In recent times, the hints are concentrating that Google has changed some parameters for the filter that is supposed to detect similar content or alternatively they could have introduced a new method of identification. It seems that many projects that have numerous similar or identical data are affected. We can take price-comparisons as a prime example for this: In the end, all of them have mostly the same or quite similar product-data, the number of shops in Germany is also limited and as far as the 20 cheapest are concerned you will usually also have the same shops that play a big part.

If, for example, we take a look at Preistrend.de, we can see in the Alexanumbers that since the end of January, the traffic has seriously decreased – additionally, the site will often be found far to the back in the SERPs. Google's statement for the site-query for Ciao.de does not really look that good either: Google only recognizes two pages with unique content, the other nearly 2,5 million pages are only listed thereafter.

The speculation in US-forums is that this could be connected with a new Google-patent. Mid-last year, the Google-employee Anna Lynn Patterson submitted the patent with the title “Detecting spam documents in a phrase based information retrieval system“ which describes how the spamdetection could be improved. It would detect phrases that are combined with other text-blocks that exist on a page and would save it together with the document. If the same combination of text-blocks is detected on multiple pages, they can assume that it is due to spam. This could be the case for the aforementioned price-comparisons in general: product informations like weight, dimensions or color as well as name of the shop and description thereof should be similar on many price-comparison websites and would therefore fall victim to this filter – if it really is in use ...