I was reading an article about how Google is going to change their algorithms to penalize sites which are overly optimized for SEO. It’s no surprise to arrive at an end state where rules of a system are gamed when pecuniary gains are available. The entire SEO industry has grown up around trying to figure out how to optimize sites for Google’s black box. To ensure an improved customer experience, this action was inevitable.
Paul Graham recently wrote that building a search engine was a frighteningly ambitious startup idea. He accurately points out that the once spartan design of the Google homepage is now an overly cluttered mélange of links and ads. PG is calling for hackers to build a better mousetrap, but what I thought of when I read his post was about Heisenberg’s uncertainty principle.
You see, Google is a completely algorithmically driven enterprise. They measure and measure and measure. In doing so, they have sought to index the world’s information and make it available to anyone. Many would argue that they have been incredibly successful. After reading PG’s post, I began to think that perhaps Google’s incessant measurement of click-thru rates and speeds has not only changed the web for the worse, but created a situation where they could no longer measure what was good for their customers; that the measurement of the one variable reduced their ability to know the value of the other.
As for the title of this post, I have taken one part Heisenberg and added in Schrödinger’s cat when evaluating the situation with Google. Their black box algorithms contain the cat (happy web surfers) and the radioactive isotope (monetization of search). They have never allowed anyone to peer inside their box, so their happy customers are both simultaneously alive and dead. Sadly, in order to inspect the state of their customers, they will need to measure them. In measuring them, they will need to open the box. In opening the box, they will alter their understanding of the happiness of their customers, while simultaneously reducing their understanding of their measurements tied to their ranking algorithms. It’s a vicious cycle.