Ever since it launched, Google has done a pretty good job of delivering on its mission promise ‘to organise the world’s information’ and we’ve been thankful for it. No-one wants to go back to the pre-Google days, where the likes of AOL and AltaVista were serving up almost totally irrelevant, messy search results.
So, what’s the latest development to come out of those brainy engineer types at the Googleplex in California? ‘Semantic search’ – a clever new service which allows Google to give you results based solely on the text/keywords that you put into the search query. In effect, Google gives you its best guess as to what your words mean.
However, even with Google’s complex algorithm, this guess isn’t always contextually correct. Enter Google’s artificial intelligence development that it has been working hard at, in order to understand your search intent and the meaning of the query you’ve entered rather than parsing through keywords like a dictionary.
With semantic search, Google will peer deeply into the relationship between the query words and how they work together in an attempt to understand what those words mean in that holistic context. Google will understand that when “New” and “York” are placed together, it changes the meaning. It will also continue to learn from your previous searches and no doubt integrate what it can find out about you from Google+.
Semantic search isn’t a new concept. Back in 2008, niche search engines were popping up that focused on natural language processing over reading keywords, but it remained just that – a niche interest. But in 2012 we can see the likes of Apple’s Siri possibly revolutionising the way we search, and as ever, Google has an answer to this with Google Assistant, its Android challenger. And now, with semantic search testing underway ahead of a full launch, we are paying attention with some excitement and anticipation.
Google’s acquisition in 2010 of start-up Metaweb Technologies is the engine of this new development. When acquired, Metaweb had an index of some 12m ‘entities’, such as movies, books, companies, celebrities, metals, wines and cars. By comparison, online encyclopaedia Wikipedia has just 3.5m English entries. Now, Google’s expanded entity set has swollen to more than 200m, partly by developing “extraction algorithms,” that can organise big data scattered across the web. They call it the ‘knowledge graph’, and this will be at the heart of the rollout of the new type of results.
What does this mean for brands? Commentators suggest that 15% to 20% of searches could be affected, and that the likes of Wikipedia and Answers.com will lose out with Google’s semantic results taking a higher ranking slot. Naturally, one would expect Google to be looking at the monetisation angle, such as PPC results within the semantic results space.
The SEO community is seeing this as a forthcoming triumph for real, original and compelling content written from a user perspective, not just for search spiders (in the old sense). As Erin Everhart observes in her eConsultancy post, brands will not just be competing against each other (in an SEO sense) for rankings, rather they will be going head to head with Google itself. Google really is taking the organising bit to a new extreme.