Reno_Chris - 8:28 pm on Aug 26, 2007 (gmt 0)

the Google algorithm is based in the academic discipline of information retrieval, and not in the world of commerce and marketing

Another thing that is often assumed about Google spam fighting algorithms is that they are perfect and 100% accurate. Matt Cutts and others do a wonderful job of promoting the idea that they always catch the "bad guy", and that they are never just flying by the seat of their pants trying to make things work. Respected geniuses never fly by the seat of their pants, right?

Unfortunately, these formulas and algorithms are simply mathematical simulations based on certain assumptions, and simulating real life with mathematic formulas and code is not as easy as it sounds even for the true genius. Sometimes the assumptions and formulas are well tested, sometimes they are not. I really think that folks should view Google algorithms (both spam fighting and site ranking varieties) more in the same light as we view weather prediction algorithms. Sometimes they work accurately, and sometimes they don’t.

We all know of spammy, low quality sites which are ranked highly by G. and we know of examples of high quality good sites which are ranked poorly. These are plain and simply failures of the G. algorithms to accomplish the goals the programmers set out to achieve.

As an example, in November of 2006, I launched a new hobby related website. I had participated for several years on a handful of forums related to that topic, and I had made friends with forum members as well as the forum owners. Since my website was of interest in that hobby, when I launched it, I inserted the URL of the site into my signature on those forums. No problem at all and I received many favorable comments on the site from the folks on those forums. I had naively hoped that those links would help me begin to build traffic to my site. I had not bought links or done anything that is spammy or violated any of G’s quality guidelines – so I had no worries. Unfortunately in the eyes of the G. algorithms, I had done evil and was caught in their net. Over they years, I had made literally thousands of posts on those forums, and what Google saw was a fairly new website that suddenly had over 4000 links pointing to it, all new in a matter of less than 24 hours. Such occurrences are identified by the algorithm as spam, and the site receiving those links is designated as one that should receive very little or no trust. Sites branded with low trust rank get very limited traffic, and the links pointing to them are highly discounted. I have been trying to dig myself out of a “trust hole” ever since.

Google algorithms, while representing a wonderful effort on the part of the programmers and engineers, are often questionably accurate. They may be better than the results of the older generations of search engines, but they are not nearly as perfect as Google’s spokesmen and promoters would have us believe. There is more than plenty of room for someone to come along and build a better mouse trap. Until such time as the algorithms are improved and better search results provided, we will just have to live with the quirks, burps, twitches and other oddities that are served up by the search engines. We have learned of the true reliability of weather predictions, and so in the same way we need to accept as something we cannot really change, the imperfections of Google.

Chris