QueryCat is a search engine that is designed to search frequently asked questions, and they claim to have the largest database of these, by using the Alexa web crawler. They have apparently indexed over 2 million questions and answers (though this information isn’t on the website) which is a fair number.
I tried lots of questions with the service and had very mixed results. ‘What does RSS mean?’ came back with a good response, but ‘How tall is Nelson’s column’ or ‘How many died in the American Civil War?’ worked rather less well, and I got much better results by repeating those questions on Google, Yahoo, Live and Ask. While it was possible to get good results from QueryCAT this variability is problematical – if I can get a good response from a traditional search engine, why bother with this one?
The problem of course is that the search engine is only as good as the sources that it has available. While a database of frequently asked questions is a good concept, and can be used to create an interesting vertical search engine, the questions and more importantly the answers have to come from trusted sources.
Many of the links provided by QueryCAT came from FAQs provided by commercial organizations, which inevitably raises questions about their particular biases. Search engines that take their data from a much wider variety of sources may well be able to provide a more rounded approach to query answering.
While it’s always a case of ‘buyer beware’ when looking for answers on the net, I think this is particularly the case with quick answers to quick questions – I would want a broad range of responses, rather than those from a rather narrower set, even if that set is 2 million questions large.
In summary, I think that QueryCAT would be a reasonable backup search engine if you were unable to get an answer from the more traditional search engines, but its variability means you’d be a little lucky to get an answer immediately if it was your first port of call.
Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.