colour founder Bill Nguyen feels misunderstood, but he admits it’s partly his own fault.The launch of his stealth startup yesterday was overshadowed by news of the surprisingly huge $41 million investment it received.
Worse, the app itself was panned by critics, most of whom ignored the prominent warning not to use the app alone.
“I think those reviews are all very accurate,” he says. “We did a poor job of telling people ‘don’t use this alone.'”
Nguyen talked to us about the launch and his vision for colour late yesterday. Here are some of the points that the early critics missed:
- colour is not about photo sharing. It’s a new way to build spontaneous social networks — and collect massive amounts of data about what people are doing and where they’re doing it — without collecting any personally identifiable information like last names, addresses, or even passwords.
- The technology is pretty amazing. colour will be able to detect when a bunch of people are pointing their camera at the same thing — say during a baseball game — and automatically figure out what everybody’s trying to capture. Then it will select the best picture and put it to the top of the photo feeds of people most interested in that image (like fans at the ballpark)
- People are downloading it like crazy. In less than 24 hours, it’s already the number-two social networking app on iTunes, after Facebook. Nguyen says he’s received emails from Japan about people using it to share photos of devastated areas with nearby rescue workers.
- Real-time news reporting is coming. The company is working on a News API that will allow journalists to post pictures and map their location, and then will link relevant pictures taken near that location back to the journalist’s story. This would be great for covering events like trade shows or natural disasters
Here’s an edited transcript of our conversation:
Business Insider: What are you trying to accomplish with colour?
Bill Nguyen: We’re trying to create a new social network really around the post-PC world. Online, we use social networking to interact with people we already know, our friends, our family, sometimes business associates, but mostly people we’ve known for a while that aren’t necessarily around us. With the power of the Internet, I’d love to build something where you got to know your community. Your neighbours, your coworkers, people at the restaurant nearby. We wanted to build a technology that fit inside your pocket so wherever you went to make the room a little smaller, a little more intimate, and more interesting.
BI: That sounds a bit like a group messaging app called Yobongo. They’re not photo-based, but it’s a similar idea where you have chat rooms based on your location.
BN: Geography is a super simplistic view of how to give us information. I’ll give you an analogy — when you look at the original search technology, Yahoo looked at the world and said, “The best way to give people the best links is to have people submit links to us, then we’ll review them and tell them what the best links are.” It was very curated, and very simplistic. Google looked at the problem very differently and said “we want to look all the data possible and get accurate information.” If it’s just GPS, it doesn’t mean much. A GPS location in Manhattan means almost literally nothing, a lot of people from different walks of life are in one place and it’s a bit random.
We actually take the social graph based on where people spend their time, how they move around with one another. So in the middle of Manhattan, we try to find the 20 people who might possibly be interesting to you out of the 2,000. There’s a lot of technology that goes to solve that.
We have DJ Patil, he came out of LinkedIn, he was their chief scientist, he is working on the algorithms to deliver this, and he’s testing them now in real time.
BI: So are you collecting data about people’s interests? How are you figuring out their social graphs?
BN: On the Web the way we would answer those interest questions is ask people. What do you like to do? Where are you?
What we do is very different. We know for example if you go to a certain place between the hours of 7 and 10 pm, it’s probably for personal entertainment, it’s not necessarily for work. If you’re somewhere between 8 pm and 5 am or 7am, it’s probably where you reside. So the phone is giving us a lot of information that helps us understand what you’re doing. What’s cool about this is we collect no personally identifiable information. We don’t collect your last name, we don’t collect a password. When we think about you, it’s not you, it’s your cell phone. So maybe we’re really a social network of cell phones.
BI: Some of the first reviews of the product have been pretty rough. I think that’s because there’s not a lot users and not a lot of data in there yet. Are you seeing a chicken and egg problem here?
BN: We tried our best — and we can always do better — but we tried to tell people “don’t use this alone, it’s not very fun.” If you download the app with someone else and they’re right next to you, you’re going to have an amazing experience because literally everything you see that they capture on their camera you’ll have on your camera. It’s a completely surreal experience.
Today when you and I go to an event and take pictures on our phone, we’re uploading to a service on the Web and people view them, but it’s as if we were never together at all, it’s a two uncorrelated events. With our application, when you launch that application with other people, it’s intimate, you’re part of a group of people.
But I think those reviews are all really accurate and great. We did a poor job of telling people “don’t use this alone, or it’s kind of pointless.”
BI: Do you expect that your ratings on iTunes will go up as more people download and test it out?
BN: Number one, more people will come on. Number two, once we do a better job explaining how it works.
