EP11 – How Mobility Data Is Changing the Urban Environment

In this episode of On Point with Korem, I sat down with Bryan Mistele, CEO and founder of INRIX. We discussed the nature of how the collection of traffic and other mobility data is impacting everything from connected and autonomous vehicles to how this might impact the urban environment. Bryan discussed how cities will be transformed with more dynamic information, like optimizing traffic signal timing but also using location information arbitrage that hedge funds are leveraging to predict stocks prices. He thinks that whether its traffic lights, tolls, or sending messages to trucks and vehicles already on the road to avoid delays of even your latest Amazon order, more dynamic data will lead to benefits for consumers.


JF: So Bryan, thanks very much for doing this. It’s been many years since we’ve spoken but as I’ve followed INRIX’s over the years, it’s really evolved quite a bit from being a company that’s primarily based on data, the traffic data side, but has really evolved into developing products for SAS and just being in the mobility sweet spot of what’s really evolved recently. Do you think that that’s a good way to characterize the company or do you think INRIX fits into a new category?

Bryan Mistele: No, I think you’re absolutely right. We started in the early days really focused on delivering traffic data globally and primarily to the automotive sector but if you look at the business today, we have now more than a thousand different customers that we’re providing analytics and SAS solutions to. So we’re really focused on this area of mobility and figuring out what’s happening around the world globally and whether it’s traffic or parking or commute patterns, trip behavior, macro-economic data, that really is where we’re focused today. So it’s quite an exciting spot. There’s a lot happening and it’s exciting to get a chance to talk about it.

JF: So, what do you think was the reason why all of a sudden we see this upwelling of interest in mobility, and it’s not just traffic data, it’s mobile trace, it’s anything more dynamic. I mean is there anything you can point to directly that really  makes that data maybe more valuable than some of the historical stuff that we used to use?

BM:  Yeah so, I think the primary thing has changed is the data has become a lot more available. So, the density of the data, whether it’s coming from cars or trucks or smartphones, has increased by orders of magnitude. Like just in the last three years we’ve gone up by more than 12 times the quantity of data that we’re processing every day. So that allows you to do things like you couldn’t do before. I mean there’s one thing for us to be able to say in urban areas, we think we know what’s going on in terms of traffic, it’s very different to be able to extrapolate to a whole population and say, for a shopping mall, we know where your customers are coming from every day. That requires a different level of density. Or, if you’re trying to retime traffic signals and stop lights, you really need to have a very high level of density and granularity to do that that thing virtually using GPS data. So that’s really what’s changed, is just the density and the quantity of data that’s now available through these devices that now are in the market.

JF: So that brings up another challenge, which I think we, in the geospatial industry always thought we faced, which is; we have a data problem. There’s so much of it and do you feel like there’s been a need for you to process more of the data that you kind of get into the cloud native big data space and then have derivative data sets from whatever you’re collecting?

BM: So, we found is… you’re absolutely right. The quantity of data has increased so much that many customers just don’t know how to deal with that kind of data, so what they’re looking for are really insights. They’re looking for answers to questions they have. So, what we found is the reason, we kind of shifted the business three or so years ago, from selling data and selling APIs to providing a robust SAS Suite, is because customers want to be able to log in and say, “how do I retime this traffic signal?” They don’t want to have to go through petabytes of data to try and ferret out that information themselves. So, the reason we kind of pivoted and focused a lot on these SAS applications because the quantity of data just overwhelmed so many people, that you know they really needed a turnkey solution, where they could just log in, get the answer to the question they needed and then move on.

JF: Is there kind of a push-pull relationship between selling data and selling software? That always seems to be a challenge, at least in our industry which is, where we put our emphasis, as the software, as a data. Do you see that as a challenge, or have you naturally progressed from one to the other?

BM: Well, I think selling just raw data has never been a great business, right! There are some folks in the space that are data brokers, and they get data from source A, and they try and sell it to others. That model really doesn’t work because a lot of customers aren’t that sophisticated in terms of geospatial data and how to correlate it and snap to road segments and use the different map databases. So, we then from our early days, then invested in building a robust machine learning platform that took all of this data and then provided products. Things like traffic data or parking occupancy or trip patterns and things like that, and then provided that at the API level. And now what we’re seeing is this evolution again, to people saying, “well even that requires a team of developers to know how to ingest an API and integrate that into my systems”. And if you’re a public sector agency in Nashville, Tennessee or Charlotte, you don’t have teams of programmers. So, again they’re looking for a complete turnkey solution, which is why we’ve turned and really focus the company around providing them.

