New Google Approach to Search Relevance

As Google became progressively more powerful, more and more people began trying to manipulate Google's "organic" search results. That aim has spawned a whole industry — search engine optimization — and in fact resulted in more manipulated links (some would say "spam") on results pages.

Now Google is considering implementing a new approach to ranking news search results that could potentially, equally be used to determine the relevance of ordinary search results and improve their quality.

This is probably old news to "geeks," but news to us "suits." (And there's already been a great deal of "grassroots" criticism of the move as favoring corporate media such as CNN vs. bloggers.)

Currently, Google uses a proprietary technology called "PageRank," which, according to Google, "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 other words, popularity is the primary determiner of relevance. And while Google tweaks the algorithm to keep it from being totally manipulable, SEOs have done a good job of keeping up with Google.

The news search algorithm would weight different factors (going to quality and credibility) rather than pure popularity. That weighted approach could be applied to general Web search — although I don't profess to know anything about how that would work "under the hood." And I would imagine that's what will eventually happen after Google test drives the system in news.

A conceptually similar "weighted" approach is what is being used by Openlist (formerly Local-i) in their local search algorithm.

Such a weighted approach might successfully prevent gaming Google's search results in the future, while boosting their quality. It's really a question of what gets weighted/emphasized.