We could have done a way better job of creating a tutorial. We did a lot of the UI things that will be the right paradigms early, but they’re a little bit challenging right now. Like not having labels — that’s because we want to be international from day one.
But it’s crazy, the usage has been insane. And the ratings on iTunes are really weird always, I look at the number of reviews and it’s not even one tenth of 1% of the number of downloads we’ve gotten.
BI: How many downloads have you gotten so far?
BN: I’m not going to comment, but if you look at social networks as an app, we’re number two. We’re only behind Facebook now.
I totally hear what people are saying, I feel bad about it, I want to do better, but put it in some context. We raised a lot of capital and we have a long way to go.
BI: Let’s talk about that capital. There is a lot of competition in mobile photo apps. So how do you justify that round of funding, what are you going to be doing that’s different?
BN: Photo sharing is not our mission. We think it’s cool and we think it’s fun, but we’re a data mining company. We are really much more about bringing these spontaneous instant social networks. We happened to begin by launching an application that captures photos and video and text.
I’ll tell you a couple things that have been amazing today. We got a whole bunch of emails from people in Japan, and they’re using the product in real time because it’s the only way they can communicate photos really quickly with a lot of other people. And it’s not because it’s trending on Twitter or anything else, it’s because they’re just around each other. That’s an example of really sophisticated networking applied in real time.
We had a crazy one where someone was taking photos in an office building in New York, then Jimmy Fallon jumped into their photo stream and offered tickets. That was pretty cool.
BI: Explain a little more about how it works.
BN: The algorithms of how you create a group are really important. So let’s say you take a traditional social graph, it doesn’t take location or proximity. If there are a lot of people in a certain geographic range that are communicating with each other, and they’re sharing similar images and similar information, we assume that data is relevant for people 20 miles from there very quickly too. It’s kind of like a clustering effect. If we see a bunch of photos being taken, we instantly expand that.
So for example, if there’s a game going on at AT&T Park, I bet it’s really interesting to people two or three blocks away, not just inside the stadium. So we dynamically extend the range of our technology. When people are in a situation that’s more of a natural disaster, we extend the range up to 4 or 5 miles. So it’s not fixed, it’s not always 150 feet.
BI: Is there some kind of filtering mechanism? If there are 1,000 people at AT&T Park taking pictures, I wouldn’t want to get flooded with too many.
BN: We took what we call the Google approach to data. All of these pictures really did happen, they were all taken at AT&T Park, but we’ll begin by showing you pictures closest to you, and the pictures you were trying to take. So if you’re pointing your camera at the third-base line, we’re going to use that information and try and find all the pictures at that latitude and longitude and deliver those to you as well. Imagine it’s a super-zoom lens of unimaginable quality.
BI: So is there some sort of algorithm that tells you which one is the best and moves it to the top?
BN: You’ve got it. It’s all algorithmic. It’s based on links too, by the way. How many people are viewing, how many people are tapping, all kinds of stuff. It’s kind of like a PageRank for images.
BI: What’s coming next? What kind of features are you going to add, and will this be the company’s only app or the first app of many.
BN: We’re trying to create some tutorials so people can understand better what we’re trying to achieve, we’re trying to tell people “don’t do this alone.”
There’s not much we can do about the money we raised, I’m not giving any of it back.
One thing we’re working on really hard is the News API. We think it’s going to be huge. It’s going to be a service where journalists can take photos with our application and click on one application called the Add colour button and click on a map to add more information. So for example if you’re at CES, you can click on a map, expand the range, and every photo taken at CES will now be linked to that article you wrote. That can be an amazing tool for people to collect news and share data in real time.
That will be a curated application, it will be available as an extension of colour, but only to journalists, and they’ll have to go through an approval process on our site.
BI: The app is free, so what’s your business model?
BN: Advertising through the app. We’re going to build a intelligent system that allows businesses to participate with their customers. So when you walk into a restaurant and you use colour, and they’re also customers through a self-service Web interface — or actually a self-service iPad interface — every time you walk into the restaurant, your [first] name will show up with your picture. The maitre d’ or receptionist will know who you are, they’ll be able to welcome you, they’ll know the last time you were here, they’ll be able to see pictures if you took them here. They’ll be able to provide you better service than they’ve ever before, that’s going to drive up their revenue by increasing repeat business because we always want to go back where we feel welcome.
BI: They would pay you for that capability?
BI: How long have you been working on this?
BN: Seven months. It’s kind of funny because we’re right in downtown Palo Alto, dead in the heart of Silicon Valley, and we wrapped our entire building in paper so literally no one knew we were here for a period of seven months. We opened all the windows today and people are kind of excited and blown away by it all.