Glancee takes up arms in the SXSW people discovery app wars

AUSTIN, Texas--Walking around the halls at the South by Southwest (SXSW) festival here this week, it seems that almost everyone is looking at their phones. But while plenty are reading the latest tweets, checking in on Foursquare, or updating their friends through GroupMe, there's a whole new category of services that has captured people's attention.

Known as people discovery apps, these tools are helping SXSW attendees make new connections. Tied to users' locations, the apps identify where they are and look for matches with others nearby, mainly by identifying common friends or interests on Facebook.

Leading the pack in this new category are Highlight, a Benchmark Capital funded startup, Kismet, and Glancee, a bootstrapped venture led by Italian import Andrea Vaccari.

Though Highlight has perhaps gotten the lion's share of the press, Glancee has been around longer, having first launched last June. But none of that matters now that it's time for SXSW: regardless of how many passionate users you have in San Francisco or New York, there's no substitution for the concentration of early adopters these services can attract in Austin--and as a result, it's almost make or break for these young companies here in the Texas capital.

As the battle for people discovery supremacy heats up as thousands roll into town, Vaccari spoke to CNET about Glancee's past, about how it is going about winning over users, and about where it's likely to go in the future.

Q: Explain Glancee.
Andrea Vaccari: Glancee's very simple. It looks for people nearby who have friends in common with you or who share the same interests. And it makes it easy to discover the same connections and start a conversation with them. So it's easy to find someone new that you don't know but that we think you would enjoy meeting.

Version 2.0 came out [yesterday], both for iPhone and for Android. We improved our speed, which was a big issue for power users with lots of friends and interests. There's also a new diary feature. Before, Glancee was about who was right here right now--people that are near you in the moment you open the app. But you were missing all the connections with all the people you cross paths with during the day but didn't have time to talk to because you were busy doing other things. So now we record all those connections and put them in the diaries. So if you're running around at SXSW, we record all the memories of all the most interesting people you were close to you and when you fly back home after SXSW, you can connect with all the people that you didn't meet but you should have.

How does Glance decide if someone's an appropriate match?
Vaccari: If you have friends in common, that's a good signal. The other thing is whether you work in the same company or you studied in the same university or if you have interests that match one another. Unlike many other competitors, that just look for exact matches between Facebook likes, where you have to actually like the same page, which is problematic because there are often multiple pages about the same concept. There are multiple pages about Apple, for example. We also try to find relevant interests that are not exactly the same, and we use Wikipedia to do that. So let's say you like Steve Jobs and I like Apple, we're going to find the Wikipedia pages related to these two concepts and if the two pages link to each other or if there is a third page that links to both then that's a good signal that the two concepts are related.

It's really more about finding an ice-breaker rather than this is the person that you should marry or work with for the rest of your life.

Where did the idea for Glancee come from?
Vaccari: In five years in the U.S., I lived in five different cities. Every time I move to a new city, I have to build a new social circle. It's a slow process, and a lot of times, you end up talking to someone for half an hour and realize you have nothing in common or that the person isn't really interesting for you. At the same time, maybe right next to them is another person who's really relevant to you.

Also, sometimes I meet someone and only after 20 minutes of talking do we realize that we have something really important in common. There is this way of finding hidden connections that are relevant that would be otherwise impossible or very unlikely normally, and I thought that these kind of situations are not as rare as we think they are and there must be an elegant way to reveal them that is not too invasive.

You're bootstrapped?
Vaccari: We are. We started playing with the idea in June 2010. Originally, we wanted to do something even more difficult: We wanted to recommend places you should go based on who's inside rather than on the food, like Yelp does. But we realized it's really difficult to understand where people are inside unless people check in. But check-ins are not as popular as we think, because only a few people actually do them a lot. So we switched to this idea and started to develop the recommendation engine and the location engine, which is the only one that does updates without running the battery on your phone. We released Glancee last June at Foo Camp. We've done some small iterations, trying to find the right balance between letting early adopters interact with each other right away but giving them the feeling that they cannot be stalked, and they still have control over their privacy.

How do you do battery optimization?
Vaccari: We have two key technologies that as far as we know, no one else has. First is the recommendation engine. And the second is the location engine. It took a long time for us to develop a smart location engine. What we can do is know where we are with a very good accuracy but without draining your phone's battery. And we did it using a series of improvements, both on the algorithms, and on the way we communicate with our servers. The most interesting bit is that our updates are not constant. They depend on whether you're moving or you're sitting in the same place, and on your history. If you're always someplace from 9 to 5, then it's very likely that you're always going to be there from 9 to 5 so we don't have to ping you every two minutes to see if you're there. Also, we chose to do push notifications only when there's someone really relevant to you.

Here you are in Austin for SXSW. How do you get ready for something like that?
Vaccari: We did two things. First, we had to get ready on the client side. That's why we released 2.0 for Android and iPhone. On the back end we already had plenty scripts to simulate spikes in sign-ups. So we can scale up very quickly. And today we were featured on the Apple App Store's Social Networking section, so we already had a spike, so we can hold up. We would actually love to crash because have too many users.

There will be a heavy concentration of users here. But does it require a critical mass of users?
Vaccari: That's another example of how we do things differently from Highlight, which is tailoring its strategy towards tech-savvy users in San Francisco or maybe New York, where it's really easy to walk a block and run into someone who's got something in common with you. We are trying to think more in general for people who don't have 2,000 friends on Facebook or live in very highly concentrated areas. That's why we have Radar, where you will get a notification if someone is close to you, but you will also see people that are a little farther away.

With Highlight, location is basically the driver. You're close to each other, so they're going to let you know. For us, location is just one of the signals, together with the compatibility, number of friends in common. So you can think of our push notification more as a recommendation rather than an automated--"Oh, you're within a certain distance, so we're going to let you know." Distance again, for us, is just one signal. So an example, if you live in an area that's not that populated, and all of a sudden, someone with 10 friends in common with you flies in, you're going to get a notification.

Where is Glancee going in the future?
Vaccari: Matching people to one another is always the trigger, but we realize that there is much more to it than just that. We think the idea of taking who you are online and making it so services, and shops, can take advantage of that is very important. So one thing we plan to do in the future is partner with major retail chains. Imagine Starbucks or Banana Republic or Gap, and as you walk by their stores, let's say your profile matches the target audience they want to address for a special campaign. Instead of receiving an ad or a pop-up that just says you have been selected, which is disruptive, we would like you to receive a message through Glancee, from an actual person inside the store, a message that says, "Hey, we think you would like this because of your interests. Would you like to know more? I can tell you about it over Glancee, or you can come inside and talk to me." All of a sudden, an ad becomes a person you are talking to.

Do you have any signed partnerships like this?
Vaccari: We are in discussions with a bunch of people. We have event organizers that are interested in using Glancee. We already tried with a couple of events, like Foo Camp. And they were successful. We can do a lot by featuring stuff, or sponsors at the top of the list, allowing sponsors to really find targeted key people they want to talk to. Or by providing analytics after an event about their audience. So they can understand your crowd, when they come in, when they leave. What kind of education they have. How they connect. Are there patterns that you couldn't otherwise find?

Why is this year when all the people discovery apps are blowing up?
Vaccari: There are two good reasons. One is that the location technologies have become good enough and don't consume too much energy to be usable if you do it right on a 24/7 basis, and that wasn't achievable before. That's why Foursquare had to do check-ins, in my opinion. Second, social norms are changing. People are starting to become more comfortable with letting others discover a bit about themselves, putting out a bit of information about themselves.