May 26, 2024

Long-Tail Innovation

Recently I saw a great little talk by Cory Ondrejka on the long tail of innovation. (He followed up with a blog entry.)

For those not in the know, “long tail” is one of the current buzzphrases of tech punditry. The term was coined by Chris Anderson in a famous Wired article. The idea is that in markets for creative works, niche works account for a surprisingly large fraction of consumer demand. For example, Anderson writes that about one-fourth of Amazon’s book sales come from titles not among the 135,000 most popular. These books may sell in ones and twos, but there are so many of them that collectively they make up a big part of the market.

Traditional businesses generally did poorly at meeting this demand. A bricks-and-mortar book superstore stocks at most 135,000 titles, leaving at least one-fourth of the demand unmet. But online stores like Amazon can offer a much larger catalog, opening up the market to these long tail works.

Second Life, the virtual world run by Cory’s company, Linden Lab, lets users define the behavior of virtual items by writing software code in a special scripting language. Surprisingly many users do this, and the demand for scripted objects looks like a long tail distribution. If this is true for software innovation in general, Cory asked, what are the implications for business and for public policy?

The implications for public policy are interesting. Much of the innovation in the long tail is not motivated mainly by profit – the authors know that their work will not be popular. Policymakers should remember that not all valuable creativity is financially motivated.

But innovation can be deterred by imposing costs on it. The key issue is transaction costs. If you have to pay $200 to somebody before you can innovate, or if you have to involve lawyers, the innovation won’t happen. Or, just as likely, the innovation will happen anyway, and policymakers will wonder why so many people are ignoring the law. That’s what has happened with music remixes; and it could happen again for code.


  1. Hello, can i use your information on my site?

  2. john erickson says

    So, I’m very interested in the relationship between Anderson’s Long Tail thesis, the copyright industries, the economy of derivative works, and esp. transaction costs.

    11 years ago I started a company (NetRights) whose premise was “making copyright easy”; I was even so bold as to call it “Copyright for the Rest of Us” (to paraphase Steve Jobs). Our core idea was to essentially “objectify” flat media objects so that (a) static metadata about the would be immediately visible in the usage context, (b) dynamic metadata could be pulled from a remote server and similarly rendered, (c) pointers to higher-value content would be given, and (c) transactions ranging from permissions requests to inqueries about commisioned work could be initiated. We would do it for a wide variety of media types and transactions, not merely the limited set that CCC and its ilk handle.

    We thought big and small publishers would be interested because it would automate what was (and is) for them an expensive process; would faciliate transactions that were previously hard and therefore lost; and therefore would represent “found money” to them — that area under the Long Tail that they were ignoring.

    But our bigger thought was that this would also enable an economy of derivative use, i.e. lots of “small providers” doing low-transaction-cost transactions for re-use of material in games, edutainment, web sites, and graph works. In fact, an example of how our stuff would be applied TODAY would be, get a video from YouTube, decide you could use it, engage in a transaction from the context of your production (i.e. from within Premiere or Photoshop or Director) and get it done.

    The problem is that the transaction costs for these micro-rights transactions are cumulative, extend all the way back to the creation event of the work, and are directly tied to (and dependant upon) the creation of the metadata that enables subsequent transactions. In other words, to even have a chance, a primary artifact must be “born” with links to attribution, a rights template (such as a Creative Commons license), etc. Even if you start the metadata wrapping when it is ingested in a repository (such as YouTube) you already have jacked up the transaction cost, although possibly not several enough to collapse the curve.

    The interesting bit is, the more “valuable” an artifact is, the more investment can be made in transaction-enriching metadata, including niche and one-off rights pricing templates. In a subsequent system (Copyright Direct) we tried this with ala carte permissions that could be re-used, but the up-front metadata costs were still too high.

  3. Ed,

    The link to the wired article is wrong:

    Instead of this:



    [Thanks. I fixed it in the main post. — Ed]

  4. To me, the long tail is not the book authors who sell 10,000 copies or the code authors who expect to make a modest living. Those people are in the middle of the curve. Lots of the mass in the tail consists of books that do (or would, if written) sell a few hundred copies, and bits of code that have only a few users or that form a tiny part of something bigger.

    Why do people write these things? Not because they expect to make any real money, but to scratch an itch, or to tell a story that they think needs to be told, or as an indirect way to further some other objective (which might be monetary).

    Many works in the long tail are produced for private reasons, and the author is willing to release them to the public as long as there aren’t any legal or financial barriers to doing so.

  5. Barry: It’s not entirely clear to me that by restricting choice, something close to all sales can be redirected. Media content is not a closed system. But then you said neither.

  6. It’s a floorwax and a dessert topping. Many of the people who supply long-tail content (and may I say that the custom of describing creative work as “content” is a symptom of the mentality that attempts to destroy the long tail) aren’t motivated entirely by profit, or are motivated by gains that are indirectly monetized if at all. BUT — the fact that they can make a decent living (or a supplement to their regular living, or a way to further their career, or whatever) by creating that stuff makes it easier or possible for them to do it. Otherwise you limit the long-tail creators to those who are independently wealthy, employed in some approximation of a sinecure, or so dedicated as to be irrational.

