|
An In-depth Look at Search Engines
How results are ranked
Different search engines rely on proprietary search technology. Not only is each search
engine's index of the Web unique, but the methods they have for finding and ranking hits
in response to a query, will also differ. So many factors are involved in the latter
process that what results is a special recipe for success.
Below is a table which compares some features of the search engines, showing
differences in the way each engine ranks the results, by boosting each factor described in
the table.
| Search Engine |
Meta-tags |
Reviewed status |
Link popularity |
| Altavista |
No |
No |
Yes |
| Google |
No |
No |
Yes |
| Northern Light |
No |
No |
Yes |
What does this mean for a searcher?
Taken together, the differences will results in different results scoring top ranking
with different engines. What a searcher needs to do is to recognise that just
because one engine hasn't returned the desired result in the first page, another search
engine might.
In practice, you can either metasearch using MetaCrawler,
Profusion, Mamma, etc, OR you
can hop from search engine to search engine with the same query. (Partnered search engines
like HotBot and Lycos make this easy by setting up the appropriate link at the bottom of
your results page). Be persistent and flexible with your query!
What does a search engine 'see' on a Web page?
Again, the exact mixture of features that a search engine will 'see' differs from
engine to engine. These features include:
- The full text of the page
- Meta-tags (which can give a description of the contents as well as relevant keywords)
- Alt-text tags (the text which describes an image)
If you are used to using online information databases, you may be used to expecting
that information is tagged to describe the contents. Although this is possible with
Web pages, only around 30% of pages have meta-tags.
|
