In one of his postings, Martinaddressed a subject-matter that I had heard a lot of rumors and speculations about, but which i never really got into before: Ask.com does not have its own index anymore, but is using data that is crawled by Google and is then sorted with their own algorithms. On the one hand, this would be beneficial for Ask because they could save time and money by not having to choose which websites are suitable and crawling them anymore. At the same time, Google is keeping a benign competitor alive, which will let them point to the working competition in the searchenginemarket when this comes up in the hearings and investigations that are likely to be held in the years to come. Another thing that helps strengthen this theory is the fact that Ask is being given a leeway by Google in their arbitrage which is not given to any other competitor.

If you crawl Asks index for certain key-words (this is something that Stefan noted in one of the comments to Martins posting), then you will quickly notice, that the index really is filled with Google-data. There are no signs, that their own crawler is contributing any new information to these results.

Whats interesting here, is that Google only passes on sites to Ask, that come from its primary index (those of you that follow this blog for a while will remember). Seeing how you are usually rather limited in crawling Google directly for these sites, which will then also be quite unstable, this might turn out to be a useful application. To test for this, I checked and compared the number of indexed sites for 1.000 domains at both Google as well as at Ask.com:

Even though there are a number of domains, where the ratio of sites from the Google index as well as from the Ask.com-(primary)-index coincide with my expectations (ciao.de: 38%, kelkoo.de: 17%, mister-wong.de: 5%, shortnews.de: 2%), there are also many domains that cannot be right (wikipedia.org: 0,8%, yahoo.com: 0,03%). In these cases, I would fancy, that the Google numbers “turned out” to be much to high, which would render this comparison pointless. For a while now, I will regularly keep an eye on those Ask.com numbers and maybe I will find some value in those numbers.