What is Semantic Search?

Back in the ancient history of the internet (at least five years ago) if you wanted to enter more than one word for a search it was necessary to put a plus sign between the words  You could search for restaurant, but if you wanted to specify the type of restaurant, you’d enter

steak+restaurant

Your search results might include a restaurant, or a shoe repair shop in a foreign country, and possibly a special on oil changes and a way to enhance your love life.

If you entered your search term in quotation marks, the search would only return items containing that exact string of text.  For example, if you hoped for a favorite kind of steak, you might enter the phrase ‘peppercorn steak.’  But that phrase would not find listings that said ‘steak with crushed peppercorns.’  This could be mildly annoying.

This was in the days of the Boolean Search, when queries returned anything that contained your keyword.

I cheered when it became possible to enter strings of search terms without plus signs and quotation marks, and the results actually had something to do with what I wanted.  I was ecstatic when search evolved to Semantic Search.  Semantic means ‘contextual.’

Semantic search is now using artificial intelligence to formulate your personal context to better interpret what you want when you enter a search phrase. Your presence online including your location, established interests, who you interact with and what you write all contribute to giving you more accurate, personally relevant search results.

This is like a new husband who knows something about your tastes, but may still produce truly puzzling birthday gifts. Over time, hubby gets to know more about what you like. Over time, Semantic Search will evolve to allow full queries.

Search will respond to language that is more like everyday conversation:

• How do I…

• Where can I…

• What does [blank] mean…

This is like talking to hubby after coffee, rather than before.  Full sentences are coming, just like when caffeine arrives in the blood stream.

Semantic Search is also evolving to from keyword-based search to concept-based search.This is like when hubby is actually listening to what you say, instead of only hearing bits and pieces.

Imagine if I wanted to know what year the Montreal Canadiens get kicked out of the Stanley Cup play-offs in the first round.  (Although I could just ask hubby.)  In the past I might enter a bunch of terms like ‘Montreal Canadiens’ and ‘Stanley Cup’ and expect results somewhere on page ten.  In the future I’ll be able to enter the question ‘What year the Montreal Canadiens get kicked out of the Stanley Cup play-offs in the first round?’ and the search algorithms will attempt to interpret what I’m actually looking for and give it to me. Sounds like a dream marriage to me.

When I tried this in a little experiment last year, I simply typed  “Senators” as my search query.

The single word “senators” returns information that is based on my location, my social proof and other criteria that Google Search believes is relevant to me. Based on my IP address  and other factors,Google knows I am Canadian,  live in Ottawa and that I talk about my husband being a hockey nut on social sites.

If I lived in California and chatted on Google+ about politics a search for ‘senator’ might return results about Barbara Boxer.

Semantic Search is becoming more user friendly both in interpreting your queries and in providing more natural, contextual results. The bonus for small business is return on your investment in quality content. As Semantic Search evolves you will no longer have to build content around a keyword or two.  The algorithms will be able to interpret your content and help your potential clients find you.