Democracy in Media

The Necessary Head of Long Tail Distribution

Chris Anderson has put up a couple of excellent posts recently (big surprise there). The first had to do with some poor guy’s misinterpretation of the long tail phenomenon, the second to do with tribal culture. What made the posts so compelling was not Chris’ clarification of what the long tail means, but rather his insight as to how consumers locate the signal amid all of the noise and how signal identification is a complex and often unintuitive process. 

We have traditionally been defined by our commonality. And while commonality is still important (we are social creatures after all), we are beginning to find commonality with very specific groups on an ad hoc basis. Online, I can be friends with someone due to a single commonality and interact with them only in relation to that interest. I can compartmentalize my relationships very effectively. 

Anyway, this has specific implications for the efficient location of long tail video. Niche content is hard to find without the help of those that have already blazed the path to it. But finding the ideal trailblazers can also be challenging. Because it’s not just in locating the individual with a common interest, it’s in identifying whether you trust that individual’s interpretation of relevance for that specific topic. He may be a master in his field, but his tastes and interests may still diverge dramatically from your own. You can have a common interest while also having a difference of opinion.

Chris sums it up nicely:

“Here's my take on what the Long Tail is doing to pop culture. Rather than the scary fragmentation of our society into a nation of disconnected people doing their own thing, I think we're reforming into thousands of cultural tribes, connected less by geographic proximity and workplace chatter than by shared interests. Whether we think of it this way or not, each of us belongs to many different tribes simultaneously, often overlapping (geek culture and Lego), often not (tennis and punk-funk). 

What's interesting is that the same Long Tail forces and technologies that are leading to an explosion of variety and abundant choice in the content we consume are also helping to connect us to other consumers, whether through Amazon and Netflix reviews, blogs, p2p networks or playlist sharing. I've described this in the past, somewhat obscurely, as the rise of orthogonal trust networks, which are the new recommendation and word-of-mouth effects that will drive demand down the tail from hits to niches.”

Another quote:

“A tail without a head is too noisy and apparently random to get consumer traction; people need to start with the familiar and then move, via trusted recommendations, to the unfamiliar. Likewise a head without a tail is too limited in choice; the odds of finding a niche you want are too low to bother exploring much beyond what you already know.”

It is critical to start any relationship with some commonality. As it relates to locating content in the long tail, I need to have watched some of the same content and found some commonality in our views of that content. I don’t need to become the guy’s best friend. These relationships can be highly segmented and ad hoc, but without a sense of the individual’s “track record” with common experiences, it is impossible to form a trust relationship for the journey down the tail.

This is the problem I see with many long tail distribution services out there (especially in video). While each has the goal of providing a simple mechanisms for publishing media, few seem to have given much thought to how published content will be located by consumers. They seem to believe that long tail video has such inherently high value that people will spend the time to find those golden nuggets; they won’t (or at least most won’t). The whole point of long tail video is that a vast majority of it will be completely irrelevant, but some of it will also be highly relevant. But the simplicity of finding that stuff defines the ultimate utility of the network, not the simple aggregation of the assets.

In most services there is no commonality, there is no community. RSS is nice, but it doesn’t make any decisions about what’s delivered, it just tells you that it’s there. Tagging is certainly a methodology that can yield interesting results, but only once the network achieves a certain scale. Getting to that scale requires people to find the service valuable. Search is also a good tool, but it still requires scale in the network to deliver real value. Try finding really great video in many of these sites and it will take some time. It’s there (in many cases); it’s just not very easy to access. 

The problem is that these networks not only lack the trailblazers, they lack any structure in the terrain. They start with no landmarks of any kind. Until these services begin to give people some reference points around which communities and conversations can form, they will have trouble reaching the critical mass that is a prerequisite to a vibrant long tail marketplace.   Otherwise, it just a waste of our time and attention.

Tags | longtail , rss , tags , chrisanderson , community

Why This Man Isn’t Paying Attention

Why does he assume that a small number of media conglomerates have been good at coming up with unique and interesting content. If the rest of us just have to overcome the “extraordinary creativity” of Clear Channel and Infinity Broadcasting I don’t think the bar has been set too high. I think we’ll manage.

Anyway, the point is not that these podcasts have to appeal to lots of people; they just have to appeal to some (or many even one). The majors will have an impact when they enter the market and a lot of traffic will gravitate to their content, but some won’t. Time and attention is shifting towards the tail. Just because it doesn’t occupy that much mind share today, doesn’t mean it won’t in the future.

Tags: longtail , podcasting

The “Long Tail” is Already Being Televised

There is a lot of niche content on TV (read: cable or satellite). Tons of it. But everywhere I read about TV and the long tail phenomenon, TV is automatically categorized as “mass media.” Well, if anyone has ever watched grainy live shuttle footage or late night Chinese news broadcasts, you know that TV has already drifted quite a ways down the long tail. It might not be the Internet of Video, but it also isn’t just a fat head.

