 |
     |
|
|
What follows is a final draft of a paper I've been working on. Comments welcome.
One original promise of the Internet was that it would render location irrelevant. That, of course, is bunk. It's true that in these early days for the Internet, the ground remains unsettled and could change tomorrow. Be that as it may, one thing can never be overlooked: Ultimately we are physical beings with physical needs. While some of our needs and desires can be made digital, most of the important ones can not. And so geography rears its head and puts the lie to great plans and forces rethinking about opportunities.
The most seductive business models of the Internet age (so far) are built on a pillar of geographic independence. Amazon, eBay, Expedia, and iTunes, by way of example, have each succeeded with an implied freedom from the bonds of geography. Tom Friedman makes the case in his latest book, The World is Flat, that globalization (not strictly but substantially equivalent to "geographic independence") has been richly enhanced by the Internet. He and others point to how the Internet has provided the structure for knowledge service providers with greater intelligence, more ambition, and lower-cost environments to succeed at servicing Western needs without geographic proximity. Fair enough. In some ways the Internet has leveled geographic inhibitors. But over-generalizing to represent proof that the Internet frees us from geography is simply wrong.
This paper is not about globalization, the forces for which go well beyond communications media. Rather, it's a perspective on why and how geography continues to matter in significant ways -- even to those bellwether businesses of the new age.
eBay probably would not be any measure as large were it restricted to the same small geographic market area upon which a flea market relies. Because sellers and buyers reside at the end of any Internet connection, eBay can host an enormous volume of commerce involving items for which there is no significant local market. It's axiomatic that the likelihood of a buyer and seller finding each other diminishes as a market space is constrained. But widely separated, unknown people with the same interest in some thing find each other on eBay.
Amazon's rich product assortment is feasible only because it holds inventory at and processes transactions from distribution centres at a cost lower than any retail store. More significantly, it can even hold esoteric items for which the market in its service area (arguably the world) is severely restricted. So Amazon is the first source for popular and unusual merchandise because it operates in a relevant market: one that is substantial and efficiently serviceable. And for a substantial part of its business, Amazon has no relevant competition. Somewhere between 30 - 50% of Amazon’s sales are of merchandise that the average bricks 'n mortar big box bookstore doesn't even carry. Traditional retailers won't typically stock such product because there is no local market to buy it in a reasonable amount of time.
iTunes sells current hits in substantially greater volume than the old and the unusual. But, one should expect that iTunes probably generates a significant portion of revenue from music that retail and mail-order music stores/services simply can't offer. Chris Anderson and Joe Krauss point out that iTunes has a library of well over one-million pieces of music and each one has been bought at least once. It's nearly a perfect model: unrestricted practical market breadth, unlimited supply, and effectively no marginal cost to distribute.
So what is this 'long tail' thing? As shown below, it is the overlooked part of the oft-forgotten 'power curve' distribution. This curve graphically represents the most common rank distribution in nature where there is uninhibited selection opportunity from a large universe of options. Whether the subject matter is individual wealth, structure of firm sizes in an economy, popularity of search terms or blogs, or through-put capacity of power-stations, the distribution invariably follows this pattern.
The power curve's essence is that in any population set there will be a very few who's riches is orders of magnitude larger than even the next largest member of the population. The measured value steeply diminishes rapidly and asymptotes from near vertical to practically horizontal. The 'long tail' refers to the unending horizontal line and the idea that while one member is definitely in first place, there is typically no end because the volume of small, undistinguished members of the population massively outnumbers the very rich few at the front end. For every Bill Gates there are a hundred million Joe Nobodys.
In the graph above, the data for which comes from Joe Krauss's accounting for a query distribution sampling at Excite in the late 1990s, the head and tail have been highlighted. What the graph represents is the order of magnitude discrepancy between the most-sought terms and the others. "Sex," the most popular term, was 100,000 times more popular than the 1,000th most popular term.
The numbers driving commercial exuberance for 'long tail' businesses are the percentages attributed to the head (3%) and the tail (97%). What's got everyone so excited is that they show in extreme relief what the businesses described above are experiencing to certain degrees. The head of the curve has the few winners, but the vast majority of members and volume in the market -- in this case all but a few percent of the total -- is in the tail. Logically, if a business has the features of those we've described:
why would it focus only on the popular? Why target only large, homogenous markets? Why carry and sell only the "hits?" It's counter-intuitive and fed by these few apparent successes and the data from absolutely constraint-free examples such as search terms.
limited market breadth
uninhibited supply
competitively lower cost to serve
What the seductive businesses should illuminate are certain conditions for bridging the gap between the 'long tail' as a distribution construct and a valid business. At least four geography-based conditions alone exist for successful transition from 'long-tail market(s)' to 'long-tail business.'
