The Long Tail: Going Local in the YouTube Global Domain
G. R. Boynton

In 2008 82 million Americans uploaded something of themselves to the web (Ostrow). It may have been a photo, a video, a comment or an essay/blog, but it was self expression going public on the web.

In January 2009 Americans conducted 13.5 billion searches (comScore). Information search on the web continues to grow at just under ten percent from year to year with Google dominating the field.

And in January of 2009 the number of live streams of video reached 10 billion (Abrecht). And "more than 147 million U.S. Internet users watched an average of 101 videos per viewer in January." (Comscore) And entertainment goes to the web in a big way.

These are very large numbers, and they are particularly impressive given the recency of these forms of communication. Blogging, writing for the web, started in about 1998 (Wikipedia, Blog). Flickr, the leader in uploading photos to the web, went into business in 2005 (Wikipedia, Flickr). Google opened up for business in 1998 (Wikipedia, Google). And YouTube, the dominant leader in video streaming, opened in 2005 (Wikipedia, YouTube). Ten years is all it has taken to move self expression, information search, and entertainment to the web along with the many forms of economic transactions that have also migrated there. It is an understatement to say the social and economic structure of communication is changing exceedingly swiftly. This is breakneck speed for cultural and economic change.

The growth is going on in parallel with the failure of the media of the past. Newspapers are collapsing all over the U.S. (Kurtz). Even The New York Times teeters on the edge of bankruptcy. Television is coming unbundled (Arnago). First it was digital video recorders that played havoc with the channel structure. But as the technology becomes capable TV shows are moving to the web where they can be viewed whenever and wherever the viewer is.

And politics is not immune from these new media developments. President Obama made his first speech to a joint session of Congress and congress responded by twittering their reactions while listening (Stone).

The presidential election: 2008 was the first YouTube election. The technology that had been inadequate and the aggregation and streaming systems that did not exist in 2004 were in active use by 2008. In the primary race up to Iowa and New Hampshire every candidate had a YouTube channel. Romney led the way with more than 600 videos, but Obama and Paul led the way in views. Between January 1, 2007 and the election the McCain campaign posted 330 videos to YouTube. During the same period the Obama campaign posted 1820 videos. If you use July 1 through November 4 as dates of the presidential election campaign the McCain campaign put 122 videos on YouTube that were viewed more than 20 million times. The Obama campaign posted more than 700 videos to YouTube and they were viewed more than 42 million times.

In the middle of September the more traditional web presence, campaign website, of the two campaigns was compared by the Pew Research Center Project for Excellence in Journalism (Pew). They reported the following.

• Since the Republican National Convention, the official McCain website, www.johnmccain.com, has substantially improved its customization and socialization tools to encourage online networking with fellow supporters and offline grassroots activity. Despite this, it still lags behind Obama's site in various ways.
• Obama's Web site, www.barackobama.com, makes it much easier for supporters to take action. They can receive up-to-the-minute campaign news, pick up talking points, download campaign posters and flyers, make computer-assisted phone calls to undecided voters in swing states, and map out door-to-door canvassing operations in their area.
• Even after the enhancements to McCain's site, Obama has more MySpace friends by a nearly 6-to-1 margin, more Facebook supporters by more than a 5-to-1 margin, twice as many videos posted to his official YouTube channel, and far more YouTube channel subscribers, by an 11-to-1 margin.
• Obama's site links to mainstream media news stories about his candidacy more frequently than does McCain's, which tends to bypass the mainstream media and instead link in its "news" section to campaign-generated press releases. That tendency has ebbed somewhat recently, as the site has begun linking to news stories about Palin.
• The word "change" -- the motto of the Obama campaign -- is now less prominent on the information pages of the Obama site than on McCain's. On the Republican's site "change" is among the top 20 most frequently used words.
• The Obama website provides far more text than McCain's, by virtue of the extensive archive of Obama's speeches (in August alone, there were 50,676 words on Obama's website versus 21,021 on McCain's). If you take speeches by both candidates out of the mix, Obama's site still features more words than McCain's, but the counts are closer.
• The McCain campaign has fully integrated his vice presidential pick, Sarah Palin -- both textually and visually -- into the website's home page, while the Obama home page displays his vice presidential pick, Joe Biden, much less prominently.

