YouTube views on a 20 minute video are a terrible metric. Many people are going to start, notice the length, and stop watching. A better metric would be to compair the average +1/like/fav and shares of posts to each site.
Views are not measured based on whether or not people watch to the end, if you click on a video a view is counted, whether you watch it for 1 second or 20 minutes is only reflected in the internal Youtube metrics.
Google has never announced how they count views. They have announced that recommended videos are suggested based on the % of videos watched. I would think some of that same technology is being used in the view count.
> Google has never announced how they count views. They have announced that recommended videos are suggested based on the % of videos watched. I would think some of that same technology is being used in the view count.
What do you mean? Surely there is no magical technology or algorithms behind view count? "If a unique user watches a video increment the counter by 1" why would they need anything more complex than this?
I know there is a lot of abuse protection but I'm not sure how this fits in with my assertion that 1 click = 1 view. There was a bug a few months back with Youtube that didn't apply anti-abuse checking to Mobile views (viewing a video from a mobile device) and a group of blackhat people abused this and sold views: they would view a video 1,000,000 times as if it was being viewed from a mobile device and Youtube would count 1,000,000 views, even if it was all from the same IP.
I'm not sure how it's debatable that if 800 people click on a Youtube video link then 800 views will be counted, what reason would they have for not counting 800 views if 800 people clicked on it?
There is a difference between a page view and a video video view. The video views really should not increment unless someone watches them all the way through (or at least most of the way through). Since Google has not announced how they tally the view count I was just pointing out evidence that they are probably considering % of videos watched in the view count.
> The video views really should not increment unless someone watches them all the way through
Although this would be better it's not what they do. I'm a Youtube Partner and from the analytics I have access to I can see that no matter how long a user watches for a view is counted, otherwise they would not be able to provide "user retention" statistics.
"Absolute audience retention helps you see how often each moment of your video is being watched as a percentage of total views. Rewinding and rewatching can result in values of higher than 100%"
They do. You generally see popular videos stuck around 300 for a while though and if it is showing 800 it probably is accurate within a few hundred views.
