Maybe not, but they are certainly making the problem easier. The basic premise – at some point, someone has posed the same question as you, posted it on the internet, and someone else has answered it. So, why not just google for your question?
The concept of natural language search has been around for a while. A perfect natural search engine would allow you to ask a question as if you were speaking to a human. Ambiguous words (such as the notorious “apple”) would be resolved by their context in the question. It’s difficult problem to solve well. There are many sites out there that have attacked it, Ask.com for one. Few are household names as a result.
Over the last decade, my consumer search techniques have evolved along with search technology. Typically, I envision the document I’m trying find (hoping it exists), boil it down to a few keywords, and try entering different permutations into Google. Lately however, I find myself skipping the first few steps and just typing a natural language question into the search box. Crazy, I know, but it’s actually becoming more and more effective and it’s not because Google released some new natural language tech (though they may have and I’m just not aware of it).
Q&A sites are popping up everywhere and their content and traffic are growing rapidly. These Q&A sites often function in one particular vertical (Avvo Answers) but the big boys are attempting to win across many categories, like Yahoo Answers. They allow users to post questions publicly and other users answer them publicly (usually for some kind of ego-boo, “points”, or just for fun). The result is a very concise page around one very specific topic (in effect, search bot candy.). The H1 is the question, the page title is the question, the meta description is the question, the details are loaded with keywords and all in the first 100 words of the page. If someone else ever searches for this specific question again (or some variation), this page is sure to be in the top 10. Now, repeat that process 10 million times or 100 million times. Toss even the weakest search engine relevancy on top (Google is more than sufficient) and you’ve got pretty good natural language search simply by the Law of Large Numbers.
It’s important to note that these are “long tail” pages, meaning that generally speaking, no one page receives at ton of traffic, but in aggregate, a site can generate a ton of highly targeted traffic. The level of targeting is also important to note. We recently received a visitor who came from Google with this query, what are the consequences of being charged with a battery?. There’s a pretty good chance that this person will be looking for a criminal defense lawyer in the near future.
So, where will this all take us? Let’s say we have a few hundred million answered questions floating around the web, indexed by search engines. “How will that change my business?”, you may ask.
Two ways:
- My guess is that we will start to see this change user search behavior. People will slowly start to move to natural language searches as opposed to keywords searches because, well, it’s easier. It takes less brain effort to blindly enter a question than to identify keywords. And it’s becoming more and more effective as the collective content grows.
- Users will become accustomed to Q&A sites turning up in their search results. Consequently overall importance (accumulatively) and popularity of Q&A sites will grow. Like most of the web, it’s a feedback loop.
Now, this whole transition won’t happen with a bang. There won’t be a revolution of natural search; it will happen slowly and quietly as users are retrained on searching the web.
“So what?”, you are thinking. “I’m a lawyer. What do I care about natural language search or Q&A sites?” Time to put your marketing hat on, but I’ll leave that for a future post. Here are a few examples to get the juices flowing.
What are the penalties of a car accident without insurance?
When can i remove spouse from my health coverage?
What are my responsibilities as trustee of my fathers estate?
PS: Granted, natural search will continue to be an important problem to crack. For instance, temporal questions like “Who beat the Lakers yesterday?” and “What will the weather be like tomorrow?” are especially tricky and are not well served by the brute-force approach. Static pages won’t suffice for questions like these; a true solution is required. But there is a huge subset of natural language questions that can be well answered by the content in Q&A sites.