Loku location discovery app knows hip from haute

There are probably hundreds of apps and sites that will help you find good places to eat and cool places to hang out. And then there's that gorilla, Yelp. But even in this crowded space, the new Loku has a secret sauce that works. It's too bad the best part of the site is not obvious, and that there's not yet a smartphone app from this company. But this is still a good site to get to know.

Loku finds news and recommends places in neighborhoods based on its algorithms that apply natural language processing and sentiment analysis to the reviews, posts, and social media updates it reads.

For example, when looking for information related to a neighborhood like The Mission in San Francisco, which is young and more artsy than, say, Nob Hill, it will give weight to stories that have language that resonates with the neighborhood vibe. "Rad" in a user review of a restaurant will help Loku place it as place good for people visiting the neighborhood to consider, CEO Dan Street says. But if you're looking at the swanky Nob Hill, the site will give more weight to stories that talk about how elegant a place is.

You can see the algorithms working hard under Loku's "Know Local" tab, where the it does a good job of creating neighborhood portals containing news, local deals, and even police reports.

But the real gem of this service is the non-default "Go Local" tab, where you can select your mood using sliders on a few nonstandard but intuitive scales like "Local," "Foodie," "Flirty," and "Swanky." You also select a neighborhood, and Loku matches your feelings against its sentiment analysis of various food reviews. I found that it does a good job of finding restaurants that mesh with what I was looking for, but more importantly is does a great job of providing a different view on options, with different suggestions, than you'll get anywhere else.

Loku is a strong service for anyone visiting one of the 15 U.S. cities it has data on. Unfortunately it's a Web-only service right now. There's no app, although one is coming. Why didn't the team do the mobile app first? "It's easier to create a lot of changes on the Web," Street told me. He's been working on this app for three years, and has changed directions with it a few times. Recently, he said, the open-source community has solved problems in three of the areas his team was struggling with: natural-language processing, ranking of data, and presentation of it. "A lot of the work we did ourselves a few years ago is now obsolete. But now that we've locked down the secret sauce and the recipe, we'll make a mobile app."

The business for Loku is an emerging standard for venue-finding services. It aggregates deals and offers from other sources, including consumer facing services like Groupon and Livingsocial, as well as from white-label deal networks like OfferEx. Street says it makes the business side the Loku equation very straightforward and effective. "There's no reason to have a sales force," he says.

That last bit may be true for now, but when there's no sales force, there's also nobody between your company and declining per-transaction revenues, so I might encourage him to find a model where Loku has more control of its own destiny. But for now, the offers aggregation play does seem to make sense.

• Product quality: Three out of five stars. The core algorithms for recommending venues are unique and very good. But the gem of the service is hidden, and the lack of a mobile app is a real problem.
• Business quality: Four out of five. There's always good money to be scraped off the table when you're recommending restaurants or venues to consumers. I would like to see the company come up with a business model that it has better control over, however.