One of the peculiar differences between the two websites is the difference in the number of viewers they sent to campaign videos on YouTube. YouTube reports the views coming from the top five sources of views. Six million of the views of Obama campaign videos came from the Obama websites. That was two-thirds of the views from the top five sources. It is a small percentage of the 42 million views, however. By contrast the McCain website was the source of only 55,000 views as one of the top five sites. That was 2.2% of the total views coming from the top five sources, and was a negligble proportion of the total 20 million views. The website sending the most views to the McCain campaign videos was The Huffington Post, a liberal blog. One might suppose it was not sending McCain supporters.

There was much activity on YouTube in addition to the postings of the campaigns. "Yes we can" by Will.i.am, an early video supporting Obama, was viewed more than 20 million times (Wallsten). The Obama Girl's love song "I got a crush . . . on Obama" was viewed 13 million times, and Barely Political, the channel/production company of Obama Girl, posted 187 political spoofs between June and November. The pro-McCain video "Dear Mr. Obama" was viewed more than 13 million times. And there were tens of thousands of videos posted to YouTube by people making their own versions of support and opposition.

The 'traditional' media were also active on the web. One ranking of the websites most frequently viewed on election day shows the mixture of traditional and new organizations the viewers were consulting (Schonfeld).

TV networks are well represented with CNN, MSNBC, Fox News and four more for eight out of the top 27. Yahoo News, at number one, Google News, at number four, and AOL News are web services ranking high in the listing. Five major newspaper websites are on the list. But the most numerous is blogging websites with 11. Traditional media and new media organizations are equally represented in the list of most viewed on election day.

Those are the big numbers, but there is another side of the web that is also important. Chris Anderson calls it the phenomenon of less is more. The long tail is one of the predominant stories about the web. While there were precursors, both the identifying phrase and its clearest articulation begins with Chris Anderson, the editor of Wired Magazine, in a 2004 piece entitled "The Long Tail." He subsequently published the book The Long Tail: why the future of business is selling less of more (Anderson, 2006) and started a blog The Long Tail. A post on the blog in 2008 makes clear how his conception differs from some others.

Anderson makes an argument about the transformations being wrought by changes in the technologies of information and distribution. He specifically analyzes the entertainment industry where he argues the movement is from an economy of scarcity to an economy of abundance. He contrasts the differences for movies, music, and books between a culture and economy based on selling more of less or hit movies, hit music, hit books and the emerging culture of "selling less of more." A book store, a music store, a video store draw on limited audiences that are geographically bounded. The stores must sell some number of each book stocked each year to pay the bills, for example. Books that sell fewer than that minimum cannot be carried by the bookstore. Even a giant marketer like Walmart must sell at least 100,000 copies of a CD each year to pay the bills, he wrote. What about the CD that would sell 60,000 if it had been stocked by the Walmart store? It will not be sold. But for Amazon with its reach around the world selling the CD that will sell only 60,000 copies is as cost effective as selling the CD that sells a million copies. Each CD that is sold costs the same: the costs of information about availability and the distribution costs are the same. The long tail is the many CDs that will sell 60,000 copies if carried but will not sell in the standard local store arrangement. This becomes even more apparent when music is transformed from CDs and DVDs to files downloaded from a computer as with Real Network's Rhapsody or iTunes. Rhapsody stocks, where stocking is a database on computers, more than a million titles as of 2008 and every one is accessed every year (Anderson, 2008). That is the long tail that Anderson argues reaches small groups with similar interests who cannot be served by the technologies of the past. And he argues there is more money in the long tail than in the head. In this new economy no one is limited to hits; abundance fosters diversity of tastes.

The revolutions in information and distribution are not limited to the entertainment industry. The web is turning over many industries -- including politics and campaigning. To carry Anderson's ideas to other domains two elements of the theory need to be carefully respecified.

Obama Videos Most to Least Views

One is the idea of head and tail. One way this might be handled is in terms of a distribution. The distribution of views of the Obama videos in the figure certainly looks like a long tail. The range is from almost 1.8 million for the most viewed video to eleven hundred for the least viewed video. The figure shows that most videos received many fewer than 200,000 views. If you were doing mathematics you might happily divide head and tail by the inflection point, which is the point where the direction of the curve decisively changes. That would be somewhere around 100,000 views. The videos that were viewed more than 100,000 times would be the head and the rest would be the tail. Or you could do the Pareto 80-20 rule; 20% of the videos have 80% of the views. In this case the 20% most viewed videos would be the head and the rest would be the tail. Both of these are ways to define head and tail in terms of the distribution. But Anderson uses a more substantive definition of head and tail.

