captions: helping users, helping yourself

YouTube now supports video captions.  Captions are uploaded as a separate text file.  YouTube also provides automatic translation of these captions into many languages.  Web tools for creating caption files are free and easy to use.  I recently used overstream to create captions for a docudrama some friends and I made.  Making captions takes time and is an art in itself, but technically it is now easy to do.

Even in its "advanced search" page, YouTube doesn't offer search limited to captioned videos.   Search limited to captioned videos would be an easy feature to implement, because the associated caption file readily indicates a captioned video.[1]  This search type would be helpful to deaf and hard-of-hearing persons.  Moreover, a person's use of it would provide valuable information for delivering useful ads to that person, e.g. social events and social networks for deaf and hard-of-hearing persons, hearing aids and accessories, etc.

YouTube has a lot of room to improve its revenue.  It could implement a premium-paid tier, which might offer videos longer than 10 minutes and better video quality.   Its search-choose model with Video ID has much potential.  YouTube has added video annotations, which users value and which over time will provide more information for targeting advertisements.  The APIs for its video platform will over time provide more information for targeting advertisements.  Despite all the hype about Hulu's revenue, I still think that worse content with higher-value advertising can beat better content with lower-value advertising.

Note:

[1] Videos with captions included within the video stream would be much harder to identify.  Unlike the "closed captions" discussed above, these "captured captions" can't be turned off and on, can't be automatically translated, and can't be searched as text.  The presence of videos with captured captions doesn't seem to me to be a good reason for not providing search limited to (closed) captions.

Tags: , , ,

YouTube's new business model

YouTube's Video ID technology points to an important new business model.   Video ID finds copyrighted content that has been uploaded to YouTube.  It then gives the copyright holder the choice to block, promote, or monetize that content.  Copyright holders benefit from being able to exploit free, decentralized distribution and promotion of their work.  Google/YouTube benefits from leveraging the value of its search expertise and advertising platform.   Most importantly,  much value is created by allowing licensing decisions for content to be made for small units of value quickly, at low cost, at high frequency, and with directly relevant economic data.[1]

Content pricing and licensing systems are extraordinarily inefficient.  Consider, for example, that the U.S. Copyright Act of 1909 included a provision establishing a two-cents per song royalty for mechanical recordings of musical compositions.  This provision was designed to give player-piano companies equal access to music for their machines.  The two-cents royalty subsequently applied to musical phonorecords.  Because the companies making the records paid the royalties, radio stations could play recorded music without having to negotiate or pay any royalties to music copyright holders.   The two-cent royalty remained law until Dec. 31, 1977.[2]  Thus this government-established price was in effect for sixty-eight years through large changes in music recording and playing technology.  This situation reminds me of iron pot that a fellow graduate student from the former Soviet Union showed me in the early 1990s.  Cast into the metal of the pot was the pot's price.  This Soviet approach to pricing kitchen pots was probably less inefficient than the two-cents royalty established by the Copyright Act of 1909.

YouTube's new search-choose model is a much more efficient business model for licensing content.  Copyright holders make licensing decisions for specific copies with knowledge about circumstances and amount of attention the copy is attracting.  Policy rules can easily be established and changed for these informed licensing decisions.  Compared to traditional licensing approaches, the search-choose model provides much more relevant information for licensing decisions and much lower transactions cost for those decisions.   This is a major, under-appreciated value of the search-choose model.

Free, decentralized distribution and promotion of content potentially has great value for copyright holders.  Attracting attention to content is expensive.  Movie producers, for example, often spend more promoting a movie than they do in making it.  Peer-to-peer diffusion of information and actions among social networks has always strongly affected aggregate patterns of behavior.  Online social networking tools make social networks even more powerful. Attempting to suppress persons' natural propensity to share, discuss, and promote content mainly pushes such behavior underground and alienates potential customers.  The search-choose model transforms a unsolvable problem into a significant business benefit.

The search-choose model better suits web video than web text.  ISPs might attempt to insert text ads into copies of copyrighted textual content found on webpages.  Web mail providers could insert (additional) advertisements in copies of copyrighted textual content found in emails.  However, ISPs don't have businesses structured to serve ads, and inserting ads into webpages problably would anger ISPs users.  Web mail providers, on the other hand, have little incentive to insert additional advertising with shared revenue.   In contrast, web video providers have an incentive to address the licensing problem, and they can do so in a way that's not likely to anger their users.

