As connected cars proliferate, the Kirkland firm could be telling all of us how to navigate congestion
It was a drizzly February morning when Ben Ascher, a BMW of Bellevue sales associate, showed me the future of personal transportation. My vehicle: a BMW i3, an all-electric coupe with the stubby and aggressive appearance of a refined bulldog. “A lot of people don’t like the look,” Ascher says. “I tell them that this thing was designed to be polarizing.”
Stomp on the accelerator — don’t dare call it a gas pedal — and the torque-heavy motor launches you to 60 mph in a mere 6 seconds. It’s quick off the gun, as a BMW is supposed to be, but that’s where the i3’s similarities with other BMWs end. The car tops out about 30 mph later, and its battery will carry the driver only 80 miles on a charge. A 400-horsepower, gas-guzzling road rocket still might signify our dominance of asphalt, but those speedy machines aren’t well designed for the traffic-light slog of city driving or the stop-and-go scurry on busy interstates. The i3 is designed for this new type of driving. The car is nimble, sips energy (the Environmental Protection Agency estimates it is as efficient as a 127-mile-per-gallon gas car), parallel parks itself, and has a light, responsive feel. It’s also very smart.
As I drive the streets of Bellevue, a comprehensive dashboard roadmap shows me nearby charging stations, up-to-the minute traffic conditions accurate within 800 feet, and a map displaying how far I can drive. If the system guesses the car doesn’t have enough juice to reach my destination, it will kindly suggest I park at a charging station and hop on a bus.
The I3’s navigation system is possible thanks to Inrix, the Kirkland company that collects, distributes, and analyzes traffic data. The firm sits in the middle of a strange crossroads for one of America’s fundamental experiences — driving.
For nearly a century, driving didn’t change. The hit-the-open-road ethos was alive and well, and gasoline taxes allowed road builders to keep pace with the growing number of drivers. Today, congestion is the norm in an increasing number of U.S. cities, and those commuting on the Eastside spend as much time driving 20 mph on the interstate as they do the speed limit of 60 mph. The open road is no longer so open.
At the same time, cars are connecting with electronic networks around them as consumers and manufacturers prioritize digitization and automation. Inrix’s stance is that leveraging data produced by cars is necessary to unclog roads. It’s clear that better understanding the way traffic moves can help alleviate gridlock, and Inrix has the data and the analytics to further that understanding. What’s less clear is whether its tools can significantly reduce traffic congestion, and whether it even needs to do so to make money.
Today, Inrix CEO Bryan Mistele is known as a key player in the local big-data sphere, but in the years following Inrix’s 2004 founding, he was better known for his venture-capital woes. The startup community still talks about Mistele’s infamous streak of 70 nos from investors, a run that nearly tanked his company before it got off the ground but one that now gives him the caché that comes with doggedness.
That magnitude of rejection wasn’t a product only of bad luck. Mistele and cofounder Craig Chapman left Microsoft to found a company very different from what Inrix is today. Inrix was originally conceived as a device maker. Mistele and Chapman were hawking a thumb-drive-size product that could be loaded with presentations and plugged into projectors, no laptop required. But capitalists weren’t interested in a company with a single product, so the duo had to broaden their line.
Devices continued to be the mantra, but Inrix’s founders also began looking at a nascent market. Before GPS systems were commonplace, Mistele and Chapman wanted to sell a small plug-in device that could alert drivers of traffic conditions. It was an early idea for the connected car, but it too failed to gain traction. After a year of VC battery, Mistele was ready to end his startup bid, but his convincing wife and analysis of investor critiques yielded a new idea.
The problem with Inrix’s Product B was a dearth of information. “We couldn’t really build that device, because the data about where there was traffic didn’t exist,” Mistele says. “The more we explored that, we ended up with the idea of focusing on the traffic data.” Devices were thrown out, data were embraced. Inrix finally received $6.1 million from Venrock and August Capital in 2005, and it began contracting with fleet companies. Inrix received data from their vehicles’ GPS systems; in return, the companies received the aggregate data Inrix collected, allowing them to better navigate and save precious fuel and time.