This is a good moment to remind everyone of the normal definition of "head" and "tail" in entertainment markets such as music. "Head" is the selection available in the largest bricks-and-mortar retailer in the market (that would be Wal-Mart in this case). "Tail" is everything else, most of which is only available online, where there is unlimited shelf space. (Anderson, 2008)

'Head' is defined by one technology of information and distribution and 'tail' is whatever distribution there is beyond that technology. He does not claim that the old technology will be completely replaced by the new technology, though it may over a long enough time. Barnes and Noble still sells books in many towns and cities even as Amazon sells books that the local Barnes and Noble sells and that the local Barnes and Noble cannot carry because the sales are not sufficient to justify the shelf space. In the transition the new technology supplements the old. And the definition of head and tail does not change. The head remains the "selection available in the largest bricks-and-mortar retailer," that is, the old technology.

The second idea is serving previously unserved markets. There are people interested in a book or subject of books, for example, who are too few in number in any geographic area for the bookstore to carry the books they would like to purchase. However, they would be a profitable market if they could be reached by technologies for which distance is no longer as important. The long tail is not just selling more books, or videos, or whatever product. 'Hits' are in the stores. More could be sold if there were more people interested in them. It is the books, music, videos that are not in the stores that are of interest to the 'niche' that can only be assembled via a different technology. And the result is

Long Tail business can treat consumers as individuals, offering mass customization as an alternative to mass-market fare . . . And the cultural benefit of all of this is much more diversity, reversing the blanding effects of a century of distribution scarcity and ending the tyranny of the hit. (Anderson, 2004)

The result is a limited de-massification of culture.

Head and Tail

Between July 1 and the election in November the McCain campaign posted just over 100 videos on YouTube and the Obama campaign posted approximately 760 videos on YouTube. And the viewers came: the videos of the Obama campaign were viewed 43 million times; the videos of the McCain campaign 20 million times. The distribution of views of the McCain videos takes much the same form as the figure for the Obama videos above; it also has the appearance of the long tail. While communicating with voters via TV ads did not stop there was clearly a transformation of communication in the election.

Anderson's ideas seem very appropriate for understanding the transformation that political campaigning is undergoing. There is an old technology for campaigning and a newly emerging technology. This is particularly appropriate given the contrast between the video work of the two campaigns. Most of the McCain campaign videos were made for TV and then posted to YouTube. There were exceptions; they posted a half dozen radio ads, and about the same number of virtual town meetings that seemed too long to run on TV. And at the end of the campaign they posted several videos that were not run on TV. However, most appeared to be made for TV. The Obama campaign also posted made for TV videos, but those were overwhelmed by the videos made specifically for YouTube. One campaign continued the 'old' technology even in the new technology, and the other adopted the new technology as an important element in its campaign.

To approximate the head-tail distinction in the new medium of YouTube head will be specified by the number of TV ads that a campaign would likely run; that number of the most frequently viewed videos will constitute the head for this analysis. The tail is the videos beyond the number of videos that a campaign might well run. In 2004 the Kerry campaign made and broadcast approximately 130 TV ads. In the 2008 campaign the McCain campaign posted 104 videos, most made as TV ads, to YouTube. The Obama campaign produced head and tail; they posted 761 videos to YouTube during the campaign. These numbers are a simplification of the activity of the campaigns. Both campaigns posted videos to YouTube that they then took down. These are the numbers that were there at the date of the election.

The comparisons in this table are views received by Obama campaign head and tail videos defined as: 130 as the head with reference to the Kerry campaign and 104 as the head with reference to the McCain campaign.

If the Obama campaign had stopped with the 130 most viewed of their videos they would have missed 14 million views, and if they had stopped with the top 104 videos they would have missed almost 16 million views. By posting the additional videos they extended their reach by one-third, which is an impressive campaign payoff.

