Long Tail TV: Part III

I promised to offer some thoughts on what the shift from Broadcast to Narrowcast to Nanocast will mean for the future of prime time. And there is a simple answer: it's sure not good for the upfront. But the more detailed answer is just as interesting. I'll start with a little history.

There are quite a few precedents for Long Tail effects, ranging from the introduction of the printing press to the arrival of the Sears Roebuck catalog on the prairie. Practically every time there's been an expansion of access to a wide variety of goods, we've seen shifts in the demand curve toward niches. The fact that everyone buys/watches/listens to or otherwise consumes the same few things doesn't necessarily mean that that's what they all want. It may be just that that's what's available.  Plenty of "hits" are artifacts of scarcity, not the triumph of common taste.

Television has already gone through one such shift. The introduction of cable TV in the late 1970s began as a rapid expansion of the number of channels available in the average American home, a trend that continues today with the switch from analog to high-capacity digital cable. In 1980, the average home could watch any one of 16 channels (probably using an antenna on the roof); nearly 80% of them were watching one of the three main networks in prime time. Today, nearly 95% of those homes have cable (AKA "multichannel" in the chart below), with an average of 110 channels each. And now only half of them watch the networks in prime-time, despite the addition of Fox as a fourth network:

But there’s another way to look at that diving market share line. A drop from 75% to 50% over the period shown means that 25% of viewers had shifted to watching non-Network TV instead. In other words, a quarter of the audience had migrated down the Tail, adding to the quarter that were already there.

Now to be fair, you might not consider MTV or HBO part of the Tail. We’re still talking about channels that are watched by millions of people each night. But as the number of channels available in the average home goes from 80 (analog cable) to 200 (digital cable) to infinity (Internet TV), it’s easy to see the shift towards less concentrated audiences continuing and more people migrating to truly niche content.

“Prime Time” is already meaningless to a TiVo user. So is “channel”. Soon “available in my area” will be, too. As the instruments of scarcity dissolve, audiences are flowing more naturally across the spectrum of available content, usually at a cost to the hits.

To flesh this out,  I spent some time with Howard Look, who runs TiVo's applications and  interface team. He was kind enough to run some numbers on TiVo users' behavior based on what they've got programmed as season passes. The first chart shows the percentage of users who have any of the top 100 shows programmed to record. The second extends that to the top 7,000 shows, averaged over windows of 100 shows (click for larger versions).

What these charts show is that the top 24 season passes make up 25% of  total season passes, and the top 80 season passes make up 50% of the total season passes. But what that means is that the bottom 6820 season passes make up the other 50%. In other words, the Tail is about the same size as the Head. Which is exactly the same ratio that emerged after the introduction of cable, but now on a massive scale.

(A definitional aside: Why draw the line between Tail and Head at the top 80 shows? In truth, because that's where the 50% marker happened to fall. But there's some logic to it, too. Nielsen only reports audience share for the top 100 shows, which pretty much defines where the industry draws the line between hits and misses. 80 is close enough to 100 to make the point.)

Is it possible that there's a fractal geometry at work here--that when distribution bottlenecks go away human nature tends to divide demand equally between a few hits and a lot of niches, regardless of the scale? That instead of the 80-20 Rule we know so well from our world of constraints, unconstrained demand is actually closer to 50-50? It will take a lot more examples for me to prove that's the case, but it's a very interesting coincidence at least.

That's enough to chew on for today. Tomorrow, I hope to finish up this series with some  thoughts about making and finding the video that will populate tomorrow's Long Tail TV.

Comments

[Mark]

I don't doubt that in general terms you're right on. But one fine point to consider: Over the time period for your TV data ('85 to '08), lots of households gained other media choices besides TV -- Internet, video games, VCRs/DVDs. So strictly speaking it's likely that a portion of the folks no longer watching network TV may no longer be watching TV at all. I would assume you could get at these numbers if you looked at network market share vs. market share for all channels.

But this is just a small detail. In broader terms, your point is exactly on.

[Robert Paterson]

Chris - Just a word of thanks - I think that you are onto a very important idea and you express it so clearly
Rob

[TW. Andrews]

So, I'm absolutely convinced that the long tail is coming, and it's going to be great. I mean, who doesn't like unlimited choices. My question is how do I get stuff I like out of it? Or more to the point, how do I get stuff I like out of it without having to invest an inordinate amount of time into getting it? Are niches mostly for people whose desire to match their expenditures to their tastes is strong enough that they're willing to sort through all the stuff in the tail that they don't like?

How does the tail manifest itself for the average Joe? Someone for whom the idea of purchasing online is appealing, but who doesn't have the time, inclination, or money to do it often enough to build up a reliable database of tastes? It seems to me that the real profit centers of the long tail will be instutions which master the complexity of the long tail for the average bloke and present him with a reasonable number of good choices with limited initial input.

[Kevin Kelly]

Chris,

Have you noticed how the Long Tail curve is the Singularity Curve on its side? I think they are connected. Seriously.

It's accelerating change which drives the plentitude of niches. And plentitude which drives the demand for faster power and more choices.

These two curves -- the Long Tail and Exponential Accelerating Change -- are the curves of our time. But I think the two also have an element of illusion in our hopes for them.

I don't think the Exponential Change curve means singularity as most commonly understood -- as in a Kurzweilian rapture. Likewise I am beginning to question whether the Long Tail curve means what you set out to claim: that the backlist outnumbers bestsellers. The curve is real; the numbers there. But you can make your claim simply by moving where the head and tail begin. The long tail becomes very long when the head is very short.

It is perfectly reasonable to expect that the neck (where the head ends and tail begins) will shift depending on the media or product or service. But rather than declare the new economy means a long tail, I'm inclined to think it means a new taxonomy.

There will be short headed, mid-necked, long tailed platforms. There will be some media who are All Tail. A very few all head, micro-tail. Some that are Long Neck. The real work will be in classifying the shape of the beast, and then to develop a predictive theory on what particulars of a product dictate what morphology its consumptivity takes.

The Long Tail, like the Singularity, and the Long Boom I would have to say, are a productive exaggeration that might mask even more useful subtleties beneath their simple curves.

[katie]

This is incidental to your main points, but I can't let it pass uncommented: that Network Prime-Time Audience Share vs. Multichannel TV Household Penetration graphic is terrible. It's got two datasets with the same units (percentages). I can't decide whether it's worse that

• each dataset has its own range, one from 50% to 80% and the other from 35% to 95%; or
• these datasets are even being represented on the same graph, since one represents percentage of people watching TV, while the other represents percentage of households.

The time axis is another mess: the interval between equidistant datapoints is variously five years and one year. And the last four datapoints in each set are conjecture.

[Chris Anderson]

Katie,

You're absolutely right, but it isn't my chart (it's lifted directly from Veronis Suhler's 2004 media report) and I couldn't redraft it because I don't have all the underlying data points. I'll be getting them and fixing the chart as you recommend for the book version.

[Robert Paterson]

Does the symmetry between popular and niche look like Dan Pink's Well Curve

[David Palmer]

Chris,
I think Kevin makes a good point, that the division between head and tail is somewhat arbitrary, and that there are numerous forms of beast (head, neck & tail combos) relevant to the current economy. Regardless, I find myself especially interested in what's going on in the tail, whether it outweighs the head or not.

As both a creator of content (paintings, music) and a consumer (esp. of music and books), I find myself interacting with the marketplace quite differently from how I did just a few years ago. I've been enjoying your posts, and look forward to the book.

[Chris Anderson]

Kevin:

I think you make a good point: the point is not the relative size of hits and niches, since the definitions of each are both interrelated and somewhat arbitrary. It's the *addressability* of the non-hit markets, allowing companies to profitably aim at any slice of the tail they want rather than all jamming up at the left side.

[John Thacker]

Mark-- it's "share," not ratings. "Share" is the percentage of all people watching TV, while ratings are the potential of people who own TVs. Since it's share, it's measuring the proper thing, and ignores people who are no longer watching TV. (And perhaps engaging in even more niche activities.)

[Alex Dante]

Chris, your closing musings about fractal geometries remind me of Duncan Watts' 'Six Degrees', which deals with the concept of 'small world' networks. One section covers how only 5 random connections will (roughly) halve the communications distance...regardless of the size of the network. It's a fascinating book, much like your site.

[Doug Stone]

Chris, when you are taking Katie's advice and cleaning up the graphs, also take into consideration standard deviations. This might make the head vs. tail cutoff less arbitrary for various media. And I recommend any of the Tufte books that cover presenting numerical information visually.
Also, I recall the internet "channel" concept pushed by Microsoft and others a few years ago. Is it still around? Is it just a means of transferring a television paradigm to the web?

[David St Lawrence]

Chris, thanks for formulating this concept.

I think we will find that the abundance of communication and distribution channels has a similar effect on employment. There will be a growing host of post-corporate business endeavors which will be small and micro businesses run by ex-corporate employees.

While the bulk of financial opportunity may occur during the brief years of corporate life, there will be significant opportunity with enviable lifestyles during this post-corporate existence. My weblog observes this area closely as I believe it will be increasingly important in the coming years.

[random answers]

A classic example of nanocast today would probably be youtube. Streaming on demand content in any niche and all of it searchable.