Wednesday, April 25, 2007

Internet Search Fundamentals

The last entry was a side track. I was side tracked by Shally’s comment. What I had on my mind to say was related to internet search fundamentals. I spend a lot of time training and developing training materials on sourcing. It is my belief that keyword development is the area that most sourcers/recruiters have the hardest time with, even though they may not realize. What I mean is that we just start keying in the keywords that are in the job description without understanding how the search engine syntax is going to interpret it.

The idea is “we just put keywords in an out comes the perfect candidate” but it is not that easy. The key to effective queries is understanding how a search engine treats your search terms. This basic knowledge will help you devise more robust queries and revise ineffective ones.

Try running this simple query in google: resume java beans (by the way this is a quick search string for finding Java Developers/designers.

Notice that it produced 938,000 results, before you try the next simple search scan thru the result summaries.

Next try running this query: beans java resume
(notice that the only thing that has changed is the order of the keywords, otherwise we are using the same keywords.)

I ran the both of them a few minutes ago and this second search string brought 859,000 results. Not only was the number of results significantly different but the order totally changed too.

Keyword location is important, order key terms with main subject first, search engines tend to rank documents that match first terms or phrases higher. Remember, the order in which you enter you terms affect both the order and pages that appear in your search results.

Before I log off, I want to briefly point out some keyword strategies. Formulate the scope of the keyword terms within the job description, identify the important concepts, identify search terms to describe those concepts (don’t rely on the keywords given), consider synonyms and variations of those terms, prepare your search logic. I’ll expand on some of these concepts later but if you stick to this you should be well on your way to writing robust effective queries.


Shally said...

The location is important because search strings are solved using the same basic math skills that we all learned in school. The search engine looks at your querry as a math problem and begins to solve it left to right, with the exception (just like in Algebra) of the OR statement in the parenthesis. What happens is that when looking for resume java beans the word resume is the first one picked up and the search engines MUX will only pull the top 999 results from its servers matching the keyword resume, then ask the servers for the top 999 results matching java, and then the top 999 matching beans. From there it picks the top 999 and that is what it shows you - except of course it makes a gross overestimate that if it had found all the pages there would be 938,000 of them. This is incorrect, which is why on the next try it gives a different guess, based on selecting a different order of sets of 999 results.

Shally said...

Oh - and check out my post on Keyword Disambiguation and Mutation Detection - very useful way to help you figure out what in the world the search engine is thinking :)

SourcingCorner said...

Thanks Shally for your very appreciated comments.

Anonymous said...

Hey, I recently added a news widget from to my blog. It shows the latest news, and just took a copy and paste to implement. Might interest you too.