Comment Articles

How much do you trust Google?

Rupert Goodwins ZDNet.co.uk

Published: 10 Apr 2003 15:55 BST

For many people, the Web is Google. The first port of call for any query, that clean, uncluttered page has replaced portals, hotlists, directories and many other half-forgotten inventions for Web navigation. But all is not as it seems behind that friendly façade.

Search engines are beasts that look simple but hide enormous complexity. What could be simpler than typing in a word and finding where it exists on the Internet? If that was all Google did, it would be above suspicion -- and useless.

Google's charm and power comes from context. Not only does it record the occurrence of nearly every word on the three billion documents -- Web pages, PDF files and so on -- it knows about, it analyses how many references there are to each document from others. Very crudely, the more links to a page there are, the better the result. And if a page with a high rating links to you, that counts for a lot too. Finally, if your site is frequently updated, then Google loves it even more.

Which is fine, if you assume a Web that's entirely unaware of the way its being searched. But the Web is symbiotic with Google -- both rely on the other -- and like any good symbiote the Web adjusts itself to make best use of its partner. So people are becoming adept at creating groups of sites whose structure and content are designed to be artificially prominent in the Google lists.

It gets much worse with blogs, which could have been custom-designed to persuade Google to disgorge the nectar of high page rankings. Highly self-referential, nests of mutually linking blogs can take a phrase, duplicate it across ten or 100 sites in a short time, and hit the big time on Google. That's not news -- Google bombing, as it's called, was first described and demonstrated last year. There's already been at least one high-profile skirmish in the page ranks between the Church of Scientology and supporters of the anti-Scientology Operation Clambake. Google is itself aware of this and has made efforts to spot and deal with deliberate attempts to tweak the ratings. It's all very Darwinist, yet Google seems sure of itself and its mission to deliver its best attempt at objectivity.

But Google has recently bought Pyra Labs, the company that started blogging as a mainstream activity and that has around a million users. Those who worry about such things approve on the whole, but the effective recruiting of a large city's worth of online content and reference-generating people into the Googleverse could have considerable commercial impact on the way the Google rankings are created.