Now, this seems kind of strange to me. I’ve always considered cable and satellite to be a pretty expensive way of moving video around. And not just because all that infrastructure is so expensive (it is), but because of the opportunity cost of all those customers. I mean Comcast has 21 millions subscribers. That’s a lot of people looking at your aggregated content (8 hours a day on average, which is also crazy). It seems kind of a waste to be occupying valuable time and channel slots for “Jade Starbiz.” (it’s in Cantonese, so I’m not sure what the show’s really about.) I mean, if Time Warner is right in describing the evils of a la carte cable, these niche networks don’t even pay their own way, but are instead subsidized by the higher traffic networks.  It would seem that TV, in its current form, should have more pressing scarcity issues.

But what really creates scarcity on TV is not bandwidth limitations, it’s:

1. Time limitations imposed by broadcast schedules: certain hours of the day, like 5PM to 11PM EST, are ideal viewing hours. Those hours are worth far more than those available during 3AM to 6AM EST, on almost every network (demand simply outpaces supply).
2. Consumer inability to easily find shows that he or she might like: time/network grids essentially suck as a way to find shows that might be interesting, leading you to gravitate towards certain channels that most frequently provide you with shows you like. 

So, in effect, certain channels have issues with scarcity, while others have bandwidth to burn. Now, TiVo and similar devices are hot on the trail of the first problem. This should eventually alleviate the inherent time scarcity within high-demand channels, but the larger problem of finding content perpetuates the false scarcity that exists horizontally throughout the entire channel lineup. This would seem the even larger issue. But even with TiVo's recommendations and Wish Lists, Yahoo!’s TV Guide and TMS data, finding great content on TV outside mainstream channels is still very difficult. It would seem that cable and satellite carriers have an inherent interest in maximizing the value of their entire spectrum, but it would appear that this interest has its limits. (This even applies to VOD where the economics should be most compelling for Comcast. Yet Comcast seems to have mastered the art of delivering what you want, when you want, without actually helping you find it.)

The worst part about this whole situation for the consumer is not that they’re paying too much for content (although they might be), but rather that they’re missing lots of content that would make their TV experience much more compelling (and far more personal). So, rather than settling down for the third rerun of some sitcom on ABC, they might find something really compelling on a channel they’ve never watched before. Combine this with TiVo time-shifting capabilities, and, all of a sudden, TV is a very different experience. A much better experience.

I’m a firm believer that TV will look radically different a decade hence. Brand new networks (if they’re still called that) will be serving thousands of niche audience communities through the magic of inexpensive IP-based digital distribution. But it’s not going to happen tomorrow. We’re talking about a massive change in people’s viewing habits that will evolve over years, not months. These people need a bridge to this long tail future. I’d be very surprised if broadcast TV (with a few evolutionary enhancements) didn’t span a significant measure of this chasm.

Tags: LongTail, IPTV, TiVo

Time and Attention Costs of the “Long Tail” and Pareto’s Bulging Neck

Kevin Laws over at Venture Blog wrote a great piece on the non-monetary costs of long tail transactions and how search and psychic costs specifically dominate when the price point of a niche asset drops below certain thresholds. While I appreciate the effect of both of these costs, it would appear that psychic costs are easier to minimize in long tail transactions through various billing mechanisms. However, search costs (or as I call it, time and attention costs) are a much tougher nut to crack.

In thinking about these issues, I had a conversation yesterday with David Sabel over at www.Upto11.net about context and its importance in minimizing time and attention investments in these transactions. David’s basic point is that while several technologies exist to minimize these costs (such as collaborative filters and folksonomies), time and attention expenditures remain high enough to render much of the long tail essentially unreachable (at least today).

I’ve spent a lot of time trying to understand how to solve these issues for CozmoTV, a product we’re planning to release this summer. Initially, CozmoTV will focus primarily on simplifying access to mass-market content, but our business model depends upon our ability to create a vibrant marketplace for niche video. And, while I firmly believe that consumer contributed content will (eventually) become an enormous component of this marketplace, resolving our target audience’s time and attention limitations remains our most pressing challenge in achieving this goal.

To understand the issue more clearly, we can look at the blog community. If you believe Technorati’s (somewhat inflated) number, somewhere around 8 million blogs are up and running today with somewhere close to 80,000 new blogs being contributed to this number every day. While blogs are still a relatively new phenomenon, the marketplace has gained enough mass to serve as a valid demonstration some of the inherent limitations of long tail content. I’ll use my own experience to illustrate my point. 

I’ve been spending a lot of time reading blogs recently. I have around 40 blogs that I check on daily and another 40 (or so) that I check on weekly (give or take). This is an enormously time intensive exercise. And, even with tools like Feedburner, Technorati and Del.icio.us at my disposal, the information yielded from my time investment is sub-optimal. It’s still (significantly) better than what I would get from watching CNN and reading the NYT, but we’re still talking about a relatively inefficient process of information consumption.