Having a lot of video on a common platform makes finding instances of copyrighted work easier.  It also makes inserting ads easier.   YouTube thus already has big advantages in offering a search-choose model to copyright holders.

Note:

[1] Google recently stated that 90% of its 300+ Video ID partners have chosen to monetize found content rather than block it.  Companies choosing to monetize found copyrighted video include major media companies such as CBS, Universal Music, Lionsgate Entertainment, and Electronic Arts.

[2] Historical versions of U.S. copyright acts are helpfully available at copyrightdata.com.  The U.S. Copyright Act of 1976 raised the royalty to  2.75 cents.  Broadcast radio stations in the U.S. have retained to the present the right to play recorded music without the need for negotiating a license or paying a royalty.  The situation is much different for Internet radio.  Copyright royalty rates in the U.S. are now typically established through the Copyright Royalty Board (CRB) for periods of four to ten years.  That CRB's price-setting process is far from simulating that of an well-functioning, decentralized decision mechanism.

Tags: , , , ,

television's moving into the toilet

YouTube is currently selling ads against less than 3% of its inventory. YouTube's advertising revenue is likely to total about $200 million for 2008. One explanation for YouTube's relatively poor advertising performance is that advertisers are concerned about what appears around their ads:

Some big advertisers, [Sean Muzzy, media director at Neo@Ogilvy, a digital ad agency owned by WPP Group's Ogilvy & Mather] says, haven't been comfortable that their ads might appear next to amateur videos. [WSJ]

In at least one major hotel in the DC metro area, CNN is projected onto the mirror in front of the sink in the men's room. That puts persons in the position of watching television ads while they are washing their hands in immediate preparation for leaving. These circumstances don't seem like good positioning for advertising.

Better advertising positioning would be on the inside of toilet-stall doors. Sitting on the toilet is similar to the position and time-relations for traditional television-watching. In the past, television has conceded toilet time to newspapers. In today's highly competitive media market, that doesn't make good business sense. You should expect to see television moving into the toilet.

Tags: , , , , ,

stories largely missing in online video

While stories are staples of television programming, the most popular YouTube videos are predominately music videos from major record labels. Among the all-time most viewed YouTube channels, the leader is the Universal Music Group channel. That channel has about as many views as the total of the nine next highest viewed channels, all but one of which are also mainstream commercial music video channels. Among all-time most viewed YouTube single videos, there's slightly more diversity in form and producers. The all-time most viewed YouTube video is the mockumentary Evolution of Dance. However, it has only about 1% more views than the next leading video, which is an RCA Records music video. Major record company music videos account for six of the top ten most viewed Youtube videos.

Online video isn't succeeding in telling stories. The popular YouTube music videos typically communicate an emotion or feeling, not a story. Video appears on many sites besides Youtube; about 49% of videos viewed are on sites with less than 1% of total video views. However, the average duration of online video viewing across all sites is only 2.8 minutes per video.[1] That's not long enough to develop much of a story.

Flickr has embraced video in a way that gives little room for video story-telling. Video on Flickr is limited to a maximum of 90 seconds. The idea of Flickr video is a "long photo"; video that's "personal" and "simple -- not overproduced or slick." Telling a story with video is much more difficult than taking a photograph. Video story-telling typically requires multiple scenes, often multiple takes and multiple actors, and usually considerable editing. Flickr video clearly is not meant for video story-telling.

Across the U.S. population, online video viewing time currently amounts to only about 3% of television viewing time.[2] Online video viewing time is unlikely to come close to television viewing time unless online viewers start to watch many more story-oriented videos.

Notes:

[1] Online videos viewed in the U.S. in March, 2008, based on comScore data. The average duration of the 20 all-time most popular YouTube videos, weighted by popularity, is 4.7 minutes. So YouTube videos don't appear to be driving down the average duration of all online video viewing.