However, it is important to note that videos constituting the head had many more views on average than did the videos in the tail. Using the estimate based on the Kerry campaign, the top 130 videos had an average of 222,207 views, and the 631 videos constituting the tail averaged only 22,408 views per video. The comparison using 104 as the head is an average of 261,854 views per video and an average of 24,038 views for videos in the tail. It took a lot of videos to get the extra third of the views the Obama campaign received.

For the Obama campaign the head was an impressive two-thirds of the total views, and the tail was long with 630+ videos and a third of the views.

Audiences

The numbers are suggestive. The distribution of views of the videos of the Obama campaign seem to be what one would expect from Anderson's articulation of 'the long tail.' But the long tail is not just about the numbers. It is about using new technology to reach out to audiences that otherwise could not be reached effectively. So, what were the audiences?

David Plouffe, the Obama campaign manager, made videos about the campaign in July, September and October to discuss campaign strategy, to thank volunteers for their contribution to the campaign, and to ask for continuing help. This is taken from his July 14 video

This campaign has been built on the backs of millions of Americans who believe in change and want to take their country back. The reason we have gotten this far -- winning a primary campaign when most of the establishment was on the other side and being in such a strong position as we head into the closing weeks of the general election is because of your hard work. The $25 or $50 you have been able to give us, the time you have spent knocking on doors and making phone calls. We've never seen a campaign at the presidential level so dedicated to belief in the strength of the grass roots, and our own success dictated by your support . . . We are only here because of your hard work and support. The great thing about this campaign is that we learn from you every day. You are out there making this campaign happen. We are going to have to rely on you each and every day to own this campaign, to make sure that our campaign for change is successful on November fourth.

In July the campaign was reaching out to local: local organizations, local individuals who had helped and who they needed to help again. Plouffe pointed out states that had voted for Bush in 2004 that had strong local organizations in 2008. It is a campaign for change, he said. Otherwise it was the appeal of "We are only here because of your hard work and support." And we need more.

In Anderson's theorizing the audience that is reached by the long tail is people who want to purchase a book, a song, a film that is not available via the head, local stores. In the Obama campaign the audience was not people wanting to purchase it was people wanting to support Obama but who could not be effectively reached via the head, TV advertising. In neither case is the long tail about creating new wants it is satisfying wants that already exist. It is about facilitating people in doing what they want to do.

And this is the puzzle. How do you go local in a global domain? YouTube is a worldwide medium. Viewers did come from everywhere on the planet. What do you do in such a communication domain to foster local organization, local effort? Plouffe was pretty clear about wanting to promote local effort. But it would take more than his video to generate the local effort. His July 14 video was viewed only 46,000 times by November 4.

One thing they did was to explicitly address local audiences in the way they gave titles to the videos they were posting: Barack Obama in Zanesville, Ohio, in Fairfax, Virginia, in Powder Springs, Georgia, and on and on. Twenty-seven, 43%, of the videos posted in July were Barack Obama in . . . Notice what this means. If you went to the Obama campaign channel on YouTube in July almost half of the videos tell you nothing about what is included in the video. All you know is that Obama visited Zanesville, Ohio. If you are from Zanesville you might well want to look at the video to see if you can find yourself or friends in the crowd. Why would someone from Iowa want look at the video? Apparently they did not. Barack Obama in Zanesville, Ohio was viewed only 15,500 times between July 1 and November 4. But there were 27 towns that Obama visited that made it to YouTube. And Zanesville, Ohio is just the 'local' that otherwise the Obama campaign could not reach with the technology of the 'head.' Even if you did not find your town among the 27 you still learned that local was in for the Obama YouTube campaign.

Geographic references did not decline as the campaign progressed, but the emphasis shifted.

The geographic references are almost constant: fifty percent plus or minus a few percentage points. The exception is August when more than 30 of the videos posted to YouTube were from the Democratic National Convention.

How 'town' and 'state' were counted for the table needs some explanation. Every reference to a town in a video title was counted as town and appears in the Town column. Many of those references to a town also indicated the state in which it was located. However, those were not included in the State column. The videos counted for the State column were videos that only mentioned the state.

Town is 42.9% and state only 11.1% in July, but the first week of November town is 15.6% and state is 36.0% Why the steady progression from town to state as the campaign progressed? As they moved into September and October registration and get out the vote became more important, and that is a state matter. Michigan will have early registration beginning October 6; a video about Michigan. Vote early in New Mexico beginning October 29; a video about New Mexico. It is a move from one 'local' to another; from towns to states. But in neither case would the videos in the 'head' accomplish this local communication.

