Human powered search. The next big thing?
This week marks the launch of the Search Wikia project. For those that don’t know, this is the new “wiki” based search engine from Jimmy Wales, the founder of Wikipedia.
The idea of humans helping to massage search results is nothing new. It is in fact the way most early search engines including Yahoo! worked. But it has seen a recent surge in attention with entries from Mahalo and ChaCha, and now Search Wikia. I think it is fair to ask, “why?” Generally innovation happens because there is an unaddressed need, a problem that is not solved with any of the current products. I don’t know about you, but I don’t really see a problem with internet search. 9.9999 times out of 10 I find what I’m interested in quickly and painlessly with Google. So what’s the point?
The Mahalo page offers this:
With traditional search engines you need to figure out the right search term and find relevant results within an unorganized list that often contains irrelevant results, spam, and some mediocre sites. With Mahalo, you can enter a simple search term and instantly get an organized page of results that only includes great links.
And ChaCha states:
Our search experts are constantly handpicking cool, hard-to-find sites so you get only the best, most relevant results.
Finally Search Wikia says that their concept is:
…that of trusted user feedback from a community of users acting together in an open, transparent, public way.
So how do they stack up against the 800lb gorilla? I made my own, extremely unscientific test, to find out. My daughter and I are embarking on a restoration of a 1964 Austin Healey Sprite, and I could use some good references. So I searched the three engines above with the term “austin healey sprite restoration” (without quotes) and compared the results to Google. The results? Google has nothing to worry about for a while.
Let’s take a look at the newest kid on the block, Search Wikia. To be fair, they clearly state on their website that they are releasing the first alpha and that “the results are pretty bad” until the community aspect starts to kick in. Still, I think they did OK. There were 128 results returned (manageable) and the top three were at least relevant. there was a personal page, a link to remarkablecars.com and a yahoo group all about sprites. not bad.
Next up was Mahalo. Mahalo has been all over the publicity trail, making a lot of noise. not surprising since it is a Jason Calacanis project. UJason is not known for his subtle approach. Mahalo’s results were abyssmal. there were no real results, just some suggested search groups I could explore. The top result was for a certain sugar water from a large cola producer. The second was a generic category for automotive restoration, but following that link led to nothing useful. The third was Austin Texas travel related. All useless. Mahalo was helpful however, in that if presented Google results right after its own!
Finally I tried ChaCha. ChaCha, like Mahalo, does not tell you how many results it finds. The first three results were “ads by Google” so I ignored them, but at least the first was relevant so they were already ahead of Mahalo. After that the top three links were all relevant, although there were really only two unique links as the second and third were links to the same Haynes restoration manual. The odd thing about ChaCha is that moments later I repeated the search and got a different set of results. OK, maybe that is some AI thing going on but I just thought it was inconsistent.
How did Google do? 64700 results. Unmanageable of course, but the top three were all highly relevant including a link to a parts source, a link to a restoration project photo history, and a link to the spridget webring were many, many other links could be found.
So for this test, Mahalo failed miserably. ChaCha and Search Wikia did OK, but none of them did anything “better” than Google in my opinion. I suppose the test results might be significantly different if I searched for something more topical like politics or celebrities, but I still suspect that within the first page of Google results most of the good stuff would come through. My verdict? Human powered search is far more hype than real. Google works, so why bother? I may play now and then out of curiosity, but nothing more.
