For those of you who don’t know already, the idea for this blog is a result of the BT Young Scientist and Technology Exhibition 2011 (or BTYSTE for short).
I did a project for the prestigious largest of its kind science fair, that has been running for 47 years in Dublin, Ireland.
The project entitled “Investigating the factors of a search engine algorithm” was relatively successful and proved an excellent starting ground for the now larger scale project.
And as you might have guessed I tried to find as many of the 200 main factors the search engines use to determine where any site ranks when you search for something on Google, Yahoo, Bing, Ask, etc.
I used a testing method that has since been called “reverse engineering to Google algorithm” despite the fact that I didn’t know what reverse engineering was before coming up with the testing method. To me it just seemed like the logical way to try and prove certain factors and come up with new ones.
I came up with a list of 157 factors that definitely are, probably are and might be factors and I tried to test to see whether they were or weren’t factors and by default find their weighting in the algorithms too.
The way I went about that was to take an individual factor, let’s say Page Speed (how quick a page loads) and compared the speed a page loads in all of the top 5 results of a Google search compared to the 25th-30th result, over 30 searches per factors.
The results I got back would be an average of how quick a page loads in high ranking sites (they are the ones in the top 5) and an average Page Speed of the sites that didn’t rank so well, low ranking sites in the 25-30 band.
I would compare this average and if there was a difference between high ranking and low ranking then that factor is impacting how that page is ranking in the search engines and as a result it can be confirmed with relatively high certainty that it is a factor and based on the difference between the two types of sites you could determine what the weighting of the factor was when we compared the difference to other factors.
The greater the difference the more impact it was having and therefore the greater the weighting.
Follow all that?
If you did, that’s great, if not your probably in the majority and I’ll summarize it. I created a testing method that could be used to confirm a search engine algorithmic factor, test a new one and return a result as to whether it was a factor and then I could tell you how important that factor is in the algorithm.
Of course there were a number of problems with this testing method.
- The project was done over a space of 2 months with time also spent on preparing presentations, report books and a project diary as well as preparing to spend a week in a hall with 500 other projects presenting my idea to the general public. That meant I didn’t have as much time to focus on the testing part as I would have liked. As a result I only tested 20 factors using this method and came up with a list of 157 factors based on intuition and other lists available.
- I don’t know a programming language capable of doing the testing automatically so I did the testing manually. I am currently learning Python which will allow me to this in the future. Therefore I only tested these 20 factors over 300 web pages, which simply isn’t a large enough sample size.
- The test was simply not scientific enough, I used no proven formula and took only a small sample size.
Despite all these problems the week at BTYSTE was interesting and inspiring, with people even paying for me to email them the list of factors. Essentially people found my half hearted efforts at a project interesting, and so I figured I would develop the project further.
In the 3 days that I spent presenting the project to the public I talked to a number of SEOs and webmasters and I have forged relationships with a lot of important people that will be able to help me as I continue to better the project.
Following the interest at BTYSTE I have been offered a number of jobs and a Google employee and high level programmer have given me advice on how to improve the project.
As a result I have come up with a fully developed system for as accurately as possible testing search engine algorithmic factors.
All in all the BTYSTE was a great week, I enjoyed talking face to face with fellow SEOs and gained a lot of contacts and experience. It also has driven me on to continue with the project, bettering it and maybe even entering next year.