Where should we eat lunch? What should we have for dinner tonight? Recommendation engines try to match our tastes to music, movies and books – why not help us answer these other eternal questions? Here are a few sites that try to do just that, whether it’s through personalized recommendations restaurant reviews, aggregating restaurant reviews or helping us figure out what to make out of our leftover rice and sprouts.
The State of Recommendation Engines is a sponsored content series by BT Buckets, a leader in personalization and onsite behavioral targeting. Check out their solutions.
Personalized Restaurant Recommendations
The number of personalized restaurant recommenders is surprisingly small, and a few services that were in this space have gone away.
It’s a shame because there’s so much potential here, particularly in group decision making. For example, an app that could help two or more people find favorite places in common (based on reviews or even check-ins) would be extremely useful. It seems like something a service like Lunchwalla could easily implement. You could even take that up a notch when traveling – use it to predict whether all or most of the members of a group will like certain places that no member of the group as ever been, based on an array of common preferences. Maybe there are some interesting apps out there we just don’t know about? Anyway, there are a few personalized recommendation services around:
Hunch is a general purpose recommendation engine can recommend restaurants, recipes and more based on how you answer sets of seemingly unrelated questions – like which sock you put on first. It has recommendations for restaurants in several major cities with filters for cuisine, dietary restrictions, price, which meal you’re looking to eat and whether you’d prefer outdoor seating. Personally, I would like to see the ability to filter by neighborhood or specific address. Once you have your recommendations, you can like or dislike them and add reviews of specific places. The current total number of locations seems a bit small right now, but the service shows promise.
LikeMe makes restaurant recommendations based on on the preferences of similar users. We’ve questioned its advertising content before. Whether it has biased reviews or not, it just doesn’t work very well. Adding restaurants and bars that I like is a cumbersome process, and the search algorithm seems weak. For example, searching for “Detour” in “Portland, OR” finds no results in Portland, but “Detour Cafe” yields the desired result. And even though I always specify what city I’m searching in, I constantly get search results in other cities. Also, even having added a few favorite places and connected it with my Facebook page, the people it suggests as being “like me” are all from other cities, making it difficult to find local recommendations. And the auto generated recommendations don’t seem to change based on user feedback, calling further into question the authenticity of the results.
Goodrec gives several filters for restaurants, including the ability to filter by only your friends recommendations. Unfortunately, that means getting your friends to sign-up for Goodrec and reviewing restaurants. If you don’t have any friends using the service, it’s just another Yelp-clone.
Non-Personalized Review Aggregators
You probably don’t need us to tell you about leading review aggregators like Yelp, Urbanspoon and Google Local, but there are a few other players doing some interesting things in this area:
MisoTrendy finds places that are trending on Foursquare not just where the user currently is, but anywhere they might be considering going. Our coverage of MisoTrendy is here.
Foodspotting lets users upload images of individual dishes from various restaurants, recommend them and discuss them. We covered them here.
Bing is working on some new hyperlocal search functions including a Portland, Ore. food cart finder.
Eating In? Find the Perfect Recipe
There are plenty of recipe sites on the Web, but here are a few that pack some extra intelligence to help you find what you’re craving – or at least make a meal out of what you have.
RecipeKey’s search interface, populated with typical ingredients from my bachelor days
RecipeKey and RecipeMatcher let you input as many items from your pantry as you’d like and get recipe recommendations from your list. Both allow you to filter by meal, cuisine, and other factors. They don’t allow you filter by percentage of ingredients you already have for a meal, but do display that information for every recipe.
Example of a RecipeMatcher search result
If you’re less worried about turning your dregs into delights and just stumped about how to satisfy your craving for something, say, “unusual and spicy” Cookthink lets users search for recipes based not only by ingredient, but by “mood” and other descriptors. It could use some quality control in its tagging system, though. I could be wrong, but I’m pretty sure shrimp isn’t vegetarian.
If you’re a wine lover, or want to be, Epicurious has a new feature we covered recently; the site uses Snooth to power a wine pairing recommendation engine.
If All Else Fails
What the F*** Should I Make for Dinner will give you a random recipe selection from recipe sites from around the Web. Hey, sometimes a little serendipity is a good thing. (Warning: strong language.)
Lead image by John Hritz
