Friday, June 27, 2008

The Brain Behind Google

The Brain Behind Google

Almost every company on the Internet these days is dependent on Google. Google now does more searches than every other search engine combined, and search is the single biggest driver of traffic for most websites.

Yet how Google’s search algorithms work is a mystery to almost everyone, especially given it changes its search algorithms over 450 times per year, more than once per day. For evidence of the power Google and its algorithms hold over business just look at, a public company that lost nearly 30% of its traffic in a single day last year and nearly a quarter of their market cap because of a change that Google made to its search formulas.

So how does Google work and how do you beat the system?

Google’s big innovation is the idea that the importance of a website is based on how many other websites link to it. And it is not only a matter of the number of links, but the quality of those links as well –the thinking being that the best websites should have many other websites that link to them.

Sound familiar? It should, because this is in fact the way the brain’s computing unit, the neuron, determines relevance of information, where the most popular and useful neurons have the largest number of connections. The best neurons, those with the richest connections, have the most links to other neurons around them.

Using links as a proxy for relevance works well for Google and for the brain. It is our natural way of behaving: we watch movies because they are box office hits and read books because they are best sellers. It is a strange recursive loop indeed, where neurons get stronger because they are strong and websites are ranked higher because they have a high rank and books are bought because they have sold well and movies are watched because everyone has seen them – we have entered a world where the rich always get richer and the popular girl becomes prom queen.

But, like the brain, Google’s algorithms are continually evolving and becoming more powerful. Google now looks at the relevance of each link and weights it based on whether or not it is similar to the website’s category itself. And this is precisely how neurons strengthen connections. A link from is more valuable to a site like than to It also looks at the quality of the website (yes, their algorithms actually read and process information), such that a link from Harvard Medical School is considered more valuable than Sonny’s Backyard Biomedical. This means that the site needs to have links to it by other relevant, high-quality sites. No longer is it good enough for a book or movie to be popular, now it has to be critically acclaimed, with a starred review or two thumbs up.

According to the gang at Google:

PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page's value. In essence, Google interprets a link from page A to page B as a vote, by page A, for page B. But, Google looks at considerably more than the sheer volume of votes, or links a page receives…it also analyzes the page that casts the vote. Votes cast by pages that are themselves “important” weigh more heavily and help to make other pages “important.”

This may not pass muster in a democratic election but it is a description that would make even the most ardent neuroscientist proud.

Just as we have this tangle of competitive interests on the Internet, so do we have it in the mind. In our cases, we are constantly weighing information, in a mental battle that we are rarely aware of. Google’s algorithms are mimicking the brain’s need to clear out the clutter and find the good stuff.

So what should you do? Follow the brain.

Stop trying to game Google and focus on building value. Improve the quality of your site, remove the clutter and focus on attracting relevant sites to link to your site. And if you are looking for a good model, look no further than Google’s website: uncluttered, massive links in to the site, and a nice big button for sending your information on.

No comments: