I was just perusing some blog feeds in the downtime on Christmas Day and came across a link on Brand New to an Economist article about the hit aspect of the Long Tail debate.

More and more articles about the Long Tail tend to devolve the theory into a general comment on the state of industries affected by Internet economics.  Whereas Chris Anderson’s original argument is very ‘niche’ and specific - describing a set of circumstances that result in a specific outcome rather than a large, generalized theory of everything Internet related.

As I see more commentary on the Long Tail, it’s like a Long Tail of the Long Tail - arguments that tend to twist and turn the original idea to suit a set of observations or opinions.  Sometimes interesting, sometimes a stretch.

But alas, that is the destiny of many a theory.  Maybe the best ones tend to devolve to this kind of reference state - their specifics long ago forgotten, but their general idea intact well enough to guide curious minds who happen along their path.

The Tipping Point is another one that has achieved this ethereal status - although I can’t but help wonder if that was the point all along.

I can’t also help wonder if, by reducing the original Long Tail argument to a reference point, a lot of the original intent and usefulness is lost.

The Economist article has a section on the success of hit TV shows despite falling ratings from wandering eyeballs.  The argument being that even with reduced audiences numbers, we are still gravitating towards these ‘hits’ - as if there is comfort still in broad and common social discourse based on shared entertainment experiences.  Yet network TV in the US is in no way, shape, idea, structure, form, construction, similar in anyway, at all, to the basic tenants that underly the Long Tail theory.

The Long Tail adds nothing to the argument other than a reference for the general idea - that TV is generally full of more choice, that some of this is useful, but we still tend to choose similar things.

Chris Anderson has largely given up the fight to defend different interpretations of his theory.  I think this is a good thing.

It’s time to let the Long Tail go silently into the night.  With the realization that in death, it’s infinitely more powerful than in life.

The Long Tail, I predict, will truly become the Obi-Wan Kenobi of Internet theories.

I am currently re-reading Here Comes Everybody, Clay Shirky’s book on the new organizing principles of today’s world.  I have been meaning to do a review of it for ages and realized that I needed to re-read it again!

In the first chapter he cites a well-know probability exercise called the Birthday Paradox (yes, there is a Wikipedia entry for it).  

If I asked you what the chances are that two people in a room of 30 would share the same birthday, you would probably say not that high.  In actual fact, the chances are very high.  It’s a counter-intuitive problem because we aren’t good at understanding how probabilities combine.  

In his book, Clay Shirky uses the example as an illustration of the complexities of large groups.  He’s right.  As groups scale in size, our ability to understand the relationships between each individual and each other individual falls dramatically.

His other, more intutive example, is about trying to find a group compromise as the group scales in size.  It becomes exponentially harder.  We’ve all felt that one.

As I read this, I realized it was another nail in the coffin for traditional market segmentation studies.  If group complexity and relationships between members scale exponentially as the group gets larger, how could we ever expect to describe thousands of users of a product or service by five or six simplified labels?

It’s definitely food for thought.

Consumerism has become a fashionable area of concern, centred on the belief that there is a compelling need to strengthen the consumer’s position in the market-place. Consumerism is essentially a societal problem, based on the efforts of the average person to come to terms with everyday companies in everyday transactions. Yet the growth in size of business organisations, together with the increasing complexity of their offerings, is such that it is becoming ever more difficult for the ordinary individual to participate in transactions with any real hope of equality and satisfaction. It is this inequality of bargaining position that has become the “centrepoint” of consumer pressures on the community to provide for intervention and regulation in a socially constructive manner. Experience everywhere suggests, however, that it is not an easy area in which to operate. The issues of consumerism are complex and difficult to resolve, concerning as they do some of the most basic human relationships: those involving exchange. 

Journal of Management Decisions, circa 1974

I’m struggling through a busy patch at the moment so posting has been light recently. My wife and I are gearing up for our move to San Fran later in the year (we currently live in NY), so there is a lot of stuff to sort out - not least of which is shifting a cat 3000 miles.

So I am taking a moment to post about something that I have been increasingly coming across - the importance of understanding ’scale’.

And by ’scale’ I mean things that scale on some type of exponential curve. The Long Tail describes the scaling process inherent in the aggregation of markets for certain types of goods. The ubiquitous 80/20 rule describes the typical scaling of customer revenue. And in Clay Shirky’s book, Here Comes Everybody, he points out that in social interactions, these Power Laws (as they are called) are everywhere. This is a Power Law:

While Power Laws have been written about extensively, I don’t think they are well understood. Mostly because, as human beings, we live in a very linear world. We’re not good at understanding things that scale exponentially. Power Laws are all about extreme scaling. Mostly though feedback and multiplier effects.