[2] The American Time Use Survey, which covers persons in the U.S. ages 15 and older, shows about 4600 minutes of television watched per person per month in 2006. comScore states that, in the U.S. in March, 2008, "average online video viewer watched 235 minutes of video." The source gives 139 million online video viewers, who are 73.7% of the "total U.S. Internet audience". Those figures imply a total U.S. Internet audience of 190 million. There are about 240 million persons in the U.S. ages 15 and older. Using this latter figure as the base implies 136 minutes of online video watched per person per month.

Tags: , , , , ,

video searching and ad targeting

Searching video is notoriously difficult. That implies a scaling problem for a large video repository like YouTube. Given relatively fixed amounts of video tagging and category information, more videos imply relatively less information for searching among videos and targeting ads to videos.

YouTube's announcement of new APIs for external use of the YouTube video platform points to better information for video search and ad targeting. Videos uploaded and played through a third-party site are segmented by the specific third-party site. Third-party sites also have differentiated text correlated with the content of associated videos. This additional information allows better video search and better video ad targeting compared to what's possible for videos uploaded and viewed on the YouTube site.

Online video services have many different niches. Decentralizing video uploading and playing provides a more scalable business for YouTube.

Tags: , , , , ,

taking rights seriously

False or excessively broad claims to rights, if taken seriously, could have devastating effects on content businesses. For example, U.S. National Football League (NFL) broadcasts include the following statement:

This telecast is copyrighted by the NFL for the private use of our audience. Any other use of this telecast or of any pictures, descriptions, or accounts of the game without the NFL's consent, is prohibited.

The claim, "This telecast is copyrighted by the NFL for the private use of our audience," is absurd. Copyrighting a telecast is not necessary for the private use of it, nor is advancing that use a credible explanation for the NFL's copyright action. The problem is not just that the NFL has not expressed a credible business justification for its copyright. The second sentence of the NFL's statement seems to imply that football fans need permission from the NFL to discuss games ("accounts of the game") that they watch on television. That's an impressive anti-social business-destroying effort.

The NFL has not yet succeeded in destroying its business. Perhaps that's because because football fans recognize copyfraud. The NFL recently has shown no respect for copyright law. The RIAA has executed astonishing initiatives to destroy the music business. If the NFL is serious about destroying the football business, it might run a few plays from the RIAA's playbook.

Shrewd and successful new media businesses seek to become platforms for users to share and discuss users' works. YouTube's terms of service state:

For clarity, you retain all of your ownership rights in your User Submissions. However, by submitting the User Submissions to YouTube, you hereby grant YouTube a worldwide, non-exclusive, royalty-free, sublicenseable and transferable license to use, reproduce, distribute, prepare derivative works of, display, and perform the User Submissions in connection with the YouTube Website and YouTube's (and its successor's) business, including without limitation for promoting and redistributing part or all of the YouTube Website (and derivative works thereof) in any media formats and through any media channels. You also hereby grant each user of the YouTube Website a non-exclusive license to access your User Submissions through the Website, and to use, reproduce, distribute, prepare derivative works of, display and perform such User Submissions as permitted through the functionality of the Website and under these Terms of Service. The foregoing license granted by you terminates once you remove or delete a User Submission from the YouTube Website. [first bold type in original; second, added here]

Google and BSkyB (Sky) have teamed up to produce SkyCast. This video service offers users a much different deal:

If you send us videos, messages or other content, we will be able to use your content in any way we like. So, we might decide to put your content on one of our other services, like TV, or give it to someone else to put on one of their services. We might even decide not to use it at all! If you decide to take your content off the Service, Sky can still use it in any way we like.

In addition:

You waive all moral rights in relation to your Content.

Moral rights, such as Article 6bis of the Berne Convention, apparently can be waived in some jurisdictions. While SkyCast filters submitted content, its terms of service declares "thou shalt not submit content" that:

1.1 is in breach or promotes the breach of any third party rights (including third party intellectual property rights);
1.2 is defamatory, offensive or libellous;
1.3 promotes racism, bigotry, hatred or harm of any kind against any group or individual or would subject any person to ridicule or cause other people to shun or avoid such an individual;
1.4 harasses or advocates the harassment of another person or persons;
1.5 promotes conduct that is abusive, threatening, obscene or distasteful;
...
1.24 refers to any arrest of an individuals [sic] or any active court proceedings.