    The other thing to keep in mind is that the tolls inhibiting innovation aren’t just $200 here or $400 and a nondisclosure agreement there. Most people in the group we’re talking about are already used to spending a few hundred to a few thousand dollars a year on equipment, internet access, hosting fees. Thanks to the current punitive (for some people) intellectual property regime in the US at least, there’s always a nonzero probability that some bright idea, if you implement it and disseminate it widely, will cost you your home, your car, your kids’ college fund, half your earnings for the rest of your career, or maybe even some guys with badges busting down your door and taking you to jail for an indeterminate time.

  7. Indeed. As a web programmer, much of the software I develop for my clients is in the long tail. Every website has a certain set of basic programs, but everybody has their specific requests which require complete, from the groud up coding.

    As far as the content industry is concerned: I agree with the article. While profit may be enough for “people can make a perfectly decent living by reaching a tiny fraction of the total market” that does not mean that is the point behind the innovation that reaches that niche market. Not only that, but I think that we are talking about a much smaller corner of the market than that “10-50,000 copies a year” for books.

    The internet blogging revolution is a perfect example. Almost all of the people that write blog posts (about mainly non-personal topics, such as this blog) do not do it for any sort of compensation – specifically financial.

  8. I agree with both enigma_foundry and paul that the following is a misstatement: “Much of the innovation in the long tail is not motivated mainly by profit — the authors know that their work will not be popular.”

    Further, I am wondering about the rationale behind your statement “But innovation can be deterred by imposing costs on it.” What is the concept behind for these undefined “costs”? My take, like that of Barry, is that DCMA/DRM is designed to add (barriers to entry) a regulatory cost, monetary cost, and time-delay cost to the niche innovator. Essentially, those who wish to implement DCMA/DRM seek to amputate the “long tail” (a buzzword I dislike) from the “head”. Artists have the ability (through the internet) to market directly to the consumer, they no longer need to go through the content providers. As Barry wrote “the Internet has the potential to be the Great Equalizer.” DCMA/DRM legislation/techologies will cripple the niche innovator.

  9. This is even more interesting when you consider that today’s content industry (specifically, movies and music) strives to exclude the long tail from the marketplace through control over that marketplace. If the long tail serves 25% of the market, that’s 25% of sales that the content industry loses out to small-time independent artists. Over the past several decades, the industry’s practices have been crafted to exclude the long tail from servicing these sales in order to force those sales back into the “mainstream” market. Media promotion is all but impossible to come by if you’re small-time, and in the face of the marketing machine for the mainstream, these niche players are crammed into an even smaller niche than would otherwise be available.

    This is part of the risk of allowing copyright to trump technology, too – the Internet has the potential to be the Great Equalizer, but the content industry, knowing that the long tail has the potential to serve such a large percentage of their current sales, is doing everything it can to burden the technology arena such that they are the only ones who can benefit from the law.

  10. enigma_foundry says

    I think you may be misstating things a little when you say “Much of the innovation in the long tail is not motivated mainly by profit — the authors know that their work will not be popular.”

    I agree, to but for different reasons.

    A participant in a long-tail market segment may be making his money by other ways–for example he may write a piece of open source software to play its part in pharmacuetical production or discovery, and make money from the pharmacuetical manufacturering process, and not care if he spends a little time writing some open source software which he can’t directly monetize–but the value of knowing he’ll get the future software improvements make it very sensible for him to contribute the code.

    So, the long tail may be monetized in a different way then the bulk market.

  11. Supposing that a long-tail innovator is motivated by profit, or at least making enough profit to support the desire to continue innovating; then, at the margins, a toll put on his ability to innovate would either result in (a) less innovation (if he complies with the toll, but can no longer make enough money to support his innovation) or (b) the toll being ignored.

    However, in your situation, I think that factors lean toward (a) because if the innovator is generating revenue s/he will be a larger target for a lawsuit or enforcement action.

    I believe that larger point is that erecting these tolls are contrary to the public interest because these long-tail innovators support around 25% or the market demand.

  12. I think you may be misstating things a little when you say “Much of the innovation in the long tail is not motivated mainly by profit — the authors know that their work will not be popular.” The whole point of a long-tail-friendly market (such as publishing used to be) is that people can make a perfectly decent living by reaching a tiny fraction of the total market. You won’t get rich writing books that consistently sell 10-50,000 copies a year, but you’ll make enough to keep writing books.

    It’s common wisdom that pretty much all forms of media have become hit-driven during the past 30-40 years. Hits yield intellectual, marketing and manufacturing economies of scale. End-user software has followed a similar evolution in large part.

    Who cares about this history? It may be important if what we’re talking about is essentially a reversion to the normal course of business after a period of anomaly, rather than something fundamentally new.