As the connected car landscape has grown, so too has Inrix’s client list. The company has 350 customers, many of which also provide Inrix with GPS and cellphone data. Automakers like Ford, Audi, and BMW provide Inrix with data and use it in their in-vehicle navigation systems. The privately held company doesn’t publish its financials, but Mistele says about 50 percent of Inrix’s revenue comes from automakers, while the remaining half is split between government agencies, fleet companies, and media organizations.
“In (transportation) circles, we talk of Inrix as a great source of information,” says Sarah Kaufman, who studies digital transportation systems at New York University. “It’s one of several, but Inrix is considered a very robust system.” In total, Inrix collects 100 billion data points each month. It monitors traffic in 40 countries on 6 million miles of roads. It’s a wealth of information that, if applied correctly, could make a dent in one of the biggest headaches people deal with each day.
As a data-rich company claiming to alleviate traffic congestion is wont to do, Inrix contributes to and trumpets studies reminding people that traffic sucks. One of its earliest publicity grabs was its traffic scorecard, which ranks world cities by their propensity for gridlock in an interactive chart on its website (as of July 2014, Belgium had the worst traffic in Inrix’s 40-nation network; Milan was the most snarled city). It also ranks specific corridors in the U.S. by their peak delays compared with free-flowing traffic speeds (I-5 southbound between 130th and Union streets is No. 6 on the despicable list; I-405 southbound between the Eighth Street and Coal Creek Parkway exits is No. 12). According to Inrix’s data, an average Seattle-area driver spends 40 hours a year sitting in excess traffic.
To reframe the issue, Inrix partnered with the London-based Centre for Economics and Business Research to release a study in October that put a price on all that idling. According to this analysis, Americans lost $124 billion sitting in traffic in 2013, and that number could balloon to $186 billion by 2030.
Most drivers are well aware of the gridlock in cities, and the problem will likely get worse as U.S. cities grow and become denser, which will leave us with situations akin to European cities like London, where drivers sit in traffic for 80 hours a year despite a robust public transit system.
Freedom-of-the-road proponents are reacting. Ford and BMW have taken to the connected car movement with fervent rebranding, labeling themselves not as carmakers but as “mobility companies.” At the Consumer Electronics Show in January, Ford CEO Mark Fields laid out his vision for the future of driving, and it’s not drivers feeling the wind in their hair in a Mustang convertible.
“The existing infrastructure for motor vehicles simply cannot sustain the sheer number of vehicles expected to be on the road in the coming years,” Fields said in his keynote. “So our roadmap has to include not only smarter cars, but smarter roads and smarter cities.”
Dave McCreadie, Ford’s manager of electric vehicle infrastructure, says Inrix’s data analysis is integral in the information systems Ford plans to roll out in its electric vehicles. “Inrix was already a data provider to Ford” — Inrix data power Ford’s Sync navigation — “so there was already that relationship, and it’s been a pretty fruitful one for both sides.” For example, Inrix is providing Ford traffic information the automaker can’t collect itself because it has access to data from its own vehicles, while Ford is providing Inrix with parking data gathered by ultrasonic sensors in bumpers used for automated parallel parking and blind-spot detection.
German automakers are particularly fond of Inrix. Porsche’s investment arm pledged $55 million to Inrix in 2014, and BMW and Audi are using Inrix data for a task that just a few years ago would seem unthinkable: telling drivers to opt for public transit rather than driving their own car.
This is the essence of the personal mobility concept. Thanks to a century of deft marketing, driving evokes a sense of freedom and power in many of us, but that’s only in effect when the car is actually moving. Thus, when traffic is bad or the vehicle is low on energy, new Audi and BMW navigation systems equipped with Inrix data and algorithms will tell the driver when it’s best to take the bus.
“We’re not saying, ‘Let’s just keep the customer in the car,’” says JonMichael SanGiovanni, the senior engineer for BMW’s ConnectedDrive navigation system used in the i3. “If they trust in us for one form of mobility, let’s look at ways to increase our mobility footprint to continue to provide premium services.”
