Bing expands natural language search: query by price

Today Bing is unveiling some improvements to the way you can search using natural language queries, adding the ability to query by price.  For example, a query for “laptops under $500” will filter out any higher price laptops, eliminating the step of clicking through to a price range in Bing Shopping.

The new feature, only available for price queries (and, need we say it, only in the US) for now, is set to be expanded to more types of queries.  Bing, working with Microsoft Research, is implementing the feature in a way that will make it easy to scale, adding in more types of queries, in more regions, soon.

So what does this mean?  You can of course type in “laptops under $500” to Google, but what you get are keyword generated results:

google-natural-query_thumb Bing

Note that while you get keyword based results (presumably from sites that have targeted the keywords “laptops under $500”), the shopping results (the images at the bottom of the screenshot) are not filtered by price, in fact none of the laptops featured are under $500.

Now on to Bing – same query:

bing-natural-query_thumb Bing

Bing shows some of the same results as Google, but notice here that all of the shopping results are filtered by price.  Clicking through to the “Shop for laptops under 500” link, not only are they filtered, but the pricing choices in the left hand bar are broken down by price under $500, too.

Does saving one click to sort by price make that much difference?  Perhaps not, but looking at the numbers of ads purchased based on a natural language query like “laptops under 500”, it’s clear that searchers are using those types of queries to find what they’re looking for.  Bing has taken a first step in trying to better serve up those results, in a way that is machine learnable and scales.

These types of queries also can help to drive traffic to Bing Shopping (and other verticals), where Microsoft is making a concerted effort to provide more than “10 blue links” into their search results. Very interesting!