Related Articles
Many of us remember how liberating mobile phones were when it came to meeting people: “Which street corner was that again?” “Should I buy tickets for both of us?” “Sorry, I’m running late.”
A start-up called Twist, though, believes phones can make the process much more convenient — by automatically handling some of those messages.
With an iPhone app, the startup’s technology tracks when a person leaves on a journey and tells the person at the destination what the expected time of arrival is based on data from Google Maps, Bing, and the company’s own algorithms that factor in things like personal driving style and GPS trouble spots. If you hit traffic or your bus doesn’t arrive, Twist will send a notification by e-mail or text message, and a few minutes before you arrive, it’ll send another alert.
It’s a modestly scoped utility, but the company expects it’ll have appeal for the many times when people are getting together.
“It’s a simple and easy-to-use app that lets friends know when you’re going to arrive. There are a lot of apps that solve where, but we’re trying to solve when,” said Chief Technology Officer and co-founder Mike Belshe, who previously worked on Google’s Chrome team where he helped create the SPDY protocol to improve Web performance.
Although Twist is an app, what it does is arguably a feature that could be built into smartphones as mapping and navigation services steadily grow in sophistication. For example, the new Google Now feature in Android 4.1, aka Jelly Bean, can tell people when they should leave for an upcoming appointment, factoring in traffic, public transportation schedules, and other information. And with iOS 6, Apple is building its own mapping service, too.
Predicting travel times is where Twist’s plan for actually making money comes in, said co-founder and Chief Executive Bill Lee. Although Twist isn’t charging for consumer services, it eventually expects to offer a business version that lets companies tell customers when to expect the food delivery or the cable guy to arrive. Alternatively, a restaurant could know better if that party of six really will arrive at 7:30 p.m.
In his meeting-infused life, “Somebody always is early or…late,” said Lee, who earlier in his career was co-founder of Remarq and Social Concepts. “The idea came out of a personal need — a better way to keep people from wasting all this time.”
The company also announced $6 million in first-round funding from Bridgescale Partners, Lee, Belshe, Eric Hahn of Inventures Group and Netscape’s former CTO, and Jeff Skoll, the first employee and first president of eBay.
Although the company today relies on Google and Microsoft for traffic data, Twist hopes hopes for more independent information from its own app, Belshe said.
“In the long term, as we get bigger and get more data, we’ll be having our own sources,” not unlike what Google and Apple do to collect location data, he said. “As we get more users on the system, we get more data. Our system already uses that in how we compute the ETAs [estimated times of arrival]. Over time, we can get more reliant on that.”
The company’s algorithms learn from individual behavior, for example whether a person drives slower or faster than average, and from aggregate data for things like how long it takes to traverse San Francisco at a particular time or particular day.
“We’re now up to about 98 percent accuracy with our ETAs,” Belshe said.
The app is supported for trips in the United States, but the company hopes to expand to Europe by the end of the year. It works with driving, walking, public transit, and bicycling trips.
Twist is launching with an iPhone app, but an Android app is planned “in the not-too-distant feature,” Belshe said.
They started with the iPhone because “it’s got a great user base, it’s a good product, and it’s a lot easier to build on,” for example with better application programming interface (API) to tap into GPS location data, Belshe said.
“On the Android product, the fragmentation problem is big. When people say ‘Android,’ they bundle together a lot of devices that really aren’t the same,” he said. “With Ice Cream Sandwich [Android 4.0], which is on the top performers of the Android market, the APIs are in good shape, but the older version really need a lot work. It’ll take awhile to make something that works well for the entire user base.”
Hardware diversity also makes Android development harder, Belshe said. “The other practical considerations of screen size and processing power are hugely different, too. It makes it a more treacherous platform overall,” Belshe said.
