August 21, 2013
Q&A with Jamyn Edis, co-founder and CEO, connected-car startup Dash
Dash is developing what it describes as “a Fitbit for cars”—a low-cost, easy-to-install device that analyzes how the car and driver are performing and then provides feedback (by voice, while the car is in transit, as well as via an app and weekly emails). The company was selected for TechStars NYC’s most recent startup accelerator program, for which JWT serves as a partner. Dash co-founder Jamyn Edis, an adjunct professor of marketing at NYU’s Stern School of Business, formerly worked in research and development at HBO and spent a decade at Accenture before that. He talked to us about why drivers might be motivated to use Dash, what kind of third parties could leverage the service and the company’s take on privacy issues.
Can you give us the quick pitch for Dash and a rundown of the major features?
Dash is a connected car platform via hardware and software products. We use an on-board diagnostics reader that uses Bluetooth or Wi-Fi to connect to our application. And to date we’ve built that in Android and iOS mobile, and we also have a Web front end, which is in beta.
So our software will simply pull the data points from the car sensor. We pair that with the smartphone sensors and with your social networks—currently we are integrated with Facebook—so we’ll get all that demographic information. And then finally, what we call ambient integration. So we use third-party APIs for weather, for traffic, for location information, businesses and attractions you’re driving past. You take all of that and you’re gathering, at the very minimum, 300 data points in real time while you’re driving. From a consumer value proposition, the product focuses on three main feature sets, which are safety, savings and social.
So how exactly is the experience personalized? How do you put so much data into understandable terms for the user?
The expression of that data is different in these cases that we mentioned: safety, savings and social. But first, everything’s personalized because you are carrying your Dash application to your car. So we’ll know the make, model and year of car and its specific data.
There is definitely a generic suite of features that can be triggered from that data. So for example, on the safety front, if the “Check engine” light on your particular car goes off, it will translate what that means into plain English. It will tell you about the severity of that issue. It will tell you how much it will cost to fix it in parts and labor and will also provide you with a list of preferred mechanics where you can get that fixed, so you can be sure you’re getting the right fix for the right price, and from the right type of business.
On the cost savings front, we will track your trips and your cost outlay for things like gas. So we use a dynamic, real-time gas-pricing database, and we’ll know how you’re driving, where you’re driving, your miles per gallon against your EPA estimate, and so we’ll be able to see how much you’re spending on gas. And so, for example, it will tell you that you had eight hard brakes in this journey—a hard brake is expressed as decelerating by more than seven miles per hour in one second. And then we’ll say, “If you reduce your hard brakes over the next month of driving by X amount, then that will translate into Y difference of gas usage and will put an additional $50 in your pocket.”
On the social front, we’ve customized your experience because you’re comparing your driving behaviors to how your friends are driving. You’re sharing your trips. The way that you can share your check-ins on Foursquare, we share your road trips—on Facebook or Twitter or however you choose to share it—on a map with fun details and added metadata. And then within the application you have a leaderboard that compares you on a number of metrics to your friends and family.
We’re building out a newsletter, like your Nike FuelBand or your Mint.com newsletter. You’ll receive it every week, and it will say where you’ve driven, how you’ve driven, how you can improve, how you can save money.
What would you say is the most interesting thing you’ve uncovered through user data so far?
I caution that you don’t want to over-extrapolate based on a small sample size; we’re in our beta test right now. We have tested 400 car models across eight markets and gathered 20 million data points to date. You get interesting behavior. So, for example, the Dash scores—which go up and down based on positive and negative signals: hard brakes, idling, sending outbound text messages, stuff like that—have shown that people in California will have a better driving score than people in New York. People with Android handsets have a better driving score than people with iOS. Women have a better driving score than men.
Those are amusing anecdotes that we’ve seen and an example of the type of data mining that we can have. Ultimately, we want to get to a point where you can really break down by ZIP code, by car brand, by time of day, by particular routes, by particular types of journeys.
Does Dash have an API to help get some of those breakdowns?
Dash does have an API, and the API has not been built solely for internal use. The phone pushes data to our servers in the cloud, and that’s what gets put out onto our API and what powers the Web content, as well. The goal would be at consumer launch to open the API to the public, not only for third-party developers but also for enterprise solutions for dealerships, municipalities, insurance companies.
One of the vendors we’re talking to wants to extract driving data to traffic management solutions for a municipality—e.g., changing the speed at which traffic lights turn from red to green to maximize the flow of traffic. Someone was working on essentially a road trip application, one that will take music information and photo information, and build a catalog based on our data. Then there are utility apps that will tell your spouse when you are five miles away from home, and based on traffic data and how fast you’re going, will state when you’re going to be home, for example. Or tracking apps for parents to see where their kids are driving. These are all use cases that we don’t necessarily need to build into our consumer products, but they can be built on the API.
Since Dash doesn’t simply monitor vehicles themselves but also drivers and their behaviors, have there been any privacy concerns?
We have to be very, very judicious with the use of the data and how it’s owned and how it’s analyzed and how it’s distributed. The No. 1 point we need to put out there is that the consumer or driver owns their data. Anything that happens to it is based off their own opt-in. As an example, I might analyze your driving behavior: I would see you’re improving your Dash driving score and you’re under a certain threshold of miles driven per year that makes you attractive to insurers. You would be part of a class of drivers that I would take to the insurance company and say, “Here are the top 20 percent of safe drivers. Are you interested in offering them your premium discount?”
If the insurance company comes back and says, “Give a 15 percent discount to people over this threshold,” I would go to that collection of drivers and say, “Do you guys want the 15 percent discount?” Then you would need to opt in before any information is shared to identify you personally. That’s the type of approach we’re taking—very, very cautious, very conservative.
In navigating privacy, where do you think responsibility lies? Is it on the consumer to navigate the consent process? Or is it on third parties to be explicit and upfront about what users are agreeing to?
The bottom line is, going into the application, nothing is shared. The user, honestly, has to be cognizant of the fact that their data is being transmitted to a server where it will be stored and analyzed. After that it simply does not go anywhere unless they opt in explicitly to share it. The default standard is to not share.
Give us a sense of what the next steps are for Dash and, more generally, the advent of connected cars?
The industry is very much embracing third-party developers and a “second screen” approach to the car. You’ll see deeper integration of apps into the head unit, and it will definitely move to a point where you can Bluetooth into the car without the OBD [on-board diagnostic] device. We are working closely with OEMs and aftermarket companies to get on their radar, as well as integrating into Ford’s OpenXC platform and GM’s OnStar API. There is an opportunity to build a universal, over-the-top platform for the road that works on any model/make of car—that’s what we’re building—and I think you’ll see automotive as the next major category in Big Data and the Internet of Things, after the home and wearable computing.