Beta Labs - Yellowpages.com

Using Distance Bands to Rank Search Results for Business Categories

Today we’re revealing a sneak peak at another twist to our search relevance algorithm — one that strikes a balance between our traditional sponsored listings approach and pure distance-based sorting, both of which you can use on YELLOWPAGES.COM today.

The new approach is simple — we take the listings that would normally be returned by a search for a business category on YELLOWPAGES.COM, and group them into five distance bands, or rings, based on their proximity to the geographic center of the search. (We already do something similar for searches for business names; we’re experimenting with this new approach only for business/product categories.)

So if you search for “lighting fixtures” in Fremont, CA, you first see the businesses that are located within 1 mile of the center of Fremont, then the businesses within 3 miles, then 5 miles, then 10, and finally within 20 miles.

By treating all businesses within a given band as equally close to the target location, we eliminate the effect of distance within the band. So a listing that’s 4.9 miles away might appear higher in the results than one that’s 3.1 miles away, but neither of them would appear before one that’s 2.5 miles away. We can think of this as a technique for reducing the impact of small differences in distance, where the definition of “small” changes with each band. (As the distance from the center-point of the search increases, the sensitivity to differences in distance decreases. Intuitively, it’s easy to understand that the difference between 0.4 miles and 1.4 miles is much greater than the difference between 10.4 and 11.4 miles).

This approach isn’t perfect. For some business categories — particularly service-oriented ones — geographic proximity isn’t much of an issue, so it’s inappropriate to put *any* significant weight on distance when ordering the results. In others, we believe this “compromise” sorting method may do a good job of serving the needs of both users and advertisers. As always, please use the comments section below to let is know whether you think it works, and why or why not!

Speak4it simplifies local search

The AT&T Interactive R&D team, in cooperation with our good friends at AT&T Labs, recently launched an iPhone application called Speak4it. Speak4it is a voice-based local search application that makes it really easy to find things in your local area. What makes Speak4it different from other speech recognition apps is that it is designed specifically for both local and mobile. So if you are out and about, and you need something nearby, Speak4it will get you what you need as easily as possible.

How it works: launch the app, push, speak and voila! Results around you will appear, and you can view them as a list or on a map. You can even see where you are relative to the results. If you want to search in a different location, you can do that too. For example, you can say “Pizza in New York.” Try it out.

We’re using the AT&T Watson speech recognition engine, and the more you use it, the better it gets. So keep using it, and send us feedback, LOTS of feedback. In fact, you can send us feedback right from the phone. We want to hear about what you like, what you don’t like, and also where you would like to see us take this next.

Check back often, as we will be launching more apps soon. In the meantime, enjoy Speak4it.

When Nature Calls

In addition to Speak4it, AT&T Interactive R&D has created and released another iPhone application called Have2P. From the name, you can probably figure out what this free app is all about…

Yep, you guessed it. It’s a restroom locator. Even we are not immune to the callings of Mother Nature and felt that something like this would be useful for everyone. So we created this specialized local search application to help those in times of need.

The application uses the GPS function on your iPhone to identify where you are and automatically searches the restroom database, which currently covers the U.S. Search results allow you to view and update the details about a restroom and add your own comments.

In order for this application to reach its potential, we need your help in collecting and refining the information.

Pay it forward! One day, you’ll be glad that someone else has too.

Alternative Local Search Algorithm Designs

Our Search Engineering team has been researching alternative algorithms for delivering highly relevant local search results for web and mobile users, and we’re pretty excited about our progress.

The existing approaches, in a simplified sense, rely upon matching keywords and/or category names based on just the name, keyword and category data associated with the available business listings. While we’ve implemented a fair number of mechanisms to identify near matches, related matches and possible disambiguation choices, there were things that we felt we could do to search our data in a more effective way.

This current round of development has yielded an approach based on a couple of interesting methods: Free Search and Geo Density. Free Search is our TF-IDF (Term Frequency and Inverse Document Frequency) based algorithm that includes all available data in our listing index and applies weights to fields based on our internal search logic. This gives our searches much greater penetration into our data set than the existing search method, providing for an information retrieval method that is more responsive to how people might casually describe what they are searching for.

If Free Search represents our approach to delivering a better “what” component of our searches, Geo Density is our new approach to improving the “where” component. When we build our indexes of our listing data, we are also calculating the relative density of various listing types in the named geographies we can search in. For this beta example, we are calculating these densities against the geography of California, but we will soon expand this to support a much more granular breakdown of density of information (i.e., the search model will be able to understand that the density of coffee shops in Santa Monica, CA is different than the density of coffee shops in Mojave, CA). This density value allows us to dynamically set the radius of our search and apply a weighting (not a hard filter) to results within that radius.

This sort of smart and dynamic distance determination is key to a friendly local search system, as it provides a more intuitive scaling of the results, based on the availability of the thing being searched for and the specificity of the search.

For example, a search for “furniture” in Pasadena, CA will return a focused and relevant list of results in and around Pasadena, while a search for “IKEA” in Pasadena, CA will automatically scale out to include more IKEA stores in the results. The search system can do this quickly and efficiently because it understands the relative density of stores of this type in California and can infer the specificity of the search from the search term used.

As another example, take this search for “coffee shop” in Pasadena, CA and compare it to a similar search for “Starbucks” in Pasadena, CA. Both searches remain well focused on the intended geography, though the area will change slightly due to the specificity of the search term used.

It is important to note that, as can be seen in the examples above, businesses of related categories sometimes occur in the result sets. This is due to the way the system includes and weights the first few categories associated with the initial query. This expansion of the search domain is an experiment to determine whether or not we can provide a more intuitive scope for searches, and will be evolving as we update the system in the near future.

We will be updating the engine in the near future to increase its accuracy and sensitivity, so please leave us feedback and check back soon for an update.

Note: Only California listings are included in our testing at this time.