YouTube is a practically infamous company by this point, when people talk about watching user created videos online, they are often talking about YouTube or some late comer to the scene like TikTok.
Part of keeping people on the site, is keeping them engaged with the content that is on the site.
YouTube is quite successful at this, they currently claim to have 1.86 billion active users on the platform, so whatever they are doing with their current algorithm is clearly working. Claiming that you have slightly less than a quarter of the World's population using the platform every month is a rather staggering claim to make. (https://www.semrush.com/blog/youtube-stats/)
This means that the mechanics of the algorithm, and the practicalities of its use have been almost exhaustively covered by a number of academics and other researchers looking to scoop some success from something that already clearly works.
A large portion of this might be attributed to the recommended videos listings that are a part of the home page for the YouTube app on both smart phones and smart TVs. The ability to tailor content to the interests of the viewers leads to a continued stream of content consumption, otherwise known as one of the sources Screen Time and for its dreaded impact on your sleep.
This infinite scroll of videos that are presented to you are produced by correlation of videos that you have previously viewed. This metadata cloud follows you across the various platforms, and your Google Searches will attempt to sell you products through Google Ad Sense, this includes learning from your YouTube watching habits.
My recent studies into Information Technology has presented a better understanding of the various forms of metadata and the mechanics on how information itself is spread. Creating a mechanism to turn the looks, clicks, and watches of a viewer is already being used to create targeted advertisements. With the upcoming addition of machine learning we as a race will be creating a new epoch of data processing that creates this meta existence of super surreal computation as floating data points turn into webs of interconnected profiles.
So the study that I have departed on will look at the recommendations that people get on their YouTube accounts. I will be doing brief surveys that will include the general demographics of the person being surveyed, what is on their recommended watchlist for their home page at the time of the survey, and a listing of the channels that they are subscribed to.
The goal of this investigation is to reflectively determine the interests of the persons by taking into consideration what videos they are watching and how this reflects their interests. Namely on what tracks of thought they are currently engaged in, topics they follow closely and a brief custom cut survey in response as well as a general survey created to determine the user (surveyed person(s)) awareness of their watching habits.
Elements for Research Survey
What are the videos being recommended (Channel Name, Video Length, Number of Previous Watches, Amount of Time Watched a Day)
The Channels that the Surveyed Person is subscribed to
The amount of awareness for the Surveyed Person's watching habits
Any recent deviation in watching habits
General demographics of the Surveyed Person (Age, Gender, Religion, Income Bracket, Criminal History, etc.)
Disclaimer
Here is a little poll to help me with site demographics. Please take the time to leave an answer, your input is valued.
Is this a worthwhile investigation?
Yes
No
Comentários