October 1, 2008

Bullpen: A Trick of the Tail

By Michael Adams

There has been a lot of discussion about the "long tail," a term first coined by Chris Anderson, and described at length in his book The Long Tail - Why the Future of Business is Selling Less of More.

The concept of the long tail is that there is a natural distribution of popularity of items so that in most retail models the focus is on selling just the most popular items. After the first 10 or so choices, individual items don't sell well enough to justify shelf space (for example, in a supermarket) and can be ignored. This has developed into an obsession with brand recognition and by blockbuster movies-where a fairly narrow choice is marketed and sold to "satisfy" the market demand. Outside of the mainstream of hits, niche markets are typically left to specialty retailers, and demand is small, driven by word of mouth.

Catalog size

Amazon changed this model for retail by creating a digital catalog where the number of items could essentially be unlimited in size and turned the conventional rules of retail upside down. The long tail theory says that as choice increases, people will tend to express their natural diversity by making more individual, personalized choices. Essentially, if there are enough items in the tail of the distribution, the long tail can generate a more significant piece of the total pie and can no longer be ignored.

This doesn't mean that blockbuster hits don't continue to drive a disproportionate share of the revenue, but it does mean that there is an opportunity in the less popular titles just because there are so many of them that they can add up to be important. And the tail can be very long - Anderson references various examples, such as subscription music services, where choices way down the popularity scale still get played at least once a month.

The interesting point here is how the size of the library influences the buying behavior. The long tail phenomenon shows itself most clearly when the library gets large enough that the normal 80:20 rule breaks down, when a significant number of purchases come from the tail rather than the head of the queue. The poster boy for this is Amazon, which has demonstrated that having a very large catalog helps to generate sales. The equivalent play in the music business is iTunes, with 8 million songs available.

iTunes and Netflix are emerging as potential examples of the long tail phenomenon in the "over-the-top" video space. Both of these content aggregators could offer a larger on-demand choice than a typical cable operator's on-demand offering. iTunes has sold 5 billion songs from a possible 8 million songs and claims its customers are renting/purchasing 50,000 movies per day. Netflix has more than 100,000 titles with currently about 10,000 available for "instant access" according to its Web site.

Pseudo set-tops

Until recently, there was no direct comparison to cable video on demand (VOD) service since Netflix ships DVDs, and iTunes downloads to PCs and iPODs. But now both have introduced set-top-like devices to drive the TV set when connected to a broadband Internet connection. These are the AppleTV, which was launched at the same time as the iPhone and almost forgotten over a year ago, and the Roku box, which was introduced earlier this year to Netflix customers at $99.

I have tried both devices, and I will focus on my personal experience of them (which may not be representative of the average viewer).

First the AppleTV. The first version really just synched to an iTunes library and therefore required a two-step ordering process. First buy a title on iTunes on your PC, then synchronize it to your AppleTV, and then play it on your TV set. This wasn't very satisfying, and soon the AppleTV was gathering dust. (Actually, it got unplugged because it developed a nasty habit of switching itself on randomly.)

When someone told me the software had been upgraded to 2.0 and that it was a much easier process to actually find and watch video, I plugged it back in and gave it another try. Sure enough, the menus are well-designed and, if you know what you want to watch, it is actually not a bad deal at $1.99 an episode for TV shows and $3.99 for a movie rental ($4.99 in HD). However, there's the rub - if you know what you want to watch. Maybe I'm just cheap, but I purchased only one complete series (The Tudors season 1). Somehow the thought of spending hard-earned cash on something I might not enjoy is a significant barrier.

I think there is also the factor of browsing; the only way to browse on AppleTV is to preview an item for 30 seconds - not long enough to make a viable viewing decision for me. It is not the same as watching something with no investment and deciding to continue watching purely on a subscription basis.

I wasn't able to find out exactly how many titles are on iTunes, but a quick check through the "All Movies" category generated more than 2,000 results. In addition, there are a lot of TV series. Apple is also providing some movies in HD format-using 720p vertical resolution.

While the iTunes model completely a la carte, Netflix is at the other end of the spectrum with a completely subscription-based service. In fact, its core service is to ship DVDs, and streaming video "over-the-top" via a broadband connection is a recent addition. Netflix grew to a million subscribers in less than four years and now has 8.4 million subscribers.

Streaming to the TV set requires the purchase of a $99 box (made by Roku). The service is standard definition. Ordering is a two-stage process - you have to go to the Netflix site and add movies or TV series to your "instant queue" - so it is familiar to Netflix customers. This is actually a great model for parents, allowing them to pre-select the titles that they consider suitable for their children to watch.

Once a title is added to the instant queue, it shows up on the very simple menu on the Roku box. When you select a title, it buffers for about 30 seconds and then starts to play. Video is encoded at 4 different quality levels:

• One dot is 0.5 Mbps.
• Two dots is 1.0 Mbps.
• Three dots is 1.6 Mbps.
• Four dots is 2.2 Mbps.

Most of the time I experienced three or four dots, which is quite watchable on my regular TV set. Two dots is pretty bad - think VHS - and one dot is just terrible.

Lessons

Of course, the jury is out on whether all of this will have any impact on cable operators. I think most people are unlikely to cancel their cable or satellite video subscription and exchange it for the vastly inferior experience of either of these services. For one thing, you need to purchase and connect another set-top box that only supports on-demand programming. However, cable operators can learn some important lessons:

• According to the long-tail theory, a greatly expanded range of VOD programming has to potential to drive increased usage and margin. (Niche programming may allow the operator to derive a more favorable revenue share.)

• The guide technology is critical - potentially a model similar to that used by Netflix could be adopted by cable operators, allowing a Web-based search and recommendation experience that allows the customer to pre-select from a large number of choices. Some operators are actively looking into this.

• When the array of choices expands so radically, it becomes very difficult for individual subscribers to find what they are looking for. This is where recommendations will become very important.

Michael Adams is vice president, Systems Architecture, for Tandberg Television. Reach him at madams2@tandbergtv.com.