Approximately fifty percent of the Obama campaign videos had the name of a town or state or both in the title. That is a very strong local focus. What were the views for those videos? The mean number of views per video was 22,717 putting them right in the middle of the long tail. If the audience they were trying to reach was a national audience that would be a disappointing number. But they were not aiming at a national audience. They were aiming at local audiences, at audiences they could not select for attention with a video of national focus. They were, as Plouffe said, intent on fostering local organizations scattered throughout the states with these videos. This seems almost exactly what Anderson was describing with long tail.

What about the McCain campaign? They ran a national campaign. Only 9 of the 104 videos mentioned a town or state. Their videos were not designed to reach out to small groups that were distinct from a national audience. There was no long tail even though some of their videos were viewed millions of times and some only tens of thousands.

Another pass at "mass customization"

David Plouffe began his July 14 video by pointing to the 2004 presidential election map of red states and blue states and doing a bit of electoral college politicking.

He did not have to explain what red and blue states are; all 46,000 viewers knew that. The objective, he said, was to hold the states Kerry had won in 2004 and to win some of the states Bush carried. Red and blue states are the result of the election. The electoral politics of a campaign oriented toward the electoral college is safe states and swing states. The red and blue map gives no hint about safe and swing states.

Reaching out to towns to foster the strength of local organizations and to states to promote registration and voting are not necessarily at odds with also reaching out to some states rather than others for building a majority of electoral votes.

Almost half of the Obama campaign videos identified a town or state or both in their title. Aggregating the videos identifying a locality by safe and swing states produces the following distribution.

The pattern is very pronounced. Safe red is defined as Bush receiving 54% or more of the vote in the state. Safe blue is defined as Kerry receiving 54% or more of the vote in the state. There were 23 safe red states and another 8 states in which a majority of the votes were given to Bush, but the Bush vote was less than 54%. There were 12 safe blue states and 7 that gave a majority of their votes to Kerry but not as many as 54%. Where did the videos go? Overwhelmingly they went to swing red states; 215 out of 398 had a location of one of the 8 red swing states in the title of the video. The next largest group was videos that had a locality from a blue swing state in the title; 81 for this set of states. Safe reds came in third, but the summary across all 23 states does not reveal how two sided this was. The Obama campaign took two 'flyers': Indiana and North Carolina. Bush received 59.9% of the vote in Indiana in 2004 and 56.1% of the vote in North Carolina. Even though they were 'safe red' the Obama campaign posted 14 videos with an Indiana location in the title and 25 videos with a North Carolina location in the title. Thirty-nine of the 55 videos went into two states. The other 16 were distributed between the other 21 safe red states. Pity the poor safe blue states that were 'recognized' with less than 1 video per state. Which safe blue states had no videos? California, Connecticut, Maryland, Massachusetts, and New York were the big ones. Five of the videos posted for safe blue states had Illinois locations in the title, and 2 had Delaware in their titles. The candidates posted at least a few from home.

Given the electoral college every campaign runs a national campaign and state campaigns. It is not unusual that the Obama campaign ran a campaign that could be 'mass customized' to reach out to a number of states. While there is some state focus when campaigns are planning their TV ad campaign it is the case that most TV ads are addressed to a national audience. They are just shown more often in some states than in others. But the Obama campaign was putting half of their videos into this effort. Half of their videos were addressed to local audiences rather than to a national audience. Most of us, for example, are not really interested in the date for early registration in Michigan. That is not aimed at a national audience. Focus on local that was impractical in the past had become practical by 2008.

Strategy and Training

There are two more forms of reaching out to an audience that could not be reached with the 'head' or normal TV ads. Neither is very large in number, but they were an integral part of the campaign of reaching out to the tail.

David Plouffe made four videos talking the strategy of the campaign: July 13, September 17, October 7, and October 14. In each he sat at his desk and computer and talked to viewers about how the campaign was progressing and what they needed to do next. In each case it was an appeal for assistance from viewers. The videos of July and September were viewed 45,000 and 41,000 times or almost twice the average number for the long tail. The two in October were each viewed about 25,000 times, which was very close to the average for the long tail videos. They are not big numbers, but they are directed to a very specific and small audience -- the most committed of supporters. They are the people who want to be 'on the inside' and who are very willing to donate time and dollars.

