Those associated with technology and its markets knows that any popular explanatory framework will suffer a semantic death at the hands of marketers, promoters and other hucksters. Sometimes sooner than later. Why, just this week...
I heard a senior exec of a large enterprise software explain why their architecture supports and enhances the Long Tail phenomenon. And shortly thereafter received a pitch from a component architecture startup saying exactly the same thing. Both rubbish. Let's review, shall we?
The potential for a Long Tail market (or attention) distribution exists when the fixed costs of production, distribution, and buying are reduced by technology or other means. Then you don't need as many units of whatever to amortize the fixed costs to a competitive point, and niche products are supported with a decent ROI. BigCo's appeal to the Long Tail falls down right here, as I didn't hear one word about how they are going to reduce the substantial (indeed, notorious) fixed costs of adopting their solutions. Megabuck antes on the buy side will likely keep down the size of their market to the point where extending the tail of developer solutions into specalized niches is problematic from an ROI standpoint.
The degree to which a Long Tail evolves depends on the degree to which costs of transactions are also reduced. Transaction costs aren't just monetary, they also include the overhead and cognitive burdens of finding, deciding on, and integrating the value of each purchase (or use of your attention), as Kevin Laws ably points out. The bigger the monetary and non-monetary transaction costs in a market, the larger the granularity of the practical transactions, the less chance for the fine grained decision making that supports a Long Tail distribution. And I'm here to tell you that the transaction costs of deciding to read or not read a blog post are in no way comparable to those implied when choosing to adopt a software module, form a relationship with its vendor, and build it into architecture and business process for the future.
Yes, technology can in some cases also mitigate transaction costs, by automating much of the search process, often replacing branded, manual aggregation. But with automatic, requirements driven programming still largely in the Real Soon Now category, that doesn't apply to software. Modular software vendors, flagged with the programming flavor du jour, have long ignored the transaction costs aspects of selling at the module level, and have all left larger or smaller craters on the landscape. The attentive will recall that I worked at Kaleida Labs - sic transit gloria. That's one lesson learned on OPM that still holds. You're not going to find my partners' money at the bottom of one of those holes, wannabe LittleCo.
So this is my moral for this weekend: Just say 'no' to bogus Long Tail pitches. This useful meme, like all others, is on its way to the shark tank. But let's at least try to delay its arrival there until Chris Anderson's book can appear and its publisher recoup his munificent advance.