Now, my feed roll is highly targeted to my interests, it’s on-demand, and it’s up-to-the-minute. These are all dramatic improvements over where I was before, but the fact is I’m still tending to gravitate towards high-traffic sources (they just happen to be “A-List” blogs rather than traditional MSM sources). This has to do with many factors (including the volume of postings, the quality of the material, the credibility of the sources, etc...), but two of the primary reasons I’m gravitating to a smaller number of sources is that:

1. I don’t have time to read any more than I currently do, and
2. finding new relevant blogs just exacerbates point a).

Now, I’m sure that every day there are dozens of other postings out there that would be even more interesting and relevant to me than what I’m reading now, but most of this content probably doesn’t reside within “A-List” blogs on any given day and is therefore pretty difficult to locate (or even if it does, it operates outside the A-list community in which I’m currently spending most of my time). So, I’m faced with a basic question, do I commit more time to looking for these ideal posts, or do I settle for posts that are less ideal but more easily accessible. Well, I’ve already answered that question.

Anyway, the point is that time and attention limitations are the primary threats to the success of a truly democratic media movement, a real meritocracy. In an ideal world, the best content could filter up from the bottom and reach intended audiences with extreme efficiency and speed regardless of its point of origin. And, while this happens far more frequently today than even 3 years ago, this process is still incredibly inefficient and unreliable. 

The fact is (as I said in an earlier post) the tyranny of a few guys in Hollywood deciding what you can watch threatens to be replaced by the tyranny of popular opinion (in which some amalgamated list of preferences decides what’s most relevant, but not necessarily to you.) This may result in a slightly better selection of higher quality mid-market content, but it is a far cry from the goal of a truly level playing field in which an individual piece of content is evaluated on its individual merits and not necessarily its proximity to other high-traffic assets.

Video is even more challenging when it comes to these limitations than blogs. Unlike text that can be easily skimmed, video relies much more heavily upon titles, descriptions, tags, and other metadata supplied by the author or the network (although closed captioning will help). Also, as opposed to just being able to scan the article in a few seconds, the file must be downloaded and watched. This whole process of finding and consuming video is orders of magnitude less efficient than blogs (searching is more difficult and less accurate and consumption of the content is far more time consuming).

The bottom line: getting consumers to venture down the long tail is very, very difficult (especially with audio and video).

I tend to agree with David that the solution to this problem lies in taking a multi-pronged approach to the issue (using collaborative filters, tagging, etc.), while also maximizing the concept of community in whatever content network you develop. I think Upto11 is doing a nice job of this for audio. Providing for the ability to develop ad-hoc communities around specific niche content is critical to driving content to its ideal audience. But even these communities threaten to buckle under the weight of their own success (one extreme) and limit their utility through overwrought stratification (the other extreme).

I’m not yet sure what the final answer is, but I am sure that relieving these non-monetary costs is critical to truly delivering on the promise of long tail content. Without it, we’re still going to wind up watching a lot of the same crap as everyone else, it’ll just be somewhere in the bulging neck of Pareto’s curve rather than the (already fat) head.

Tags: longtail, weblogs

Can Video Diamonds be Found in a Sea of Coal?

A recent posting at Buzz Machine has me thinking a bit more about the difficulty in locating multi-media content that is relevant, quickly and easily. The Web of today is largely a text-based phenomenon. This is so for many reasons, but as the quantity of content has exploded, so has the importance of advanced search tools that efficiently sift the irrelevant from the relevant. These tools are getting better, but even so, finding that ideal nugget of text that is relevant to your query involves scanning half a dozen documents (or more). While this is irritating, it almost always yields a satisfactory result within 2-10 minutes of effort. (It might even lead to some interesting tangents along the way.) This process is only improving over time as search technologies improve, but what about audio and video?

 The problem with searching for a specific nugget of audio or video is that this scanning process is much less efficient. First of all, many current A/V search efforts scour metadata like “title” and “description” as opposed to the content itself (imagine searching for specific chunk of text that meets your requirements if all you had to search were the author’s contributed title and short description… inefficient to say the least). But let’s assume that this problem is resolved rather quickly and that Google, Yahoo and MS are all successful in penetrating the content to revel a text-based transcript. (While closed captioning, voice recognition, and the like make this possible, it will certainly take some time to perfect, especially once you begin to penetrate the unstructured world of content beyond CNN, ABC and HBO.) But even with these transcripts, searching through A/V content would involve scanning transcripts, identifying items of potential interest, downloading/streaming the feed, and essentially repeating the scan in a different medium. This is a rather complex and involved process to say the least...

 “Long tail” content only emerges from obscurity if interested audiences can locate the content to begin with. If more of this process cannot be automated, the risk is that the tail will be cut short and that a larger and larger segment of information will effectively disappear; searchable, but unsearched due to the limitations of human bandwidth. Does this mean that while text-based communications are less efficient that audio and video communications, written text will remain dominant due to the inherent difficulty of efficiently scanning richer mediums?