Moreover, in conjunction with the opportunity to offer their work to SkyCast, users are required to accept liability to SkyCast and third parties:

5.5 You will reimburse Sky and any third party who provides services to you as part of the Service for any losses, costs or damages incurred by Sky and/or any third party, on demand, arising out of:
5.5.1 your use of the Service, or anybody else that your [sic] allow to use the Service using your SkyCast Profile; and/or
5.5.2 your breach of these Terms of Use.

How would one assess the financial risk of this liability given the terms of service?

I cannot imagine that any rational, informed users would actual agree to submit work to SkyCast. Put different, if users take seriously their rights as currently set forth in SkyCast's terms of service, I think SkyCast's business is worthless.

Tags: , , , , ,

novel content creation

In the middle of the eighteenth century, a new form of content creation grew rapidly in Britain. The new popular content was long, realistic but fictional narratives of ordinary individuals whose lives nonetheless were put forward as significant for everyone. These works were called "novels."

Novels were YouTube in the eighteenth century. Authors of novels included roughly equally men and women, "leisured gentlewomen, high-profile aristocrats, obscure vicars, and curates, sea captains, destitute merchants' wives, reformed and some unreformed prostitutes, over-archieving adolescents, and pious autodidacts." A leading novel publisher in Britain explained in 1769:

all we have hitherto published have been sent to us unsolicited from their authors, without any stipulated pay, promise of reward, or previous agreement whatsoever, either by ourselves or any other person for us.

Most authors didn't even have their names attached to their work: about 72% of new novels published in Britain and Ireland, 1770-1799, were published anonymously.[1]

Most novelists received little monetary compensation for their works. Sometimes authors funded publication of their works, or assumed liability for losses from publication. When authors sold their copyrights, the typical payment was low but payment variance was high. For example, in 1787 a publisher bought a copyright from an obscure novelist for £5, while in 1794 the same publisher bought a copyright from a well-known novelist for £500. The median payment to British novelists among surviving copyright sales receipts, 1770-1799, was about £29. That was about the annual earnings of building craftsmen. By 1860 in the U.S., only 216 persons declared their profession to be "author." In contrast, 3,154 persons declared their occupations to be the newer occupations of daguerreotypist and photographer.[2]

While authors of novels typically did not earn enough money to sustain themselves, novels quickly dominated popular book reading. In the late eighteenth century, purchasing books would have been a financial hardship for most persons. Social and commercial libraries, however, made books much more readily available. At the end of the eighteenth-century in Britain and in the U.S., novels comprised 40% or higher shares of titles in commercial circulating libraries. Limited evidence from circulation records suggests that the share of novels among books borrowed was probably higher than 50%.

Novels had well-recognized popular effects in the second half of the eighteenth-century. Commentators observed that a rage to read (Lesewut) was gripping the German lands. Reading the novel The Sorrows of Young Werther [Die Leiden des jungen Werther, 1774] prompted young men to dress like the character Werther. Reportedly about 2000 young men committed suicide in sympathy with Werther. In France, the novel Julie, or the New Heloise [Julie, ou la nouvelle Héloïse, 1761] created a large body of weaping readers. American readers of the highly popular novel Charlotte Temple (1794) tended her purported grave in New York City.

The prevalence of imitations, mock sequels, and parodies among late eighteenth-century novels marked them to contemporaries as a "faddish, superficial make of literature." Following Henry Mackenzie's popular Man of Feeling (1771) came the anonymous and forgotten Man of Failing (1789). Only one year after Hannah More's highly successful didactic work, Coelebs in Search of a Wife (1809) came Coelebs in Search of a Mistress (1810), under the likely authorial pseudonym Sir George Rover. Many now-forgotten novels of the late eighteenth and early nineteenth centuries were hastily written, poorly crafted works.[4]

Many influential persons in the eighteenth and early nineteenth centuries considered novels to be wastes of time, causes of ill health, and corrupters of virtue and morals. In 1794 a London reviewer described a new novel, Widow, as "fashionably vicious." The reviewer warned against reading such novels and implicitly proclaimed the importance of the reviewer's own work:

O! for a warning voice to prevent those, at least, in whom age has not yet destroyed the capabilities of improvement, from dreaming away their hours in turning over publications like these.