JF:  So, you just introduced the INRIX IQ Solution. I’m curious to hear more about that, and I looked at some of the interface and I thought, “okay, again it does have a real niche place in the industry”. Is it something you’re finding more interest from the public sector? The private sector? Both? What do you think…how the product is positioned?

BM: Yeah so, INRIX IQ has about 12 different modules today in it. Things like; traffic and real-time road analytics module, it has a parking module, it has modules targeted around traffic signal timing, a different module focused on macroeconomic information and trip patterns, as well as a location analytics module helping retailers do things like, figure out where to put stores. So, there’s a bunch of different modules that we’ve created and integrated together and  to (answer) your question, what we found is like I said; providing that kind of integration and providing it wrapped in the data is fairly unique in the market. So, I mean there are analytics packages that exist but then cities or enterprises have to get their own data and figure out how to put the two together. So, we’ve kind of created this turnkey solution, first targeted to public sector agencies because we’ve gotten a lot of interest a lot of traction in the public sector space but also we’ve now found with enterprise customers, folks like Discount Tire, who turns out, they’re using our macroeconomic tools to figure out how much people drive because that helps them gauge demand in terms of what they need to order and manufacturing and things like that. So, there’s a lot of enterprise customers as well that once you kind of move beyond complex APIs and data sets, now you’re just providing answers and solutions and charts and graphs through a turnkey solution, then it becomes interesting to them.

JF:  Yeah, let me go back to the issue around traffic data and just mobility data in general. I mean, we’re sort of waiting for autonomous vehicles, EVs, to become more part of what we can buy and expect. Are you finding that, at least from the autonomous side, that INRIX is getting more attention or are we still a long way off from really autonomous vehicle?

BM:  Well, you know obviously that’s the 25 000 question in the industry. I think you’re probably aware that Tesla now is testing their autonomous update, right. So, they’re in beta right now with an update to existing Tesla’s on the road to take them to level four economy. So, being able to navigate you from your house to your work in most conditions. So that’s quite interesting because you could be in a situation where at some point in the next couple months you’ve now got a million, or however many vehicles Tesla shipped so far, that have the capability. Where they could get that update and all of a sudden they’re level four autonomous. So, I think this is continuing to happen faster than most people think.  Obviously, level five is a different animal but in most cases, the eighty percent of the time where you can be autonomous, and not have to drive yourself from your house to your parking lot at work, I think that’s coming sooner than people think. And obviously, what we’ve seen is from the partners we’re working on, that creates a new set of problems, right. It’s one thing that the color road segments, different colors to show traffic on a map, but now they need something very different. They need to know about very specific road closure information. They need to know about when I arrive at a parking lot. Specifically, where are the entrances? Where are the spots? Where can this autonomous vehicle park? So, different level of granularity that we’re now working with a variety of cities on, through our INRIX road rules product. About 50 cities now have adopted it, are using it to digitize their roadways and their rules, so that whether it’s on Thomas vehicle or even a TNC, like an Uber or Lyft, or a micro mobility scooter, people know where to drop them off, where to pick them up, where they can park, where they can’t park, those type of things. So, that really is where I think most of the action is happening today, is in that creation of the HD map, the parking restrictions, the different layers that are needed digitally to move those autonomous vehicles around the roadway.

JF:  Yeah, and even (Elon) Musk is pretty direct about his comments, about his own technology, and obviously he needs improvements. Does that put any pressure on the ability to collect data at a much more higher granular level? Like, we need curb heights or something almost at what I would call an “engineering level specificity”, for digital road center? Do you see that happening, or is that going to be needed, or is what we have today just fine for the autonomous vehicle segment?

BM:  Well, so there are things that are needed right, in terms of that vehicle. If it’s going to do a self-park on street, it needs to know that it can park there, and what the restrictions are, and that it’s not…that eight to nine o’clock in the morning, there’s garbage pickup or something, and you can’t park. So, the vehicle needs to have that kind of information. It needs to know, again, once it goes into a parking garage; where it can park, how it can pay, those kind of things. There are levels in the HD map that are important, right! Things like, how many lanes there are. Lane configurations. Things like that are important to the autonomous vehicle. But I think that the need for the centimeter level accuracy, right, things like the curb heights, or the centimeter level accuracy in terms of lane markers. Most of that probably isn’t going to be needed because the autonomous vehicle makers are now relying on cameras. So, I don’t need to know digitally that the curb height is four inches. My cameras can pick that up. So, I think, like you said, there’s a need for additional data but it’s not probably as detailed as most people think. Because, again, the algorithms are going to navigate primarily through cameras and lidar and it’s not necessarily going to be relying 100% on a super high resolution data set and mapping a geospatial map set, to basically move that car around the roadway.