1. Product must be non-perishable. A fundamental characteristic of the 'long tail' is patience. While it ought to afford high overall volume, generated in small bits from a very broad base of customers, business in the 'long tail' is precisely the opposite of that at the head of the curve. Each item may take ages to move; inventory turnover could be well outside normal business levels. That could be a problem if the products offered have a limited shelf life. Non-perishability is critical in another way. In the seductive business models, the inventory of product is ideally infinite: a sale on iTunes does not create a replenishment need; eBay leaves inventory fluidly in the hands of sellers; a book can wait in a cheap warehouse -- as opposed to dear retail space -- to be sold by Amazon.
2. Cost to hold additional inventory and process more orders must be truly marginal. Enough said. Without access and capacity to hold vast amounts of product at a competitive cost advantage, it's not worth it. In addition, the cost to process a transaction must be favourable. This includes not merely electronic v. physical checkout but credit card processing fees, system costs, logistics, and so forth. Just as the revenues for 'long tail' businesses would come from selling ones of many rather than many of one product, so too will its cost structure be inverted relative to traditional businesses. Amazon and iTunes, for instance, appear to have achieved mastery of inventory and ordering processes and costs as compared to the traditional geography-bound alternatives. One could speculate that at some point though, that will not be enough.
3. Distribution costs specifically must be limited or off-loaded. Amazon, like catalogue retailers, charges the buyer for delivery. With eBay an agreement is made between buyer and seller as to who will pay the shipping cost. iTunes's delivery cost is bandwidth, which is (a) trapped within baseline infrastructure overhead and (b) gets more cost-effective with increasing volume. (Some traditional retailers, such as REI, have begun to use their retail locations as 'warehouses' and 'shipping destinations' for online orders. The seductive businesses thus externalize one of the principal barriers to distributed-market business models: moving the merchandise -- even the digital sort. Businesses that can't offload this critical cost -- and they are the majority -- will have difficulty benefiting from the 'long tail' model or the Internet in the obvious ways.
4. Legal jurisdictions are geographically-defined: accommodate borders. Despite globalization and regionalization in the form of partnerships and trading blocks or unions, the world is and probably will be politically defined by nation-states. Each nation will persist in asserting its sovereignty with unique, controlled legal systems. In theory the magic of the 'long tail' (especially) Internet-based business models is the absence of boundaries, including geo-political ones. In practice though, the geo-political realities of currency differences, taxes and duties, customs clearance time and fees, and so on are immediately visible beside shipping costs. Hidden just below the surface are more sinister legal issues like, well . . . legality. Consider:
"As the eBay community expands around the globe, we are encountering different laws and different points of view as to what constitutes illegal, offensive or inappropriate items," said Mike Jacobson, eBay General Counsel. "Given our expansion, as well as feedback we've received from our users, we reviewed our policy and concluded that these changes are appropriate."This was the global response from two large, 'geography-independent' online businesses to the public outcry and subsequent legal proceedings in one jurisdiction. Then there's iTunes. This truly global business (i.e., the music and its form is globally acceptable without structural obstruction) is trapped by borders. Its geographic challenge is that copyright and licensing laws for its product (music) are different by nation. That is the primary reason why the iTunes music store was unavailable in Canada for nearly two years after it launched on the south side of the 49th parallel.eBay is not the only auction site to suffer from public ire raised over the trade in Nazi-related items. Yahoo!'s Yahoo Auction site was sued in a French court over "alleged justification of war crimes."
Here's the rub: Most businesses, even the most seductive of 'long tail' Internet business models, do not satisfy these criteria sufficiently to render them truly geography-independent. Not only must there be sufficient value in the business proposition to overcome the constraints of space, but at a minimum the operating model has to accept and account for, not avoid, the impositions of geography.
Before business planning oneself into a virtual or 'long tail' model -- for the sake of investor, fad, or fashion -- make sure the basic business principles work. And, like one of my History professors would say, "Always go back to the geography."
Posted by Grayson at July 11, 2005 03:16 PM