Another author more directly warned against reading for amusement and diversion:

To read a book merely in order to kill time is an act of high treason toward humanity because one is belittling a medium that was designed for loftier purposes.

Others described reading as a cause of masturbation and other injuries to good health:

the obligatory position, the lack of all physical movement when reading, combined with the violent alternation of imaginings and feelings [create] limpness, bloatedness and constipation of the intestines, in a word hypochondria, which has a recognized effect on the genitals of both sexes, particularly of the female sex [and creates] coagulations and defects in the blood, excitation and exhaustion of the nervous system, as well as conditions of langour and weakness in the whole body.

The effect of novels on manners and morals was an acute concern. Novel reviewers in London publications in the late eighteenth century described reviewed novels as "one of these pernicious incentives to vice that are a scandal to decency"; "utterly repugnant to every idea of delicacy and honor"; and, "Written solely for the use of circulating libraries, and very proper to debauch all young women who are still undebauched." A popular American author of conduct literature noted in 1831:

Of late years, the circulating libraries have been overrun with profligate and strongly exciting works, many of them horribly exciting. I have deep prejudice against the whole class. The greater the genius displayed, the more dangerous the effects. The necessity of fierce excitement in reading is a sort of intellectual intemperance; and like bodily intoxication, it produces weakeness and delirium....They have a most unhealthy influence upon the soul....

From a less evangelical, more republican position, novels were described as "murdering of freedom of thought and the press." Similarly quotations from late eighteenth and early nineteenth century sources could be multiplied endlessly. All the ill effects ascribed to television, video games, and the Internet in recent years apply equally well to reading novels two hundred years ago.[5]

Novels and tabloids changed persons' relationships to printed words. Historians of the book have described a "reading revolution" (Leserrevolution) -- a shift in the distribution of reading from intensive reading (reading a book, particularly the Bible, carefully and repeatedly) toward extensive reading (reading one new novel after another). More generally, the rise of empirical science shifted authority from a bounded text to an unbounded corpus of evidence. The expansion of print functioned like science in the realm of imagination and culture.

About 1854, a man who grew up in Ridgefield, Connecticut recalled his youth:

Books and newspapers -- which are now diffused even among the country towns, so as to be in the hands of all, young and old -- were then scarce, and were read respectfully, and as if they were grave matters, demanding thought and attention. They were not toys and pastimes, taken up every day, and by everybody, in the short intervals of labor, and then hastily dismissed, like waste paper. The aged sat down when they read, and drew forth their spectacles, and put them deliberatively and reverentially on the nose. These instruments [spectacles] were not as now, little tortoise-shell hooks, attached to a ribbon, and put off and on with a jerk; but they were of silver or steel, substantially made, and calculated to hold on with a firm and steady grasp, showing the gravity of the uses to which they were devoted. Even the young approached a book with reverence, and a newspaper with awe. How the world has changed![6]

Yes, the world has changed. The world continues to change.

The history of the novel helps to provide some perspective on current media developments. Today major media companies are struggling to set up user-generated content divisions to foster production of user-generated content. At the same time, author and blogger Andrew Keen is promoting his new book entitled, "The Cult of the Amateur." He recently changed the book's subtitle from "How the democratization of the digital world is assaulting our economy, our culture and our values" to "How today's Internet is killing our culture." Without a whiff of amusement, a leading blogger laments, "There's no food for thought in this book." Even more seriously, a business intelligence company recently reported that user-generated videos "made up 47% of the total online video market." The report proclaimed, "consumer usage exploded in 2006 but revenues will prove slow to develop. The honeymoon period for user generated content is over."

The entertainment business is as strong as ever.

* * * Notes and Sources * * *

[1] Raven, James (2000), "Historical Introduction: The Novel Comes of Age," in The English novel, 1770-1829: a bibliographical survey of prose fiction published in the British Isles, gen. eds. Peter Garside, James Raven, and Rainer Schöwerling (Oxford: Oxford University Press), vol. 1, pp. 17, 51, 42.