JF: What do you think this means then for urban mobility and how urban planners are thinking of downtown central business districts, and of course, ultimately, what does that mean for retailers or commercial real estate? Has anybody delved into that, that you know, of all this autonomous vehicle stuff is really going to change the way we view cities and shopping and whatnot?

BM:  Well, so Joe, I spent a lot of time talking about what are called “The Aces”; Autonomous, Connected, Electric and Shared vehicles. And people kind of think of them as independent trends but they’re really not, right. The vision I think most people have in the industry is not necessarily that you’re going to own an autonomous vehicle and it’s just going to be a replacement for the vehicle in your garage. The real benefit is in in most urban areas, having basically mobility delivered as a service. Where I can get from point A to point B, it’s going to be done through an electrified autonomous vehicle but it’s going to be shared. It’s not necessarily something that you own. And that really does change the way you think about cities. If all you’re doing is replacing a human driver with an autonomous vehicle and I still own it. It still goes from my garage at home, to my garage at work, it doesn’t really reconfigure the city. But if we’re now operating mobility as a service then it does! Because I need less vehicles on the road, I have to rethink parking, because we don’t need as much parking, because more vehicles are shared but I do need to rethink things like; where they do pickups and drop-offs, right! Now in New York… as you know Uber and Lyfts will stop in the middle of Fifth avenue to do a pickup or drop off, and that blocks traffic. If you have a lot more of those, that creates more problems! So, they need a dedicated pick up and drop off, things like that that are pretty important. So, that’s really our vision. I think you can have a dramatic impact on cities. You can reduce the number of parking garages. You can reduce congestion. Obviously reduce pollution through the electrification of the drivetrain and reduce the number of vehicles on the roadway, because now more and more are shared. But it is going to require some cities to rethink how they do things because obviously it’s a different world if you’re delivering the majority of your daily trips as a service through these mobility service providers, as opposed to individually owned vehicles.

JF: So, how does that then play out or how do you think it plays out? If mobility is a service, then do, we introduce an entirely different traffic management system, much like we have with the airline system, where there’s a centralized traffic management center and we’re platooning vehicles, we know when they’re going to leave, when they’re going to arrive. Is that too far in the future? Is that too futuristic or is that around the corner?

BM:  Well, so I think it’s around the corner. We’ve gone from a world where when we started INRIX 15 years ago, literally state of the art was flying a traffic helicopter, looking out the window to figure out what was going on. Now you’ve got 20 plus percent of vehicles that are now shipping are connected and they can report data to the cloud and where they’re at, and how fast they’re going, and that gives cities a lot of granularity that they didn’t have before. They can now understand what’s happening, not just on major freeways but side streets arterials, and really across an entire roadway system, even in rural and suburban areas. So, probably our most successful product in INRIX has been something called, Roadway Analytics, which is targeted at cities and helping them get a good view of what’s going on. Now you’re starting to see cities say. “Hey, now I want to influence what’s going on. I want to change traffic signal timing based on what’s happening. I want to be able to change toll rates based on what’s happening, and I want to be able to send signals to cars that are autonomous to be able to guide them in certain areas, or away from certain areas, that may be dangerous”. And that’s where we see a lot of the focus right now. So, we’ve created this product called, Road Rules, which helps digitize the infrastructure. We now have these safety alert messages that we now send to vehicles and autonomous vehicles can respond to certain events that are happening on the roadway. So, you’re starting to see that kind of two-way communication. Not just from vehicles and mobile devices to these cities but now cities are starting to want to send more and more messages, more and more directions, so I really do think that over the next, say three-four years, this will be a huge trend. And that is, cities being able to influence what’s happening on the roadway and guiding these vehicles to the point where ultimately, you’re going to view managing traffic just like you view managing traffic on a router, right. Where that router is optimizing the traffic, it’s optimizing how it delivers messages across an entire network and the same will be true in the physical network of a roadway.

JF: So, I’m going to change subjects on you a little bit, because I saw recently that you had been monitoring the Sturgis Bike Rally, and I think you put out a press release on that about how people are looking at that as a super spreader event with Delta. I’ve been through Sturgis several times. I used to live in South Dakota and oh, it caught my attention. How was that data collected? Was it just cell phone, or was it…it just seemed like an odd thing to cover but obviously that data was out there somehow.

