Classification of Web Video Sites (Video Search Engines)

As users browse the Web, they are likely to encounter video on almost any site. If we focus on the sites that appear to be video portals or claim to offer video search, we can begin to discern several categories of video search sites. To complicate the matter, there are hundreds if not thousands of such sites, and of course the Web is evolving rapidly, with business models and content presentation strategies in a constant state of flux. Some Web destinations are amalgams of several differing approaches. In spite of this, in this section we attempt to point out a few general classes of video Web sites that users may encounter and that employ some form of video search capability.

Content Originators and Traditional Broadcasters

Examples of origin content Web sites include major TV networks, affiliates, major league sports (™) as well as an emerging class of Internet-centric producers such as Rocket-boom™ and CNet™. Content on these sites is typically from a single source, but due to co-ownership of content, the user may observe several different “brands” such as television network call letters often owned by the same company. Content is usually posted to the Web after it has aired with the business intent to extract additional advertising revenue from the content. However, we are seeing a trend to simultaneous release of content on the Web and on traditional distribution mechanisms. A second goal is to generate more viewers for the next episode to be aired. Going forward, the affiliated local TV stations may become disintermediated and suffer revenue loss as viewers move from the channel based consumption model and get the content directly from the national “broadcaster.” Network affiliates post local news and other content, and may receive content from the network for their site.


Sometimes called “Internet Broadcasters,” aggregators act as centralized repositories that give users a wider range of content sources, with the goal of providing content providers with more viewers for their content. Samples include Brightcove™, ROO® Media and the FeedRoom™. These sites can be “white labeled” and branded by others such as broadband Internet service providers so users may not recognize these names.

A second, more widely recognized class of aggregators includes MSN™, AoL™, Google™ and iTunes™. Business models vary widely and include rental, purchase, advertising, and subscription. The primary access method is HTTP streaming, but some sites also offer higher quality video via managed download.


Movie and video download sites such as MovieLink, Akimbo, and Cine-maNow support search and allow users to rent movies. The media is good quality and is typically downloaded to local storage using a download manager which must be installed on the users’ local machine. In addition to managing media file transfers to local storage, the client supports DRM which prevents copying the content and enforces the business rules for rental periods (e.g. keep for up to 30 days, play for 24 hours.)


In addition to sites featuring professionally produced video, there are a large number of sites designed for sharing user generated content (UGC), or more precisely, “user contributed content” (UCC). The most widely known of these is YouTube™ which was purchased by Google™ for $1.6B in stock in 2006, however there are many others in use. In some cases these are not much more than network storage, but most add features such as transcoding to a common format, entry and search of author-supplied metadata, social tagging, and even e-commerce. (e.g. Putfile, Vsocial). The content on these sites is mainly consumer generated such as video blogs, but may include a range of qualities and genres. Unfortunately it may also include pirated copies of copyrighted material as well.

In addition to the aggregators, video download and video share sites, the traditional search engine model involving Web crawl for content discovery is evident. However, blind crawling has taken a back seat to feed based content pull using various syndication formats which describe the media at a high level via XML. Also, top search sites now cross-index to broaden coverage (e.g. Google™ video results will appear in MSN™ searches.) There are hundreds of Podcast search sites which offer none of their own content, but rather direct users to a broad range of content providers.

Application Specific

Vertical search sites that cater to a specific audience or to a narrow range of source content may include more structured metadata since the source of video is more controlled. For example,™ offers video content from baseball games that have been highly annotated with detailed metadata indicating the player, the game situation, the stadium, etc. All of this data is accessible using an HTML forms interface to create very specific queries. Other sites such as IMDb or TVGuide® focus on video metadata such as movie information or guide listings, but may only contain preview clips rather than the full video content. In some cases, there is very little video available at all, and the sites may contain only related media such as photos of actors, box art, etc. The goal of these sites is to generate rentals or purchases in the case of movie sites, or to plan TV watching or schedule TV recording. On the other hand, the Internet Archive has a long history of offering both metadata and video for content in the public domain, or with very liberal copyrights.

Other Video Systems

There are an increasing number of IP applications for viewing Internet TV such as Joost™, or Miro™ which attempt to organize Web video feeds as channels whether they are in fact live streams or published as feed based media files. Video search is a key element for content selection, along with other means such as promotional placement or popularity rankings. Other video search applications may offer Web-based front ends, but require subscription. Media monitoring services allow subscribers to search and browse content as it was aired for various purposes including advertising verification, and corporate public relations. Media asset management (MAM) and digital asset management (DAM) systems are used for “in-house” production and archiving applications. These include work flow automation and may support Web-based distribution and monetization features, but at a minimum include a Web-based UI for administration, asset browsing and retrieval using metadata search.

Next post:

Previous post: