The Myth of the Long Tail (and the Importance of Blockbusters)

I read an article about 4 months back in the WSJ ("Study Refutes Niche Theory Spawned by Web") regarding a debate between traditional marketing theory where businesses focus on producing and distributing blockbuster products (that have significant marketshare) vs. a new paradigm where businesses shift production and distribution to a vast array of products tailored specifically to the individual tastes and needs of consumers.   The latter theory is generally referred to as the "Long Tail", but a recent study by a Harvard Business School professor seems to refute that theory.  Adding to the discussion, I just watched a fascinating interview with Eric Schmidt, the CEO of Google, in The McKinsey Quarterly where he talks about why businesses should not focus their strategy on the long tail.

First, what is the "long tail"?

We can summarize the long tail as follows (here we draw on a few sources including HBR - The Long Tail Theory in Short, WSJ - see link above, and Wikipedia - The Long Tail).  A market demand curve tends to have a distribution pattern where a few products claim the majority of demand (e.g. 80%) while a large remaining number of niche products comprise the rest of demand (e.g. 20%).  Much of this pattern can be attributed to traditional distribution channels.  Imagine, for instance, that you run a book-store.  There are literally millions of books that you can choose from to stock your shelves.  With a physical bookstore, however, you'll likely limit your selection of books to the inventory that sells the most - the blockbuster books making up the 80%.  The remaining vast array of books in the world that make up the other 20% are not in your store because the demand of each individual book is so low.  Consumers don't have ready access to these remaining books.  The 20% at the end of the distribution curve can be referred to as the "long tail".  Another name for this distribution pattern is a "power law" or "Zipfs law" - basicly a declining curve that tails off asymptotically (see the diagram to the right).

With the introduction of online retailers, however, you have an opportunity to have the entire long tail in your inventory.  An excerpt from the WSJ:

The Long Tail theory, as explained by its creator, Wired magazine editor Chris Anderson, holds that society is "increasingly shifting away from a focus on a relatively small number of 'hits' (mainstream products and markets) at the head of the demand curve and toward a huge number of niches in the tail."

The reason involves the abundance of easy choice that the Web makes possible. A record store has room for only a set number of titles. ITunes, though, can link to all of the millions of songs that its servers can store. Thus, said Mr. Anderson, "narrowly-targeted goods and services can be as economically attractive as mainstream fare." Managers were urged to adopt their business plans accordingly.

This can be described graphically as well:

The idea is that the entire demand curve over time will flatten, resulting in the "long tail" becoming fatter, making up more than 20% of demand.  In the long-run potentially, the tail could represent more demand than the blockbusters themselves.  As consumers are drawn by the possibility of finding and demanding niche products that cater exactly to their specific needs, the marketshare of concentrated blockbuster products will be shifted towards the niche products.  For businesses, that implies that you should focus on the tail.  As Anderson's theory goes, demand is going to shift there over time, so you should make it a core part of your strategy.

Now that we understand the long tail theory, we can examine what's wrong with it.  So what is wrong with it?  Why doesn't it hold up in the real world?

The study mentioned by Prof. Elberse at HBS mentioned in the WSJ article refuted this theory using data from the same companies Anderson used (Amazon, iTunes, etc.).  Here's an excerpt that touches on why the theory doesn't hold:

But Prof. Elberse describes research showing that even in our cultural consumption we tend to be intensely social folks. We like experiencing the same things that other people are experiencing -- and the mere fact that other people are experiencing and liking something makes us like it even more. Far from being cultural rugged individualists, most of us are only too happy to have others suggest to us what we'd like.

Building on this idea, Eric Schmidt emphasizes in his interview that the "long tail", although an intellectually appealing concept, is not the reality in the business world.  He says, instead, that you have to focus on the "head" or blockbusters because that's where the lion's share of revenue is.  The tail is also important (i.e. it's still 20% of demand), but it should definitely be a secondary focus.  For instance, there may be innovations in the tail that may turn into blockbusters in the head, but you can only fund R&D for those items if you are generating cash flow from those blockbusters.

This touches on the old growth share matrix (or product portfolio) model used by the Boston Consulting Group since the 70's.  In the model, you have "cash cows" that fund your ability to turn "question marks" into "stars" (by purchasing marketshare) and eventually into future "cash cows" (as blockbusters are more self-supporting and don't need funding to protect their marketshare).  A more recent HBS Working Knowledge article titled "Long Tail Economics?  Give Me Blockbusters!" touches on the same topic, drawing on examples from the movie and pharmaceutical industries.

The key take-away here is that consumer behavior is social in nature and hence there will always be a high demand for blockbusters - the "power law" distribution is what you should expect.  And, counter-intuitively, rather than the Internet flattening the demand curve, it's more likely to steepen it!  Schmidt believes that the global distribution medium of the Internet allows brands, media, products, and information to take advantage of the network effects of larger and larger networks.  So, you should expect to see increased concentration and blockbusters in those things as the network effect takes advantage of global scale.

Another key take-away - I guess we can't quite throw out the old paradigms yet!