Why Another Search Engine?
We use Google everyday and to a large degree, we’ve come to the conclusion that the results Google returns for our searches are “best in class”. Thousands of engineers high on unlimited caffeine and red bull at the Googleplex are refining that algorithm as we speak, thrwarting spammers, getting rid of phising sites, and improving the relevancy of the results overall. But imagine if we added in a social component that would allow for democratic participation in the rating of results based on human-perceived qualities that machines cannot determine easily — factors like quality of content, grammar and syntax, writing style, and even presentation and design.
Overview of Current Social-Ranking Search Engines
I’ve looked at 2 of the most similar search engines out there - Kratia.com and Wink.com. To a large degree, most people have not heard of either although Wink.com does have good coverage and have recently begun to focus more on the People search aspect of their engine and less on the “thumbs up, thumbs down” ranking system.
However, I think there are some shortcomings to both of these engines which hamper their ability to make socially rated search engine results a success.
- Both of these engine’s ranking system is not persistent as you navigate through the search engine results. Once you leave Wink.com, you have to return to Wink to provide a rating. On Kratia.com, the toolbar is persistent but in my testing on IE7 (it doesn’t work with on FF above 1.5), the toolbar was not aware that I was navigating search results for my search for “cat food” as I moved from site to site.
- There is no apparent effective spam/abuse system built-in. On Wink.com, that may not be as big of a problem because they determine results from a combination of both algorithm score + user ratings. On Kratia.com, their philosophy of “Equality: In Kratia every vote has the same value. There are no distinctions, no classes, no authorities” means that webmasters and spammers have the potential to upvote their own site rankings without penalty.
- One size fits all for voting - In the case of Wink.com, it’s a thumbs up or thumbs down and on Kratia it’s a “Like or Not Like”. Users value different aspects of websites which means that any rating system should consider some broader but logical aspects of a website to rank on such as content, quality of writing, presentation & style (quality of images, videos, etc.).
Key Ideas in this New Search Engine
- Leverage existing Google search results for best-of-breed algorithm base.
- Ratings toolbar should be a no-install to ensure maximum adoption. Can be achieved through “Netscape-like” persistent frame w/ AJAX/Flex UI for seamless unintrusive integration into surfing experience.
- Toolbar must be connected to the search engine to provide for navigation across search results even when a user leaves the main search site. This will ensure that users can provide feedback on the spot and move to the next search result w/o having to return to the search site.
- Votes should not be created equal - Merit based system would reward users who consistently rate close to averages to have their scores weighted more heavily. Spammers and those who abuse the system will have their rating’s relevance weighted downwards across ALL of their ratings, eventually diminishing their influence on the search results.
Why would Users do Something for Nothing?
People are inherently good.
Contributing to a more human search and web experience is truly something that can be achieved easily through the power of social networks.
Everyone wants their voice to be heard.
Everyone wants to be recognized. (We’ll have a Top Users board)
People hate BAD Search Results.
Finally, this would force web designers, businesses and individuals to focus on optimizing for People and not for Bots in creating their web experiences. Concepts such as keyword density, hidden text, black hat, white hat, etc. would become irrelevant. What becomes critical to success is relevant, well-written content that that is presented in a clear, easy-to-read, easy to understand format, with great supporting materials such as photos and videos when appropriate.
the Idea in Graphical Format
