7.1.10.1. YouTube as a Search Engine
7.1.10.1.1. Backgroud
https://www.youtube.com/watch?v=vQ0SKTcxjfI
YouTube Search Engine Issues to Consider:
- What video formats are acceptable for
- Uploading
- Downloading (the YouTube players: desktop, iPhone, android)
- How does YouTube identify videos
- What does YouTube know about the video
- How does YouTube rank search results
- What factors are used
- How are the factors weighed in producing the final ranking
- YouTube Recommendation System
- How does YouTube determine what to recommend
- How Does YouTube distribute videos
- How does YouTube monetize its website
- YouTube’s ContentID system
7.1.10.1.2. YouTube is Not a Search Engine in the Conventional Sense
- YouTube does not crawl the web for videos
- YouTube only searches the content that has been uploaded
- YouTube only returns videos, not web pages or pdf files
- Perhaps YouTube is best thought of as a content aggregator
7.1.10.2. Uploading to YouTube
- Youtube gathers information : title, description, tags,...
- Translations
- License and ownership, Syndication ,Caption , Embedding , Age restrictions, Categories
YOUTUBE UPLOAD CHARACTERISTICS
- 8 formats for uploading : MOV, MP4 (MPEG4), AVI, WMV, FLV, 3GP, MPEGPS, WebM + Aspect Ratio: the standard aspect ratios are: 4:3 or 16:9. When the video is uploaded to the site, YouTube will either leave it as-is (for 16:9) or add vertical black bars (for 4:3)
- The maximum file size you can upload to YouTube is 128GB.
- By default, you can upload videos that are up to 15 minutes long, though that can be extended by visiting https://www.youtube.com/verify
- there is no time limit for videos to remain on YouTube, unless
- You delete the Video.
- You delete your account.
- You violate Copyright or community guidelines
7.1.10.3. YouTube provides filters
- UPLOAD DATE
- TYPE
- DURATION
- FEATURES
- SORTBY
7.1.10.4. YouTube Ranking Factors
YouTube uses the following metrics for ranking search results:
- Meta Data
- video titles, descriptions and tags are core ranking factors
- include links to a website and social profiles
- Video Quality
- HD ranks higher than low quality videos
- Number of views, likes, shares and links
- Subtitles and Closed Captions
- captions can be crawled by the YouTube search engine and used for ranking
What is not known is how YouTube weights the individual factors to make up their final ranking
7.1.10.5. YouTube Recommendation System
Elements that are considered are the following:
- First, a user’s personal activity is initially determined (watched, favorited, liked, people he subscribes to, videos he watched to the end) – Next, an expanding set of videos is produced by traversing the co-visitation graph (other videos people watched after watching a given video)
7.1.10.6. Association Rule Mining
7.1.10.7. Google Search is Biased Towards YouTube Videos
A video rich snippet means that when someone searches for something on Google, you can have a small tiny video show up next to your result to let the user know that particular result (yours) has a video to help
7.1.10.8. Content Delivery Networks
A content distribution network (CDN) consists of a large set of content servers and a means for dynamically selecting servers based on knowledge of the location of the user and possibly the content being requested
7.1.10.8.1. YouTube’s Content Distribution Network Basics
7.1.10.9. YouTube Upload Architecture
7.1.10.10. Monetizing YouTube
YouTube challenges in the early days:
- YouTube had no way of making money, and its infrastructure is very expensive
- YouTube was being sued by content creators, as many of YouTube’s videos were uploaded illegally
- YouTube solved both problems at once, by
- Developing a system for spotting copyrighted content
- Allowing the copyright owner to decide if he wants to keep the content on the site and let ads appear, splitting the revenue with YouTube or taking the content down
7.1.10.10.1. Content ID
- YouTube’s solution was to create a fingerprint database of copyrighted content, called Content ID
- ContentID technology creates an ID Fingerprint File for copyrighted audio and video material, and stores it in a database.
- YouTube solicited cooperation from content owners asking them to submit copies of their content so YouTube could fingerprint them
- There are millions of reference files in YouTube’s Content ID database.
- When a video is uploaded, it is checked against the database, and flags the video as a copyright violation if a match is found
- When this occurs, the content owner has the choice of blocking the video to make it unviewable, tracking the viewing statistics of the video, or adding advertisements to the video