[2] Id. pp. 52-53, which records 51 copyright receipts. In 1757, a journalist complained that a bookseller-publisher "never paid to any author for his labour a sum equal to the wages of a journeyman taylor." Quoted in id. pp. 50-1. Building craftmen in Southern England, 1736-1773, earned about 24 pence per day, or about £30 for a full year of work. See B.R. Mitchell, British Historical Statistics (Cambridge: Cambridge University Press, 1988) p. 165. The data on occupations is U.S. Census data, collated and discussed in Galbi, Douglas (2003), "Copyright and Creativity: Photographers and Authors."

[3] Raven (2000) p. 85-6, 93; Winas, Robert B. (1975), "The Growth of a Novel-Reading Public in Late-Eighteenth-Century America," Early American Literature, IX.

[4] Raven (2000) pp. 15, 34. Garside, Peter (2000), "The English Novel in the Romantic Era: Consoliation and Dispersal," p. 58, in Garside, Raven, and Schöwerling, vol. 2.

[5] Raven (2000) p. 119; Wittmann, Reinhard (1999), "Was there a Reading Revolution?" in A history of reading in the West, eds. Guglielmo Cavallo and Roger Chartier (Amherst : University of Massachusetts Press) p. 301, quoting J.A. Bergk, Die Kunst, Bücher zu lesen (1799) p. 69 and Karl G. Bauer, Über die Mittel, dem Geschlechstrieb eine unschädliche Richtung zu geben (1791) p. 190; Raven (2000) pp. 17, 114, 101; Lydia Maria Child, The Mother's Book (1831), Ch. VII; Wittmann (2000), quoting, original source not cited; see Dmitri Williams (2003), "The Video Game Lightening Rod: Constructions of a New Media Technology,1970-2000," Information, Communication & Society 6:4 pp. 523–550.

[6] Goodrich, Samuel G. (1857), Recollections of Lifetime (New York: Miller, Orton & Co.) vol. 1, p. 86, quoted in David D. Hall, "The Uses of Literacy in New England, 1600-1850," in William Joyce et al., eds, Printing and Society in Early America (Worcester, Mass.: American Antiquarian Society, 1983) p. 21.

Tags: , , ,

YouTube's success

How has YouTube succeeded? YouTube makes uploading and sharing videos free and easy. It has hosted some popular high-quality clips and stimulated calls for a copyright-brawl-of-the-millennium. These are well-recognized aspects of YouTube. But consider the outpouring of sympathy for Martin. YouTube has also succeeded as an innovative communications service.

Social networks and communications services are closely related. Robert Young, an insightful industry analyst, recently noted:

communications ultimately serves as the anchor feature and the driver of retention and growth. …when dealing with an online community, that one lasting activity is almost always communications. … Social networks, which are rapidly becoming the portals of the next generation, must place high strategic priority on their communications functionality if they wish to continue their pace of traffic growth, usage, and retention.

Mirroring this remark, Norman Lewis of Orange, a mobile communications service provider, suggested at the Telco2.0 Industry Brainstorm:

Any future applications which do not have a social networking aspect to them will be irrelevant. If we don't understand that, we won't have a business in the future.

YouTube shows communication service providers that video can be important driver of communication. On the other hand, sharing video, like providing VoIP, is a service that many providers potentially could offer. If YouTube doesn't link itself closer to enduring real-world social networks, it may not have a business in the future.

Tags: , ,

YouTube and indecency

No nudity, no violence, no profanity -- can you f@$%^&* believe this $#^+! YouTube has declared that the Galbi Brother's Epic 800-Meter Challenge video "may contain content that is inappropriate for some users." So they want all the sports fans to register before they watch the video (also available without registration here and here).

Indecency is a major communications policy issue. How this issue will play out for online video sharing isn't clear. Don't do evil is a good principle both for service providers and users (see truth #6, which extends to users having fun, too).

I sent YouTube a polite email requesting that YouTube reconsider the appropriateness of the Galbi Brothers' Epic 800-meter Challenge video. That was on Friday, March 3. YouTube hasn't yet responded to my email.

I think that respect for users implies that YouTube should have some fair process for reviewing "appropriateness" classifications. The same goes for copyright rule enforcement. This isn't just good business practice -- it's also common decency.

Tags: , ,