The Long Tail

A physical record store could stock maybe 10,000 albums. Spotify carries over 100 million tracks, and the ones after the first 10,000 matter more than anyone expected. Chris Anderson popularized the long tail in a 2004 Wired article: digital distribution kills the shelf-space constraint, so when you can carry everything at near-zero marginal cost, the aggregate revenue from millions of niche items can rival the blockbusters. Amazon proved it first in books, stocking millions of titles where Barnes and Noble stocked tens of thousands. Netflix proved it in catalog watch-hours, Etsy in handmade goods, YouTube in the oceans of content made by creators with modest followings.

For founders and operators, the long tail is seductive precisely because it sounds automatic, and it is not. The aggregate of a million tiny niches only adds up if something connects niche supply to niche demand, and without that you do not have a marketplace, you have a warehouse full of inventory nobody finds. Spotify's Discover Weekly, Amazon's "customers also bought," and Netflix's recommendation engine are as load-bearing as the catalogs themselves. The piece of infrastructure that makes the long tail actually pay, and how to tell whether your tail is worth serving, is what the app holds back.