Many small bets, and a clear way to pick the winner
The method has two halves. Most teams remember the trying and forget that someone has to decide, by a stated rule, which result counts as best.
You can drastically increase your chances for success if you try lots of things and bring the best forward.
Jonah Peretti learned early how poorly anyone predicts what spreads. In 2001, as a graduate student, he asked Nike to print the word “sweatshop” on a pair of custom sneakers, and when the company refused he forwarded the email exchange to a few friends. It circulated to millions of people and into newspapers. He had not planned any of it. Much of his later work, at the Huffington Post, which he helped found, and at BuzzFeed, which he started in 2006, can be read as an attempt to turn that accident into a method.
The method is in the line. Since no one can reliably pick a winner in advance, make many attempts cheaply, watch which ones people respond to, and put more effort behind those. The first half is volume. The second half, bringing the best forward, is selection, and it is the half that takes discipline.
Online media is close to the ideal setting for it. A headline, a list or a short post costs little to make, feedback arrives within hours, and a failure is forgotten by the next day. Ten attempts where one takes off beat one polished attempt that might. The same logic runs a lot of web product work: small variations of a signup page, a dozen ideas for a feature tested with a few users each.
The method weakens as each attempt gets more expensive. A drug company cannot run dozens of clinical trials for fun. A hardware startup gets perhaps two or three chances to build a product before the money runs out. A restaurant cannot open ten versions of itself. Where each try costs a year or a large share of the budget, the advantage shifts toward thinking harder before trying.
The selection half has its own trap. Bringing the best forward requires a definition of best, and whatever is easiest to measure tends to become the definition. Clicks, shares and signups are quick to count. Whether readers trusted the piece, or whether users stayed a year, takes longer. A team that tries many things and always promotes the one with the most clicks will drift toward whatever gets clicked, which is not always what it wanted to build.
So the line is most useful read as a pair of instructions: make attempts cheap enough to run many of them, and decide in advance what result would earn one of them more attention.