At CMT, we have developed the DriveWell program, which we use to build smartphone apps to make drivers safer. These apps work by giving users scores on each of their drives, taking into account several factors (hard braking, at-risk speeding, phone distraction, harsh acceleration, and harsh cornering). They also provide an overall score (typically over a rolling two-week duration), and a variety of features to engage users and incentivize safer driving, such as personalized driving tips, a leaderboard where they can compare their scores (to their families, friends, neighbors, co-workers, town, state, etc.), and occasionally cash prizes for safe driving.
A natural question to ask is: how well do our apps work? Do drivers benefit from these apps and actually improve their driving behaviors, by using their phones less while driving and by being more attentive and reducing the rate of hard braking?
Before I answer that question, you may be wondering if behaviors like phone distraction and hard braking are actually correlated with crashes. Numerous studies have shown them to be. For example:
- According to a study by the Virginia Tech Transportation Institute , “text messaging increased the risk of a crash or near-crash by two times and resulted in drivers taking their eyes off the road for an average of 23 seconds total” and “activities performed when completing a phone call (reaching for a phone, looking up a contact, and dialing the number) increased crash risk by three times.” One suspects that other mobile apps, like Waze and Pokemon Go, are similarly risky.
- A study by the NHTSA  showed that unsafe drivers are about twice as likely to brake harshly per mile driven. The NHTSA study defined “harsh braking” as a deceleration of more than 0.3g (10.6 km/hour per second, or 6.6 mph per second); this intensity doesn’t make your tires skid, but is enough to make you feel like you are being jerked forward when you stop. Intuitively, lots of harsh braking suggests that the driver may often not be paying enough attention to the road, or anticipating early enough. Furthermore, slamming on the brakes increases the risk of being rear-ended, and of misjudging the distance and hitting the car in front.
The DriveWell program measures these risky driving behaviors using sensors on the phone. DriveWell’s phone distraction focuses only on significant handheld distractions, such as picking up a phone and talking, or using any app by moving the phone and tapping on it, and does not penalize mounted use, hands-free use, and so on. The goal is to identify and dissuade significant distractions that likely take the driver’s eye off the road, rather than penalize things like hands-free calling, which are problems because of their cognitive impact, but still not as harmful as the behaviors that take both the mind and the eye off the road. For braking, acceleration, and cornering, we take the accelerometer data from the phone — which measures the force the phone is experiencing in the two axes parallel to the surface of the phone and one axis perpendicular to the surface — and transform this signal into an estimate of the acceleration of the vehicle, while ensuring that the movement of the phone relative to the vehicle itself does not contribute.
Analyzing The Data
To understand how driving behavior improves with time, we analyzed data from over 40,000 users in multiple countries including the United States and South Africa. The incentives for these users to improve driving include:
- for all users, “social gamification” in the form of leaderboards where a user can compare their driving with peers (e.g., friends and family, town, state), and also personalized driving tips provided using the measured data
- for some users, insurance discounts provided for safe driving.
With this data, we answer the following questions:
- How much do drivers improve on average?
- How much better are “good” drivers than “average” drivers?
- How long does this “improved driving” effect last?
How Much Do drivers improve on average?
To answer the first question, we measured the number of hard braking events and the time spent using the phone while driving on day 1. We found that the average user user is significantly distracted by their phone for 91 seconds per hour of driving, and has 3.5 hard braking events per hour of driving. We compared initial performance (end of day 1) on these measures to performance after 7, 30, and 60 days of driving.
From this chart, we can see that our users show rapid and sustained improvement in their driving habits. By Day 30, we see a 35% reduction in phone use and a 20% reduction in the number of hard braking events. For phone use, the numbers continue to get better, while they degrade very slightly for hard braking events.
What this shows is that users of safe driving apps, do in fact improve once they are given one of these apps. What’s surprising is that this isn’t even the population of users that our program identifies as good drivers — it’s across all users!
How Much better are good drivers?
If we focus our attention on the top 15% of drivers, we find that they are:
- distracted by their phone about 40% less than the average driver, with 50 seconds of phone user per hour of driving, and
- have 50% fewer hard braking events, or 1.7 events per hour of driving.
For these users, the improvement over time is even more dramatic than across all users:
By day 30, these users have a 40% reduction in hard braking and a 70% reduction in phone use!
How Long Do These Effects Last?
To answer this question, we looked at about 20,000 users who have been using one of our apps to obtain monthly insurance discounts for safe driving. We found that even after 200 days of usage:
- the average user still shows a reduction of more than 25% in both phone use and hard braking, and
- for the top 25% of users, we see a 2x reduction in braking and a 3x reduction in phone use.
I.e., the positive behavioral change persists!
The Bottom Line
User’s of the apps built with CMT’s DriveWell program show a significant and sustained reduction in phone use while driving and in the number of hard braking events. Specifically:
- Across all users, on average, users reduce hard braking by 20% and phone use while driving by 35%; phone use continues to reduce steadily through Day 60 and Day 90 as well.
- The top 15% of drivers, are distracted by their phone 40% less than the average driver, and have 50% fewer hard braking events.
- Drivers persist with their improvement continuously, even beyond 6 months.
These improvements don’t just reflect in the numbers but in real testimonials from our users. Here are a few emails we’ve gotten recently:
- [The app] is a great challenge that I enjoy daily now. There is no question that my driving behavior has been modified knowing that your app is keeping an eye on me!
- I am amazed at how good this app is technically and beneficially. Thank you for it; I’ve become a better driver with its help.
- I enjoy the app cause it keeps my mind even more in that space of being mindful of everything I need to do as a driver.
- I have to admit something that I’m not proud of. I have been a distracted driver. I could come up with excuses as to why, but they would just be lies…. When I started, it was hard to realize that I needed improvement, but the numbers didn’t lie. However, I am happy to report that all my areas of driving have improved and my stress level, on the road, has been drastically reduced ! I still have my moments and people that drive slow in the passing lane STILL drive me nuts, but I know I’m a safer driver now and if you practice something long enough, it becomes a habit. Good or bad. Thankfully, this old dog is breaking bad habits and creating good ones and everyone is better for it. This is one of the “badges” I earned, of which I am particularly happy. (I’m actually up to 30 trips without picking up my phone!) Mission Accomplished!
 Klauer, S. G., Dingus, T. A., Neale, V. L., Sudweeks, J. D., and Ramsey, D. J. Comparing Real-World Behaviors of Drivers With High versus Low Rates of Crashes and Near-Crashes. DOT HS 811 091. February 2009.