Posted on
February 23rd, 2012

Can Search Engine Results be Trusted?

Is it really possible to accurately measure the popularity of someone or something by the number of results generated via a Google search on the internet?

A simple name search, run on three different computers in the same office and within feet of each other, brought up figures varying from 285,000, to 325,000 to a massive 835,000 results. Try this search using your own name in your own office, on different computers and see what happens.

What you have to understand is that there are many factors which will determine the results that Google will produce. Did you know for instance that there are a number of copies of Google across the world and the version that is the quietest at the time your search is made will be given the task of executing your request? The computer you are using will have an impact on your results too, as will any previous searches you have made – results are personalised and even where you are based will have an effect.

Estimations

If a search comes back stating there are 57 million pages to be viewed, we would tend to believe the results and we aren’t likely to start counting but this data is not necessarily accurate.

A recent study ran a query across three different search engines which brought up less than 1000 results. It showed that all the search engines were only providing estimates and even the best search engine was approximate in its results.

Researchers discovered that searches for one word queries were fairly accurate but the more search terms that were used, the less accurate the results became – a one-word search provided results with only 10% of errors. When a two-word query was run, the accuracy reduced by almost 50%. Even the highest performing search engine couldn’t correctly match documents within 10% of the actual number, more than 20% of the time.

Experts believe that it is due to the speed that the results are returned that means users put up with this inaccuracy. The computers that run the queries behind the scenes have to work incredibly quickly, ignoring spam and dismissing duplicate pages – and all of this is done in an instant. Bearing in mind how big the web is to search, the engines can only cover a small percentage of it.

There is no definitive list of all of the web pages, or even the websites, in the world and the search engines could search forever and not find every web page that is out there. In a nutshell, the search engines will begin by looking at a few of the bigger websites, mostly by following links on well established sites, which means if you have a reasonably new website that hasn’t been linked to yet, then it is highly likely that Google wont even know of your existence.

< Back to Blog