A good example of this mis-understanding is the Sprint campaign that was widely criticized on blogs and social media sites. One of the criticisms leveled at it was the auto-response email you received if you tried to email the CEO (who gave out his address at the end of the TV clip). It’s a fair criticism. A personal appeal from the CEO doesn’t feel very personal if you get a canned response back. But it’s not surprising. On the graph above, Sprint occupies a position near the steep part of the curve. It has millions of customers. The CEO can’t have a conversation with each one of them.

I heard people comment that Sprint should have just hired more people to respond. And that if smaller companies can do it, why can’t they? Customer numbers follow a power law - the more customers you have, the more you get, the more you spend on getting more, etc. Social media tactics - two-way conversations, dealing with customers as individuals, having tailored conversations, etc. - don’t work as you climb that curve. They can’t. You can’t add resource to deal with those conversations at the same rate as you acquire customers because you can’t add resource exponentially and stay in business.

A lot of Social Media pundits talk about the 2-way nature of conversation these days not understanding that what works for a 100 person outfit is not going to work for a company ten times larger with 1000x the customers because of the way customer numbers and resource scale. Joseph Jaffe’s Delta Skelta debacle is a good example. As is the Target example.

In Jaffe’s case, he’s right to want to claim compensation for what happened to him, but wrong for thinking Delta can somehow treat him differently on the merits of his individual issue. He’s probably one of a few 1000 people they need to deal with on a weekly basis. A simple policy for his situation is the most efficient way to deal with it. A full-blooded, tailored conversation for his individual needs is not. The fact that he got a direct response from Delta is more a reflection of his standing in the blogging community than his value to them as a customer.

Now I’m not saying that having a canned email or a standard policy letter is ‘good’ - in the sense that it is the best type of customer interaction. It’s not. All I’m saying is that it’s a realistic response to this issue of scale.

Shirky makes this same point when talking about weblogs in his book:

As is normal in a power law distribution, most writers have few readers. Such readers and writers can all pay similar amounts of attention to one another, forming relatively tight conversational clusters… As the audience grows larger, into the hundreds, the tight pattern of ‘everyone connected to everyone’ becomes impossible to support - conversation is still possible, but it is in a community that is much more loosely woven… Once writers start getting more attention than they can return, they are forced into a width versus depth trade-off.

Essentially, as reader numbers scale exponentially, the blog writer has no hope of increasing their ‘attention resource’ in a similar way. You can’t add attention exponentially. You don’t have enough of it to start with!

The Sprint CEO knew this before he set out. There was no way he could have a conversation with everyone who saw that ad. Hence the canned email. And to be honest, he was silly to try. He held out the promise of such an interaction knowing he could never deliver. That’s not a great tactic.

Companies like Sprint and Delta and other large entities that exist on the steep part of that curve need to get away from this notion that they can profitably sustain 2-way conversations with their customer base. Companies that exist further down the curve absolutely need to keep those conversations going.

Where ’scale’ starts to be an issue, other types of strategies need to be employed. Social media holds out huge promise for self-sustaining group collaboration by customers - initiated by companies. Company as enabler makes much more sense than company as sounding-board.

Sounding-boards are only good to rant at. There is no future in being a sounding-board.

If you read this blog a lot (ad thank you for those of you who do!), you know that I have been following the work of Chris Anderson for a while - his Long Tail theory and the newer Free movement.

Chris recently posted a rebuttal to a Harvard Business Review article by Aniti Elberse criticizing some aspects of the Long Tail theory. Anti is a Professor of Marketing at Harvard Business School.

The debate boils down to the fact that Elberse’s analysis of new data found that the Long Tail is increasingly flat (i.e populated by less and less popular titles as inventory grows) and that consumers, even consumers in the Long Tail, prefer the ‘hits’ more (the popular stuff).

Both to of these findings seem pretty obvious to me. Online retailers are always going to keep adding inventory as long as inventory costs are close, if not at, zero. And consumers will always tend to gravitate to popular titles - at some point. Popularity is multiplicative. The more popular something becomes, the more popular it becomes.

However, I believe she misses the point of the original Long Tail argument when she says that these results undermine the importance of the Long Tail as a business phenomenon.

The Long Tail is an observation of optimal product portfolios for two very different delivery mechanism. The two things that make the Long Tail work are close to zero inventory costs and a system for recommending and searching content. Both of these facts predict her findings. More inventory is inevitable, and with recommendations still playing a big role, popular products are very likely to remain popular.

Anderson’s theory was never about the death of the ‘hit’. It was about the rise of the ‘tail’. These things aren’t mutually exclusive.