Tracing the Path of Online Video

In recent years, video has been blowing up in popularity across the Web and on mobile networks and most of us just watch it without thinking much about what it takes to bring it to our screens.  It’s simple to assume that it’s just ABC, Hulu, or MSNBC.com publishing streaming files to the Web, and yes that’s partly true.  Savvy users may be aware of companies like Brightcove that build white-label players and distribution platforms for name brand video providers.

But, as I’m writing in my dissertation right now, some of the most powerful and well-positioned players—obligatory passage points, if you will—in the online video distribution chain are companies that most people don’t know exist.  Here are three of the most essential Web video companies you’ve never heard of:

Transpera

The mobile market for video is huge, but it’s far from standardized.  It’s difficult to count on different phones and mobile devices supporting the same filetypes and codecs.  Even the display sizes of mobile devices are all over the place.  If you’re a video publisher, though, you don’t want this mess to diminish the viewing experience for end users, or worse, limit the market for your video to some minority of devices.  Circumnavigating this problem means converting your video to multiple formats, autodetecting the sort of device a video is being loaded on, and serving the right file to the right phone or tablet without the user even being aware that there’s a dialogue going on.   That’s a hard and expensive technical problem—especially when the roster of mobile devices and video standards out there seems to change on a monthly basis.  It’s also the problem that Transpera specializes in.  Publishers like Disney and CBS News give Transpera a feed of the video they want distributed, and Transpera takes care of all the messy standards issues, conversions, and autodetection.  On their way to you, Transpera packages the videos with advertisements that the company has sold against the content, giving the proceeds back to the publisher and taking a cut.

YuMe

The great thing for users about embeddable video is that it can go anywhere.  The horrible thing for advertisers about embeddable video is that it can go anywhere.  Publishers often maintain domain and IP blacklists, along with other technical restrictions, to keep their embeddable video from showing up on porn sites and unauthorized mirrors or being viewed in countries that don’t recognize international copyright.  But as big as these restrictions are, they still leave a lot of leeway for users to do with videos what they please.  This is appropriately celebrated as being great for free speech.  But the lack of control over the context in which videos appear tends to scare the crap out of advertisers—they want their commercials to appear next to wholesome, brand-friendly content and nothing else.  Think back to all the times you’ve heard about ads being pulled from traditional television programs when video advertisers found the content objectionable and you have an idea of how horrifying the Web can be for them.  This has made embeddable players, as popular as they are, extremely difficult for video publishers to monetize through advertising.

YuMe made its name by offering a solution to both parties.  Advertisers buy embeddable ads through YuMe, which packages them with the embeddable players of major video publishers.  Wherever an embedded player appears, YuMe’s service surveils the sites on which videos are served, automatically “collect[ing] detailed information about the in-page environment of a syndicated video player when it makes an ad request.”  Pages and sites deemed unacceptable are added to the company’s “constantly growing blacklist of more than 1.6 million Web domains that contain inappropriate content. When we receive an ad request from a video player, we use our domain-detection capabilities to confirm that the player is not embedded on a blacklisted domain before serving an ad.

Finally, the subjective quality of sites where a player is embedded is ranked by YuMe based on the information harvested by the ad service.  This information is then used to score the performance of publishers in terms of the general quality of the sites on which their players are embedded.  YuMe claims to then “regularly encourage publishers to offer greater campaign performance to advertisers by removing their players from lower-performing domains.”  In other words, YuMe threads a needle between the demands of publishers on the one hand and advertisers on the other, creating a compromise in which players can be embedded anywhere, but advertising is only packaged with them in “brand-safe” environments.

Nexidia

It’s widely acknowledged, if not accepted, that we’re now in a search-centric Web, where content has to rank well in search results to get circulation.  Social media are complexifying, but not necessarily altering this situation fundamentally.  We still search for tweets, after all.

The problem is that search currently operates on text, which makes indexing and finding video problematic, unless it’s packaged with excellent textual metadata.  Transcripts of video can be a tremendous help here, but transcription is a painfully slow and/or expensive process, and ultimately transcripts may help find the video, but not the part of it that users are looking for.  All this can put search-generated revenue out of reach for video publishers.  Nexidia provides software that can make a 60-minute video searchable in 17 seconds.  It breaks speech on the audio track into individual phonemes (the basic units of speech), which can be recognized by the software as words and matched against search results.  The software can also generate likely keywords for the video by identifying oft-used phrases.  After it’s done creating this metadata, the video is not only search-optimized, but the video itself is searchable; with the transcript matched to the location of phonemes in the video timeline, users can cue the playhead to a particular part of the clip simply by searching for it.  In short, Nexidia aims to fully integrate video and audio into the search economy.  Of course, media isn’t its only revenue stream—the company also markets the same software to intelligence agencies for data mining of telephone conversations.  So they’ve got that going for them, which is nice.

•••

Some of the technical problems solved by these companies may eventually fade away.  For example, we may eventually have a universal mobile video standard, and/or cell phones capable of playing a wider range of video formats.  Likewise, as some of the patented technologies behind YuMe and Nexidia become more widely available, publishers may be able to roll their own solutions to the problems these services provide.  But all of them are thinking long term.  Both Transpera and YuMe for instance, have used their position to develop large advertiser networks with high-profile clients ranging from PepsiCo to Unilever.  With these in place, they’ll have something to market even after the technical problems they negotiate vanish.  Nexidia, for its part, seems to be aiming for stability by pursuing a diverse portfolio of clients for its patented software; it’s the sort of company I can imagine being purchased by a larger firm down the road, both for its client roster and its technical resources.

In closing, though, I love to think about these companies through the lens of heterogeneous engineering—something I do at great length in the dissertation.  Here’s the gist: The path that video takes on its way to us is a series of compromises, threading a needle between the demands of different constituencies like advertisers, users, and competitors.  John Law once wrote about the volta, a Portuguese trade route to India as being the literal inscription on the map of compromises between “Portuguese system builders and their adversaries, that is, the winds, the currents, and the capes.”  Moreover, the forces shaping the volta were not merely physical ones, but social, commercial, and political demands by groups ranging from royalty to sailors to navigational astronomers.  Digital distribution, and the path that content takes to us online, is much the same.  It’s a path made possible by the lashing together of heterogeneous resources from technologies to social media accounts to advertisers to lobbyists, but it’s also one that wends its way around the forces imposed by different interests.  And at the moment, for many publishers-cum-system builders, part of that path is running through companies like these.

Postscript: I should add, for posterity, that I’m not endorsing any of the above companies—there are others like them that could have served equally well as exemplars.  My aim was rather to point out by example the numerous and largely invisible intermediaries that are currently growing up along the online video distribution chain, both enabling the transmission of content and impacting where it goes.

[Image Credits: Blackberry; Transpera Logo; YuMe Logo; Nexidia Logo; Portuguese Ship]