Software Robots Determine Content Quality
July 15, 2009
ZDNet ran an interesting article by Tom Steinert-Threlkeld about software taking over human editorial judgment. “Quality Scores for Web Content: How Numbers Will Create a Beautiful Cycle of Greatness for Us All” is worth tucking into one’s folder for future reference.
Some background. Mr. Steinert-Threlkeld notes that the hook for his story is a fellow named Patrick Keane, who worked at the Google for several years. What’s not included in Mr. Steinert-Threlkeld’s write up is that Google has been working on “quality scores” for many years. You can get references to specific patent and technical documents in my Google monographs. I just wanted to point out that the notion of letting software methods do the work that arbiters of taste have been doing is not a new idea.
The core of the ZDNet story was:
Keane is at work on figuring out what will constitute a Quality Score, for every article, podcast, Webcast or other piece of output generated by an Associated Content contributor. If his 21st Century content production and distribution network can figure out how to put a useful rank on what it puts out on the Web then it can raise it up, notch by notch. This scoring comes right back to the Page Rank process that is at the heart of Google’s success as a search engine. “The great thing about Page Rank in Google ‘ s algorithm is … seeing the Web as a big popularity contest,’’ said Keane, in Associated Content’s offices on Ninth Avenue in Manhattan.
Mr. Steinert-Threlkeld does a good job of explaining how the method at Mr. Keane’s company (Associated Content) will approach the scoring issue.
My thoughts, before I forget them, are:
- Digging into what Google has disclosed about its scoring systems and methods is probably a useful exercise for those covering Google and the businesses in which former Googlers find themselves. The key point is that the Google is leaning more heavily on smart software and less on humans. The implication of this decision is that as content flows go up, Google’s costs will rise less quickly than those of outfits such as Associated Content. Costs are the name of the game in my opinion.
- Former Googlers are going to find themselves playing in interesting jungle gyms. The insights about information will create what I cool “Cuil situations”; that is, how far from the Googzilla nest with a Xoogler stray? My hunch is that Associated Content may find itself surfing on Google because Associated Content will not have the plumbing that the Google possesses.
- Dependent services, by definition, will be subordinate to the core provider. Xooglers may be capping the uplift of their new employers who will find themselves looking at short term benefits, not the long term implications of certain methods.
I think Associated Content will be an interesting company to watch.
Stephen Arnold, July 15
Comments
2 Responses to “Software Robots Determine Content Quality”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
[…] the original post: Software Robots Determine Content Quality : Beyond Search View admin’s Profile Subscribe via RSS RelatedBookmarksTags […]
[…] is the original: Software Robots Determine Content Quality : Beyond Search View admin’s Profile Subscribe via RSS RelatedBookmarksTags […]