To understand the future according to Inrix, one must first understand that vehicles rolling off assembly lines today are loaded with sensors. There is a GPS, an accelerometer, a fuel gauge, multiple radar systems, traction control monitors, temperature and humidity sensors, and more. It’s not unusual for new vehicles to have 60-100 sensors — luxury vehicles can have hundreds — and most of them are producing data that are not being used by any commercial entity.
“The sensor data exists” on vehicles’ internal computer networks, Ford’s McCreadie says. “It just has to be pulled off of the network, and it has to be interpreted for what it’s saying … and then it would be sent to the cloud. Commercially, that’s not happening today, but the data is there.”
Inrix is utilizing GPS data from cars and smartphones, but the company’s growth will hinge on it being able to use data from the myriad other in-vehicle sensors. At a big-data conference in Seattle in March, Mistele discussed how vehicles could be used as weather monitors. Traction-control sensors in one car could alert other drivers to ice on the road, while windshield wiper-speed data could help gauge precipitation. Combine these, and you’ve got millions of roving weather stations that could provide far more accurate data than the local DOT’s webcam.
Once Inrix can better handle and analyze that data, Mistele envisions the company becoming somewhat of a backbone for a more productive transportation infrastructure.
“It starts while you’re at the breakfast table,” he says. “The phone lets you know that you don’t have enough gas to get to work and you have to leave early because you have to stop for gas.” Watch the morning news, and your TV — if you have a cable package from Inrix customer Comcast — lets you know that construction will slow down your commute by 15 minutes. Once you’re in the car, a navigation system connected to your personal calendar will automatically have directions pulled up to your destination, be it the office or an offsite meeting.
The scenario Mistele lays out can directly help consumers, but governments and municipalities have to jump on board for Inrix’s effect to reach its zenith. “You could see more efficient routing, fewer traffic backups, fewer disruptive forces on the route such as blind curves and confusing intersections,” says NYU’s Kaufman. Parking and express-lane price tweaks also could be enacted based on real-time data.
Mistele says it has been difficult convincing public agencies to invest in Inrix rather than developing their own systems or eschewing big data altogether, a sentiment echoed by Kaufman, who previously worked for New York’s Metropolitan Transportation Authority. “Public agencies tend to be risk averse, so jumping into a new technology comes with some hesitation,” she says. “At the same time, these technologies aren’t perfect. … As sensors improve, the data will improve, and that will make it easier for public agencies to get involved.”
Cities invest heavily in traffic monitoring — cameras and weight sensors embedded into asphalt are at nearly every intersection, and other sensors are used on arterials and high-traffic sections of highway. But Inrix makes cars themselves the data points, not just the sections of road that happen to be driven over. If an accident causes a traffic jam on a rarely traveled road, Inrix data will reflect it; sensors in the city center won’t.
Inrix Insights, an analytics platform the company released this week, should help the company better court public-sector clients. It’s broken into two products — Volume shows a client where traffic is at any given time, and Trips explains where that traffic is coming from and going to. “We’re automating something cities have always done with pen, paper, and surveys,” says Steve Banfield, Inrix’s chief product officer. “Volume will tell you how many cars drove over a piece of asphalt without installing a (mechanical) counter. Then you can combine that with Trips and know where they started from and where they are going.”
Mistele says a city that receives data from and shares data with cars can reduce traffic by 30 percent. That claim isn’t backed by studies, but research does suggest traffic data placed in the hands of drivers can have a significant impact. A UCLA survey found that 30 percent of cruising vehicles in downtown areas are looking for a parking space, and that figure largely remained consistent from 1927 to 2001. In 2014, MIT outfitted 10 drivers with an app that guided them around traffic congestion, and the drivers experienced an 8 percent increase in rush-hour speed. If cities use car-borne data to better understand where traffic is and how to price parking stalls and express lanes, it’s reasonable to believe a double-digit percentage improvement in traffic delays could occur.
Inrix finished its October report with four suggestions to mitigate U.S. traffic woes. The first three were services Inrix already offers: multimodal navigation systems, navigation that directs drivers to the closest available parking and the cheapest fuel, and real-time analytics at the city level. Inrix-enabled navigation systems like that in the BMW i3 check off the first two criteria, while Inrix Insights accomplishes the third. But the fourth suggestion peeked a little further into the future: Cities should consider infrastructure that supports autonomous vehicles.
