impressions.

the algorithm has to choose a video for the top spot. Video A must have better quality content than B. But asssuming all other factors are the same, YouTube should choose video B for higher rank at least because they would earn more ad revenue from it.

Definitely all factors are important. Using similar extemes, you could produce an example where a video with high quality content produces more view time and views than a video with low quality content. So high quality content (better retention rate) is still important. But i think click to rates must play some factor in the analysis. It must have some value. And Better titles and thumbnails would have the most direct impact on click to rates.[DOUBLEPOST=1457206230,1457204383][/DOUBLEPOST]

Interesting question, I have no idea what the answer is, but it would make sense that the algorithm would favor more clickable stuff. At any rate, it makes sense to always be making the best thumbnails & titles that you can, without being misleading.

Yup. Definitely better to have appealing titles thumbnails that are relevant instead of click bait. Cheating is a short term gain. Youtube will probably target them during their updates.

By the way ive noticed a lot of cheaters yesterday. People who completely steal videos from other top youtubers. Even going as far as stealing their name and profile pic for the new channel. If you search for "giant gummy bottle" and then sort by "upload date" theres so many stolen videos uploaded. I dont know how theyre even able to monetize.

 

[Flyination]

I also have over 1000 subscribers. I have a video that has over 200 000 thousand views as well.

What I have noticed is that shorter videos (videos under 5 minutes) tend to have a harder time getting to higher view counts through 'suggested videos' and 'recommended'

 

[Sabrina]

But I have some videos with great retention rate for my category 60 to 80% that get no time in any suggested video section. One of the largest youtubers (3 mil+ subs) in my category once mentioned their avg retention rate was less than 30%.

I just posted this in another thread because I read it today on YouTube's help page about watch time:

"The algorithm for suggesting videos includes prioritizing videos that lead to a longer overall viewing session over those that receive more clicks."

That means videos with low retention can still rank high if they lead to a longer viewing session overall. I wish we could see session watch time in Analytics, it would save us a lot of speculation.

 

[FamilyToyReview]

I also have over 1000 subscribers. I have a video that has over 200 000 thousand views as well.

What I have noticed is that shorter videos (videos under 5 minutes) tend to have a harder time getting to higher view counts through 'suggested videos' and 'recommended'

I have a 3 and half minute that has ranking in suggested video from a popular video. But yeah I have also noticed generally short videos havent helped me much and dont add as much to total minutes watched. So i think total minutes watched > retention rate and longer videos have more impact.[DOUBLEPOST=1457208502,1457207588][/DOUBLEPOST]

I just posted this in another thread because I read it today on YouTube's help page about watch time:

"The algorithm for suggesting videos includes prioritizing videos that lead to a longer overall viewing session over those that receive more clicks."

That means videos with low retention can still rank high if they lead to a longer viewing session overall. I wish we could see session watch time in Analytics, it would save us a lot of speculation.

Yeah there was a period where I made a bunch of shorter videos with higher retention rate. I thought a super high retention rate would make a significant difference but it didn't. The high retention rate only netted around 2 minutes per video because the video was too short. I realized shorter videos weren't adding much to longer session time so I stopped doing that for the most part. Now I try to make them a little but longer.

Turns out a poor 15% retention rate on a 7 min video is slightly more total view time than higher 50% retention on a 2 min video. So if I can improve retention rate on that 7 min video then i am much better off.