Craig Hordlow put forth an interesting concept publicly that he’s been working on for awhile noting that Google’s growth is slowing as more and more people go directly to other sites such as Craigslist or Youtube or Wikipedia to find information within their respective niche as those sites become the authority over time and as such, people begin to not use Google’s search engine as much. He explains it as such:

What I mean by that is that, for example, people are less likely to go to Google and search for “apartments in San Francisco”, or “cleaning services in San Francisco”. They’ll either go straight to Craigslist.org, yelp.com, cityspace.com, or another destination.

This is because as people become more familiar with the web’s flora, they don’t need search to guide them to obvious destinations. So they use search increasingly for eclectic reasons and less for simple navigation.

[...]In 2005, the top 5 sites were all search engines. Now there are only 4 search engines in the top 10, and we have 4 social networking sites (none in 2005).

As much as I dislike having to disagree with him, here are the basic reasons why his facts are off:

  • Graph does not directly show a correlation
  • He’s using Alexa for the data
  • One graph cannot show that the reason for peaked growth is due to a change in search habits

More details after the jump.
Firstly he is incorrect about no networking sites in 2005, at least they were not called social networking sites then. I joined AsianAvenue.com in 2000 and that was comparatively late in joining that service then. So ’social’ networking sites have been around for awhile, just not quite as popular as they are nowadays.

Secondly, I’m not really convinced by the graph he provides:

Alexa Growth

Leaving aside the use Alexa as a trustworthy source of data (who actually installs their bar for them to track this data?), let’s assume that for this case, Alexa is worthwhile. Now, admittedly I am on the fence about the data, but I have to agree that his concept makes sense: The more you hear and see a site in Google for a type of search, the more likely you will stop going to Google and just to the respective site instead for that type of search.

Still, the graph visually does not show a direct correlation between a Google plateau and the rise of other sites; I see Wikipedia rising as Google was. I also see no remarkable improvement in Craiglist after Google’s plateau. A correlative analysis for this will only find a relationship between a change in Google affecting another site’s growth rather than people searching for it enough and going to another site. Therefore a correlation analysis can only show that there is a moderate correlation for Google’s rise as with Wikipedia and Craigslist (eg: Google search results showed a lot of Wikipedia and Craigslist listings and thus more people went there), and a weak one for Facebook and Youtube. Yet, to be able to show a Google plateau relationship with increased usage of the other sites? That would require some econometric analysis to fully speculate without doing some kind of survey research.

Furthermore, even if there was a correlation for Google’s plateau and Wikipedia’s rise, there would need to be data to back up the concept that it was due to people getting used to seeing Wikipedia and going directly there instead of Google’s lack of innovation on how the search engine displays results (it’s still very text based and not widget/web 2.0 friendly for those not using iGoogle).

Mr. Google Sucks has more about the impact of what he calls “destination sites” and is certainly worth a read.