In my long years of working with video I have been encouraging everybody who publishes video to also provide textual representations of video, which includes subtitles/captions, but also includes metadata and hyperlinks that will enable video to become part of the content networks of the Web.

The key advantage for me is not accessibility, but it is to increase the value of the content. Content that knows more about itself and can expose that to machines is inherently more valuable than content that is just a dark collection of bits.

Added Value 1: automated translation

Once a time-aligned transcript such as a caption file is available, the video can expose this to a translation engine and provide itself in any language. This capability is now available for some videos on YouTube, e.g. the following winning Eurovision song of this year:

You have to click on the triangle icon on the video player bottom right while the video is playing, and then follow the red “CC” menu to go to “Translate” and turn on subtitles in a chosen language.

Even if Google’s automated translations are not 100% accurate, they still make the content accessible to a much larger audience than if they were not available. And all of this basically “for free” through the automated translation engine.

Added Value 2: increased user attention

Interestingly, recent research has shown that captions and subtitles don’t only make content more accessible to the hard-of-hearing, but also to well-hearing people. Where a video file has captions, 80 percent more people watch the entire video to its completion.

Achieving complete views is one of the most difficult challenges video publishers face, since people loose interest and attention fairly easily in our modern world of media over-stimulation. So, anything that can help people focus their attention longer is great news.

Just imagine the increase this can bring e.g. to the value of post-roll ads and to closing titles that contain the brands that sponsored the creation of the video.

Added Value 3: video SEO

Typically for a video the only text that is available and indexed by search engines are “title”, “description”, “tags”, and “categories”. These are fairly limited when you consider all the action and information that is inherent in a video.

Once a time-aligned transcript such as a caption file is available for a video, search engines are able to index that text together with the minimalist other text related to a video, thus making the video a whole lot more discoverable.

It is a shame that YouTube’s caption files are not yet indexed by Google, but do not fear: Google already has the technology and is using it on the Google video site:

We can only expect that it will be available on YouTube soon, so if you want to give your videos a huge SEO boost, think about uploading a caption file.

Further applications will certainly emerge to make better use of the annotated video content, such as automated summarisation, search that points us directly to offsets (see media fragment URIs), and automatically created mashups based on keywords.

If you want to do the captioning yourself, there are now some nice tools that work with YouTube. In Australia we also have captioning services that can do it for you, such as the Australian Captioning Centre and caption.it. Also check out Media Access Australia who have a large collection of resources and information about captioning in Australia.

David Hallerman from eMarketer reckons that “the sharp escalation of professional video content on the Web—mainly from TV networks—is creating a viable base for brand marketers”.

Add to this the standardisation efforts of the IAB for in-stream and companion ad formats for video and you can foresee a lucrative new advertising market emerge around online video.

• Linear Video Ad With or Without Companion Ad
• Linear Video Interactive Ad
• Non-linear Overlay Ad
• Non-linear Non-Overlay Ad

The document is interesting since it summarises the state of art in video in-stream metrics – metrics that we are also preparing to graph in the Vquence metrics mix. While it specifically addresses ads, the metrics are really about measuring a user’s interaction with a linearly playing video.

The following default metrics for in-stream video ads have been defined and are prescribed as the basis of every reporting:

• video impression (or view)
• companion ad impression (or view)
• video click-through
• completed video play at normal playback speed
• time spent viewing at normal speed in sec
• percent of un-duplicated video viewed at normal speed

The following additional events may be reported:

• audio mute / un-mute
• video window collape / expand
• video pause /resume
• video rewind

Further, for non-linear ads that run in parallel to the video, the following metrics are defined:

• ad impression
• click-through on ad
• accept invitation to view additional ad aread
• minimise ad
• close ad
• overlay view duration

It seems to me that we are really only at the beginning of understanding what we need to measure around video and video advertising. It is good to see that the IAB have summarised the current state of the art. These metrics provide first standards towards calculating a CPM-type value for a video advertising campaign. Yet, it also states taht the value of watching a video ad is not just defined by the number of impressions, but also by a complex set of duration and event measures. I expect we will see a new kind of standard measure emerge over time for estimating the value of a video ad from these metrics – a measure that includes impression, event and duration measures.