Ressources numériques en sciences humaines et sociales OpenEdition Nos plateformes OpenEdition Books OpenEdition Journals Hypothèses Calenda Bibliothèques OpenEdition Freemium Suivez-nous

Google Panda : quelques paramètres du nouvel algorithme

“Google Panda est un nouveau filtre mis en place par Google début 2011, déjà déployé dans les pays anglophones, et qui sera étendu prochainement au reste du monde. Google a travaillé depuis des mois, voire des années pour mettre en place Google Panda, en utilisant ce qu’on appelle un « algorithme d’apprentissage ».”

Tags:

Is Google’s Panda only biting the competition?

google_panda_algorithm_indexLast month, Google implemented a major update to its search algorithm, dubbed “Panda”. The point of this update was to deprecate so-called “content farms”—sites that turn a profit by posting lots of low-quality content that is used for earning ad revenue, search engine optimization, or both. These sites’ content is generally worthless in terms of actual relevance to the terms of the search, and ends up cluttering results pages with useless links.

One of the features of this update is that having a sufficient number of low-quality pages causes all pages on a site—even those of “higher” quality—to take a hit in search popularity.

Today on TechCrunch, Paul Edmondson, CEO of user-generated content site HubPages, posted a complaint that Google is not applying this formula fairly across the board. He points out that some of Google’s own user-generated content sites, such as YouTube, have quite a lot of spammy and worthless content, too—but YouTube (and other Google-owned properties, such as Knol) consistently rank higher in search engine results than non-Google properties (such as HubPages). (And he complains about Google not responding to his questions about how to fix his site.)

It’s easy to see this as the operator of a content farm complaining about being penalized for being a content farm, and hoping that throwing a public fit will get Google to take notice of him and answer his email—and indeed, a number of follow-up comments to the articles do accuse Edmondson of being just that. And others point out that comparing a text-based site like HubPages to a video-based site like YouTube is comparing apples and oranges.

Still, it does raise an interesting question about the possible conflict of interest inherent in a site that does both site searching and content hosting. Given how much of the search market Google controls (Edmondson puts it at 2/3 of the market in a comment), it could have a pretty significant advantage over non-Google-owned content sites if it chose to exercise it. What if it is exercising it? Is a Google-based content farm more deserving of traffic than a third-party site’s? And since Google won’t show anyone else its algorithms, how can we know for sure whether it is unfairly favoring itself?

These questions will have to be answered sooner or later. They may have to be answered in a courtroom or FTC hearing room.