Leverage, the respectable kind of laziness

Google’s early ranking method is a good illustration. In the late 1990s the main way to organize the web was by hand: Yahoo employed editors to sort sites into a directory. Larry Page and Sergey Brin, then graduate students at Stanford, noticed that the web already contained millions of human judgments in the form of links. A page that many good pages linked to was probably good itself. By counting and weighting those links, PageRank put other people’s work to use, and it improved automatically as the web grew. Nobody had to edit anything.

That is the kind of laziness the line recommends: an unwillingness to do by hand what a system could do, and a habit of looking for the existing force that will do most of the work. A developer who spends a day writing a script to replace an hour of weekly manual work is being lazy in this sense. So is a company that lets its users generate the content, or a salesperson who finds a partner already selling to the same customers.

The line comes from the co-founder who takes over as Google’s chief executive next month, and its original setting has not been traced. It should not be read as a description of how Google itself works. The company employs tens of thousands of people and a great deal of its work is careful, manual and slow.

Leverage has limits. Some work does not scale and should not: a founder’s first sales calls, the early conversations with users, the hiring of the first team. Trying to automate those too early loses the information they provide. Leverage also magnifies mistakes. A system that does the work for everyone also repeats its errors for everyone, and financial leverage, as 2008 showed, multiplies losses as readily as gains.

Leverage, then, is a way to spend effort, not to avoid it. Finding the lever usually takes more thought and work up front than doing the task by hand. The laziness comes afterwards.