Note to all PR people handling the launch of a new search engine. Top priority: avoid over hyped headlines and especially any reference to terms like ‘Google Killer’ or ‘the next big thing’. Of course, that’s easier said than done. The first thing any journalist wants to know when a new search engine launches is ‘what does it do that Google doesn’t?’.
And so it was with the pre-launch hype around Wolfram Alpha this week. It’s not really a search engine at all in the sense of crawling the web for relevant information. It’s a ‘computational knowledge engine’ – which is a phrase that means nothing to most people but one which accurately describes the fact that Wolfram Alpha is designed to ‘compute’ answers to specific questions. There is some debate over how much of this is really computational and how much is querying a database, but that’s another story.
At VCCP Search we’ve had a preview of the product and had a good look around. Problem is we probably shouldn’t have read all those ‘Google Killer’ stories beforehand because that really didn’t help to manage our expectations.
So what’s it like?
It’s pretty slick at some things; it returns some very good data in response to factual questions which is, after all, what it’s designed to do. Questions with a factual answer such as ‘price of oil’, ‘height of Everest’ ‘speed of light’ return some very rich results, as do searches for species of birds or animals and scientific theories. Its geo-targeting and customization of results looks good, you can download results as a PDF, it does on-the-fly conversion, for example, from imperial to metric. The idea of it being ‘computational’ is very interesting, which in theory means that rather than returning a web page which mentions the phrase ‘average rainfall in London during Wimbledon’ it would actually collate the historical average rainfall data in London with the Wimbledon tournament dates and be able to deliver a definitive answer. However, this is where we stray into dangerous territory because this is big ticket stuff and it won’t take much searching to find examples where it will fail to deliver.
Similarly, once you move away from questions to which there is an accepted factual answer and ask questions on which there is any degree of ambiguity things start to get more complicated and the results get much weaker. We searched for ‘ice age’ for example and Wolfram Alpha returned the movie first – hardly appropriate for the academic and scientific community which the product seems tailor made for. Even ‘lunar cycle’ was a failed result.
Wolfram Alpha is obviously a work in progress. I really believe that specialist search products have a big future – whether that’s a technology specialism such as semantics or a vertical specialism in a specific niche or vertical sector. Wolfram Alpha will suffer from the hype and headlines comparing it to Google when, in reality, it’s a very different product with a very different target audience. At a much lower level, another search engine, Hulbee.com also launched this week though with much less hype and looking at the product it’s very easy to see why. Cuil was the next big thing back in the summer of 2008 with all the front page headlines and ‘Google Killer’ claims. It’s currently just inside Alexa.com’s top 15,000 websites and probably hasn’t registered very high on the Richter scale at Mountain View.
So we wish Wolfram Alpha and all these new search products the best of luck. It’s great to see new technologies being developed to keep up the pace of change in our market. But a word to the wise: Cool off on the hype, reign in the PR and whatever you do, don’t claim to be a Google Killer.
Revolution Magazine
