Sunday, March 3, 2013

Thinking Like a Search Engine (part 2)

These diverse results are to be expected. What you do not want, however, is for a number of business opportunity sites, credit card sites, long-distance service sites, and porn sites (at least those that don't have stars named Saturn) to come up also. If those sites come up under a search for Saturn, then your search engine is not very efficient and your visitors will become frustrated and quit using your search engine. (If your visitor loses valuable time needed to write his report on the planet Saturn while plowing through all of these other non-relevant sites, he or she will find another search engine to use next time. If this happens often, there goes your zillion dollars!) You also want a means to identify the most valuable sites and rank them to come up first in the results. So, what you realize is that you need a means to exclude irrelevant sites and to rank relevant sites according to their value. The better you do these things, the more people will like your search engine and the more successful it will be.

The first thing you have to figure out as a search engine operator is how to associate keywords with sites. You will not have time to examine all the sites on the Internet and write keywords for them. You will either have to create a program that can read sites and make the keywords for them, or you will have to get the site publishers to do that themselves. There are at least two ways that you can get the sites' Webmasters to do this. One (employed by Yahoo!) is to make the Webmaster submit keywords in the process of submitting their site to your engine. This does not work for the "spider" engines, which go out and find the pages themselves, however. Since you want as many sites on your engine as possible, you do not want to wait for the Webmasters to submit them. You want to go out and find them. What you could do is support a standard whereby a certain meta tag included in the HTML code of all Web pages contains the keywords the Webmaster thinks are appropriate for his or her site. That is indeed what has happened. HTML supports the meta tag keywords for that very purpose.
When knowledgeable Webmasters build their Websites, they use the meta tag keywords on each page and insert the relevant keywords for that page. The tag looks like this:
<META Name="keywords" Content="saturn, planet, planets, planetary, solar system, astronomy">
Thus, when your search engine indexes sites, it automatically grabs the keywords from the keyword meta tag, and you are good to go. . . Or, are you? What if the Webmasters cheat when coding in their keywords?
Why would Webmasters want to cheat, and how would they go about doing so? Let's switch gears away from our Saturn example for a moment to explore these questions. Although the statistics are now changing, over the last several years the Internet has been used mostly by young men. Thus, the "hot" search terms—the ones most frequently searched for on the search engines—have been things of interest to young males. Because of this, as you might expect, search terms relating to sex and nudity, rock music, famous female stars, and outlaw-type sites (called "warez sites") have been high on the list of popular search terms. Some Webmasters have kept themselves aware of the current popular search terms and used them in their keywords, even though they may not be relevant to their Website. They do this to increase the probability of their site showing up in a search and, therefore, to increase their traffic.

from George Little's Internet Income Course. Register to SFI for free and get immediate access to the complete 79-part course.

No comments:

Post a Comment