The connected car discourse is progressing at very different paces in the public and private sectors. While cities are sticking with mechanical sensors to monitor traffic, companies are producing working prototypes of self-driving vehicles. So whether cities are ready for them or not, autonomous cars could soon be on public roads. A recent study from McKinsey & Company says vehicles will be fully autonomous by 2040. Google’s ballyhooed autonomous cars have logged 700,000 miles without a crash, and a Wired magazine writer drove, er, sat in an Audi that piloted itself from San Francisco to Las Vegas.
“You need two things for a truly autonomous vehicle,” Mistele says. “You need sensors on the vehicle to detect that there’s a vehicle ahead of you … but you also need real-time information at a citywide level. Carmakers are doing a good job of building sensors into the car. At the same time, these vehicles have to be connected.”
As more connected cars hit the road and their technology progresses toward automation, a virtuous circle is established: The cars produce more data as they operate, and they need more data to operate. Inrix is poised to sit in the middle of the cycle.
It’s important to make clear that Inrix’s data and analytics cannot make traffic conditions on I-405 in Bellevue mirror those on I-25 in Cheyenne, Wyoming. Even the most brilliant transportation managers and data scientists cannot overcome the primary drivers of traffic — population growth and economic activity.
It’s in this realm that Inrix’s message is tailored, and perhaps misleading. Inrix explains the need for its products by emphasizing the economic cost of traffic congestion, but it’s economic activity that propels traffic. Mistele doesn’t deny that traffic is an economic indicator, often citing a 20 percent dip in traffic during the global economic recession and the ensuing rebound once the economy improved. But if drivers and cities ignorant to data aren’t the root cause of congested roads, then how much of a difference can Inrix make?
The answer could be why Inrix was valued at $550 million when it received the Porsche investment: It doesn’t matter. Inrix’s greatest strength as a business could be that it’s trying to tackle a problem that, if the economy is strong, is unsolvable. Data-driven navigation and smarter cities almost certainly would improve traffic flow, but only to an extent. When Inrix numbers showed Seattle-area drivers spend 40 hours a year stuck in traffic, local media were aghast that we were spending a week — a whole workweek! — each year sitting in our cars. But 40 hours a year represents 10 minutes per workday. If a system based on Inrix data analysis cuts traffic 30 percent, that’s just three minutes shaved off the average Seattle commute.
Shaving three minutes can still be stimulating, though, as Inrix’s best-known competitor demonstrates. Waze, an app Google purchased for $966 million in 2013, taps into the GPS systems on users’ phones to provide navigation services and update Google Maps’ traffic readings. It re-routes drivers based on updated traffic conditions, playing on the blissful psychology of unexpectedly discovering an improved alternative, even if it is marginally better. Furthermore, Waze users can report fuel prices, police positions, accidents, and other traffic events, making traffic monitoring an active (if distracting) endeavor.
“Waze has become extremely popular internationally,” Kaufman says. “People are really relying on it now, and it’s a really good source for public planners to draw information about certain locations.”
Waze’s overwhelming popularity is helping it make inroads in the public sector. The app has been downloaded more than 50 million times, so even though active users sometimes fudge inputs — Los Angeles residents have falsely reported accidents to route traffic out of their neighborhood — and Inrix receives data from more diverse and reliable sources, the city of Boston decided in February to partner with Waze for real-time traffic information.
An absence of humans gaming the system isn’t Inrix’s only advantage over Waze. Though 60 percent of Inrix’s data comes from phones, it’s the 40 percent that comes from vehicle GPS that is most important. If a car is transmitting data, it’s being driven; phones transmit data even when their owner is at a standstill.
Nevertheless, Inrix can’t fix traffic. But that could be the key element of Inrix’s business model.
“It was a conscious decision to sell aspirin rather than vitamins,” Mistele says of the founders’ decision to focus on traffic. “Vitamins are nice to have, not a must-have. Aspirin is a must-have.”
No traffic would mean no headaches, and nobody takes aspirin unless they have a headache.
A version of this story titled “Can Inrix Fix the Traffic Problem?” appears in the May 2015 issue of 425 Business.