In addition to Plouffe's strategy videos they also posted videos about the strategy in individual states. These were videos aimed at swing states. They were all posted in October as the leaders in 8 states talked about the thrust the campaign in their states during the last days.

Another type of video reaching out to an audience not reachable with television advertising is instructional. myBO was the Obama website for volunteers. Neighbor to Neighbor was the organizational mantra for volunteers. "myBO: Neighbor to Neighbor" was posted to YouTube September 8 and was viewed 110,000 times. And "Neighbor to Neighbor How-To" was posted on October 1, and viewed 43,000 times. It was very straightforward instruction. Step 1, step 2, step . . . It was instruction and social reinforcement for people who had not participated in canvassing before. And the Obama campaign posted a CNN news video about Neighbor to Neighbor attracting 110,000 views. The 'instructional video' watched most frequently was "Jill Biden Phonebanking at Obama/Biden HQ" with 225,000 views, which was Jill Biden talking about and calling in 'living color.' There were videos of opening campaign offices and one about talking to your parents, "The Talk: Convince Your Family to Vote Obama," that was viewed 82,000 times. Collectively these instructional videos were viewed 790,000 times. That is an impressive reach for instruction and social reinforcement.

Conclusion

The limits on communication fade with the growth of the web as a medium of communication. And campaigns move from an 'economy' of scarcity to one of plenty. And the long tail, as specified by Anderson, becomes feasible. The Obama campaign took full advantage of the opportunities presented by this new medium and economy of plenty. The McCain campaign worked within the tradition of television and adapted only to the extent that they also posted their made for TV videos on YouTube. In politics nothing is copied more assiduously than a winning strategy. It seems very likely that the long tail is the future of presidential elections.

References

Chris Albrect, "Nielsen: Total Video Streams top 10 Billion" http://newteevee.com/2009/02/26/nielsen-total-video-streams-top-10-billion

Chris Anderson, "The Long Tail," Wired, October, 2004

Chris Anderson (2006) The Long Tail: why the future of business is selling less of more, Hyperion.

Chris Anderson, "Excellent HRB piece challenging the Long Tail," June 27, 2008, The Long Tail, Wired Blog Network

Tim Arango, "Broadcast TV Faces Struggle to Stay Viable," http://www.nytimes.com/2009/02/28/business/media/28network.html?_r=2

comScore Releases January 2009 U.S. Search Engine Rankings, http://www.comscore.com/press/release.asp?press=2729

comScore YouTube Surpasses 100 Million U.S. Viewers for the First Time, http://www.comscore.com/press/release.asp?press=2741

Howard Kurtz, Under Weight of Its Mistakes, Newspaper Industry Staggers, http://www.washingtonpost.com/wp-dyn/content/article/2009/02/28/AR2009022801964.html

Adam Ostrow, 82 Million User-Generated Content Creators and Counting, http://mashable.com/2009/02/19/user-generated-content-growth/

Pew Research Center Project for Excellence in Journalism, "JohnMcCain.com v. BarackObama.com" September 15, 2008

Eric Schonfeld, Hitwise Ranks Election Traffic to News Sites, http://www.washingtonpost.com/wp-dyn/content/article/2008/11/05/AR2008110504025.html

Emily Stone, "While Obama Speaks, Congress Tweets," http://www.myfoxdc.com/dpp/news/022509_congress_twitter_obama_speech

Kevin Wallsten (forthcoming) "Yes We Can": How Online Viewership, Blog Discussion, Campaign Statements and Mainstream Media Coverage Produced a Viral Video Phenomenon" Journal of Information Technology and Politics

Wikipedia, Blog, http://en.wikipedia.org/wiki/Blogging

Wikipedia, Flickr, http://en.wikipedia.org/wiki/Flickr

Wikipedia, Google, http://en.wikipedia.org/wiki/Google

Wikipedia, YouTube, http://en.wikipedia.org/wiki/Youtube

@@@@@@@@@@@@

KDPaine and Partners listing of campaign videos and views http://www.measuresofsuccess.com/Default.aspx

http://nav.diydashboard.com/landingclassic.aspx?diylogin=PEFUX/TFNJw8=P&s=mrrtxfymanbnjn45uwwy4i45