Yesterday, the New York Times picked up on an essay written by Oren Etzioni, in which the computer scientist suggests that the keyword searches of today could, and should one day be extinct. As true as that may be, that day still doesn’t look immediately imminent.
Etzioni espoused the virtues of “natural-language searching and answering”. In many quarters, this has become known as the Semantic Web. It’s not a new theory, and has been something that has been discussed at length here and elsewhere. So why is it still only theoretical and, perhaps just as importantly, newsworthy?
Keyword search has a purpose. Whilst it is relatively complicated, it’s still far from perfect. SEO probably has a fair amount to answer for in this regard, but ultimately it is the furthest extent of the technology and knowledge that we have currently.
Small steps, such as Schema.org, have helped to push us towards better understanding of semantics within content. Unfortunately, search engine spiders still aren’t intelligent enough to digest the entire context of every page they index though, which is why keyword searches are still very much in vogue.
The integration of real-time data also helps to provide greater relevance within SERPs. With Google developing the Plus project, including both the social bookmarking and networking facets, it will invariably have even more data at its disposal. Theoretically, the more knowledge they amass, the greater their understanding of search intentions.
A smarter search system will have undoubted benefits. Questions can be answered instantly be reputable sources, cutting out much of the fluff that bogs down most queries. However, beneath it all, there will still need to be some form of complex algorithm to calculate where everything fits in. Google can’t provide all the answers on its own (not yet at least) so it will always have to funnel traffic through to secondary sites within its rankings.
Where there’s an algorithm of course, there will always be optimisation. But in a world where quality and relevance are key to dominating the top spot, as would (presumably) be the case if we were to adopt a semantic style search, the onus may move sharply from links and towards content.
There are search engines that provide answers, as I’ve covered many times before. You only need to look at Wolfram Alpha to see how connections between words and equations are then turned into a meaningful answer. But in order to become a truly semantic or next level search solution, the intention of a search would need to be identified and an accurate response found – this is where the difficulty lies.
Keyword based search is essentially a kind of educated guess. It’s not perfect and never will be. The author of the essay mentioned previously is actually involved in developing the technology that could one day drive this kind of logical search. Therefore he is speaking from experience and personal knowledge. But it isn’t an overnight solution. There isn’t a timeline and we don’t even know who the pioneer will be.
Whilst the assumption would be that Google or Microsoft will power the semantic web (more likely the former), there’s every chance that a new group will come along, just like Google did in the 90’s and blow the whole thing apart. There have been big strides in recent months and years within the development of search and intelligence therein, but this simply represents the tip of the iceberg.
But if you’ve invested in SEO, targeting keywords and rankings, you’d be well advised not to panic. Whilst a fast, semantic search might be the end goal, it is still some way away from being achieved. And whilst it may be able to answer questions, it probably will never understand “where can I get this DVD for the cheapest price with next day delivery?” or “what’s the best restaurant in Kettering?” As mentioned, keywords serve a purpose, and whilst algorithms need refinement, they won’t be defunct for quite some time.
Question Key via BigStock