[chenli]

I am the person talking about OmniPlaces.com. It's an amazing experience and Paul was right-on with his questions. Given the time limit, I was not able to answer all the questions. Our site provides powerful search features (instant, error tolerant and local), which are important for location-based apps and mobile apps. In addition, we provide aggregated links to other important sites such as Yelp, Foursquare, and Gowalla. Our core value is the backend technology and product developed by Bimaple, which are also ready to be used by other companies.

Chen Li

[durga]

Chen,

A few suggestions: 1. Prepare answers to basic questions about your business before you go on stage/interview. Practice with someone. 2. Actively listen and try to understand what the interviewer is asking for. Don't answer question B when interviewer is asking question A. 3. Think about what use cases your product could quickly become a hit with. Keep some examples ready.

Having said that, I tried out your webapp, and really like the speed of search as well as the quick zoom in and out based on how many search results you found. Get someone to help you with interaction and visual design so you can polish the look, feel and experience. Also try to think through all the steps a user would go through to achieve their goal when they come to your website.

[chenli]

Durga,

Thanks a lot for your suggestions. They are all very valid. (1) I should admit we didn't do any preparation before the interview since we didn't think we could get selected. (2) When I watched the video tonight, I also noticed that I should have let Paul finish his questions before trying to answer them. For Harj, I had a hard time hearing his questions possibly due to the distance, even though you guys could hear him clearly through his microphone. (3) We will improve the UI of our site.

I should emphasize that we currently use OmniPlaces.com to demonstrate our backend technologies. Our main focus is companies who can use our search technologies. Our goal is not to compete with Google Local. We want to develop OmniPlaces as a site that can aggregate information from other sites, such as Yelp and Foursquare, which can provide valuable information to end users.

Thank you.

Chen Li

[spdy]

Did not work for me at all. Tried to search for Starbucks near Hamburg and compared it to Google Maps.

Got nothing on Omniplaces and everything i needed on Google Maps and as pg pointed out on stage you fight against Google in search and there is no compelling feature im even interested in on your site.

Overall it looks like a clone of Google Maps with less features and information sorry to say that.

Maybe adding some social layers (friend recommendations etc.) would help.

[chenli]

So far the backend is using more than 17 million business listings in the US, licensed from Factual.com.

Chen Li

[tathagatadg]

Wondering what is your business model - i.e. how do you plan to make money, given that maps app comes preinstalled and does instant search (I only typed in ghi and the suggestion include Ghirardelli Chocolate Co, Chicago). Though there isn't a link to yelp or yahoo review in the balloon, on selecting it I do get reviews, images, ratings from the site I'd care for. How much more does it add to the user experience by mashing up gowala and foursquare checkin data?

Have you thought of something like bringing together wishlist + local deals for places - say I add to list Ghirardelli and when there's a deal available there - I get notified about the deal. You can provide the businesses with a deal-hype building based on which they can plan their deals ahead of time. More fun stuff can be thought of with these on the user side too ... as which places are getting hotter and might be deal prone soon

[icandoitbetter]

First of all, fuzzy search is no biggie for Google. Google Search Instant works great, and I can't see any reason why the Instant technology wouldn't work just as well for Maps--which leads me to believe that they have a reason for not releasing a Maps Instant. Even if your fuzzy learning algorithm is superior to Google's, Google has way more data, and more data > better algorithm anytime.

Your backend technology may have promise, but you probably need to apply it on a more niche and less traversed market than location search, one that BigCos like Google wouldn't consider.

In any case, good luck!

[chenli]

Agreed completely. I am glad you noticed the value of our backend technologies.

[benmccann]

Hi Dr. Li, I saw your presentation on BiMaple at the Irvine Incubation Center, so it's nice to see the technology out in the wild. My initial reaction to omniplaces was that it was slightly harder to use than Google Maps because it does not guess at your location based on your IP address. I think if you implement that it will be easier to use than Google Maps due to the instant fuzzy search (though as Paul Graham said, just being slightly better might not be enough.) Good luck!

[lawnchair_larry]

I tried it. To me it felt like a different skin to google maps. The only difference I noticed was that it provides the icons under each place that link to yelp etc. That feature would be much more useful if it could pull the data instead of link you off site.

Google's big advantage is that I can click "directions to here", which would prevent me from switching. If you are targeting mobile devices, this is especially important.

[chenli]

Please check my reply to Durga's message. We use OmniPlaces.com to demonstrate our backend technologies, which are our main focus currently.

[joshmlewis]

I check it out. It worked nice but still needs something to set it apart I think. Even if you have something that's a little better than Google I might be willing to still choose Google over OmniPlaces just because I'm already familiar with Google and probably already have it open. Just a thought.

[suking]

Take his advice and do something different. Google Local is very good.