Beyond Search is Suggestion

In our previous posts we discussed how the lack of personalization and lack of incentives for a review to be honest and fair, reduced the value of local search reviews by themselves. At theSUGGESTR.com we’ve decided to use a technique called collaborative filtering to provide a more personalized experience by being able to provide personalized suggestions not just search results. Our suggestion engine starts by calculating similarity scores between every pair of users and every pair of businesses. This allows us to ’score’ how similar the taste’s of two users are as well as score how similar two businesses are. These scores are then used in our proprietary algorithm to predict how well our users will like a given business. We believe, as do others, that the future of local search is more of a personalized experience and if you can personalize searches then you begin to enter the realm of suggestion.

Let’s look at a simple example of how this may work. Let’s say User A has rated 5 restaurants in New York, and User B has rated the same 5 restaurants with the SAME SCORE as User A. That would make User A and B have a high similarity score. Now let’s say User A from NY is going to San Francisco for the first time, but User B has been there, theSUGGESTR.com will be able to suggest restaurants to User A based on User B’s rating. Now imagine that process for every combination of users and businesses in our database. The data is combined using our unique algorithm so that we can accurately predict your ratings.

Collaborative Filtering in Action

We realize that our vision of local suggestion will be a marathon not a sprint. We do think we’ve made some great initial steps with features like our similarity meter which allows registered users to quickly see how similar their tastes are with others. Imagine being able to quickly identify which reviews come from people with similar tastes to you.

Also our predictive search allows us to predict how much you’ll enjoy a given business and even explain WHY we’re making the prediction.

We know their are lots of improvements still to be made and we should have some great new features available at theSUGGESTR.com shortly! In the meantime let us know what you think we should do to improve our site.

Share This

Tags: Local Search, Recommendations, Suggestions, theSUGGESTR

What’s Wrong with Local Search?

The local search market has become saturated with sites that basically do the same thing. They allow users to comment on places, give them ratings, sometimes tag them and if your lucky provide additional data that is useful about a local business. While this does generate a ton of content the problem is the quality and accuracy of the content is questionable at best. Often a review will contain many contradicting opinions, and why shouldn’t it, everyone’s tastes and experiences are different. My opinion of a great steak house is probably different than that a vegetarian.

There are lots of reasons why a person may like or dislike a given business, especially restaurants. The majority of the current local search sites don’t provide any personalization when it comes to the searches and results performed by their users. Two users with opposite tastes will get the same results if they enter the same search. Even more troubling is when reviewing the 100’s of reviews posted for some of the most popular business the user is often left to decide if they want to read the one that says that the restaurant is the “best place ever” or the one that says that the “service was horrible.”

Most sites do not provide any way to easily distinguish which user’s reviews are most likely to be shared by you! We believe, as do others, that the future of local search is not really search at all but personalized suggestions.

We’ll lay out our vision for local suggestions in a couple of days.

Share This

Tags: Local Search, Recommendations, Suggestions, theSUGGESTR