Saturday, October 22, 2005

QTSaver Meta Search Research

Now that we have a new Meta Search tool we started checking what its worth.
The tests are not scientific enough - they were meant only to show trends.

We executed 100 queries that we took from a list of names of famous people and from a list of names of diseases and rated the results as:
No results - 1.
Bad – irrelevant -2.
Ok- relevant but not rich enough - 3.
Good- relevant and gives a good picture of the subject – 4.
Very good - relevant and gives an excellent picture of the subject -5.

We checked QTSaver-Wikipedia and QTSaver-Dmoz.

Than we took another list of unique phrases and checked the QTSaver Meta search.

Results show clearly that Metaserch is better than each search engine alone.
(in previous research QTSaver-Google gave 85% relevance. We never checked QTSaver-Yahoo).

In Metaserch there are no cases (in this research) of no-response (while in each of the other engines there are meaningful percentages).

In Meta there are less "bad" results and more "good" and "very good" results.

Sum of Meta Ok+ results is 99% which means that the user gets a reasonable answer for almost every query that he has and in most cases he gets a satisfying answer.

Checking the first 50 queries from this Meta list shows that the different search engines "help" each other in giving the answer, and on 24% of the answers were found in all search engines.

Wikipedia answered 50% of the queries and "raised" the average quality of the results.

Dmoz answered only 32% of the queries.

To sum it all up - the results are better than expected. In other words - I'm happy...

1 comment:

Naveed said...

You seem to be very poetic and artistic in your approach to innovation. I like that :-)
I'll keep visiting your blog, so that I can draw inspiration from you :-)

- Naveed Ahmed (