BM: Yeah, so we collect literally data from hundreds and hundreds of millions of vehicles, or data from vehicles mobile devices and mobile applications. So, you know this is anonymous, we don’t know who the rider is or the driver, but yes, it’s collected from cell phone apps, and it’s collected from GPS devices, and it’s collected from cars. So, in the case of Sturgis most of it was mobile app-based data and we’re looking at basically trip patterns. Where people come from. Where they go to. And then, we correlated that on the study with where we saw basically cases breaking out, in terms of the Delta variant. So yes, those kinds of analyses where you’re either looking at an event like that or you’re looking at say New Year’s Eve celebration, you’re trying to understand everybody coming in and everybody leaving, or a big you know concert, or the Michigan football game last weekend. It’s really fun to look at those patterns and understand where people are coming from, how they’re moving around, and then using that to kind of correlate that with interesting things that are happening in society.

JF: Yeah, I looked at that football game and I thought, “there’s 106, 000 people in that stadium and I’m not sure I want to be there but good for the Michigan fans, I guess. Before I let you go and I want to take you back when you and I first met at the Location Intelligence Conference back in 2006, and I remember what you spoke about was location-based services, right, and how it would be the killer app. So, I’m going to ask you to look into the future and tell me what you think the next killer app is?

BM:  Well, so we’re very focused on you know these B2B scenarios, again the level of data, the density of information, the granularity, is very different than what it was even five years ago, let alone ten years ago. So, the killer applications right now are around the public sector. Specifically, understanding things you couldn’t understand before, right. How to optimize traffic signal timing. You used to have to hire a consultant and put tubes across the road and count cars and it would be a six-month study and then you’d create a report, and they’d retime traffic signals. Well, why can’t that be done dynamically. So, when the Michigan football game gets out on State Street, all the signals change dynamically to make sure that the traffic is optimized based on that kind of event. Not just if you kind of look at other scenarios, right, things like in real time. Retailers understanding who’s coming to their store if they run a promotion or if their competitors run promotions? How does that influence the traffic over at their competitor’s store? Those kinds of scenarios are now available because of the level of information and the density that now exists. Looking out in the future, again you kind of have to view the world as a digital version of Sim City, right. This video game that my kids used to play where you’d kind of build the city and based on what was happening you can kind of optimize what’s going on. That’s really the world we’re going into, where you have a digital representation of a city, you can see what’s happening and you can dynamically change things. Whether it’s the toll rates, whether it’s traffic lights, whether it’s sending messages to vehicles on the road to avoid certain areas, or to target certain areas where there’s lighter traffic. So, you really are in a case where the city planners, the city governments, can really optimize their infrastructure and optimize traffic on the road. So, I think those kind of scenarios are the killer apps over the next three, four, five years. I think the same thing’s true in enterprise, where enterprises take more and more advantage of location-based services to make decisions. Whether it’s hedge funds…we have a couple hedge funds that use our data to look at truck volumes in and out of distribution centers. So, things like Walmart distribution centers, and Target distribution centers, they’re using that to understand. There’s a very high correlation with what the actual performance of those companies are, so they can make stock bets. We see the same thing happening with hotels. You can actually look at all the traffic around hotels and get a very good view of how they’re performing before earnings announcement. So those are certainly the killer apps over the next five years and beyond that, I certainly do agree with you, that you’ll have pervasive at that point autonomous vehicles on the road. You’ll have pervasive mobility as service providers, where that digital communication from and to the car is the most critical element, right. So, all these things that we used to think about on navigation applications and Google Maps, and those kind of things. Those kind of go away because consumers really don’t care. The car starts to care but it needs that information in a very different way and at a very different level than like you said, what we’ve seen before. So, those are really, I think, the key areas over the next say five, six, seven years where we’re investing, and I think where the market’s gonna be going.

JF:  Yeah, just one quick question because you brought up an interesting point. Is there…what’s going to be the catalyst to make that happen? Is that going to be the pervasive use and implementation of a 5G infrastructure, for example? Is that going to support what you just kind of predicted? It sounded like that was where, maybe you were heading. That there is a nest, there’s a needed infrastructure to make that connectivity happen. Is that what you see?

BM:  Well, so 5G is nice. It decreases latency and obviously increases bandwidth, but the reality is everybody building an autonomous vehicle today is currently using 4G, and in most cases that’s going to be good enough for the 80% case. So, II don’t think that the deployment of autonomous vehicles is dependent on pervasive 5G networks. I really do think most of the processing is going to be done locally on the car itself. It’s going to be done through cameras and lidar sensors and edge processing is going to happen on the vehicle, and the amount of information being off-boarded can be done in a very different way. So, 5G is nice but it’s not a dependency. It’s not a precondition to the AVs on the road

JF: All right, great. Well, I guess we’ll leave it there. Thanks Bryan, that’s great insight and kind of gives us all a little bit of a dip into the future, so I really appreciate your insane perspective.

BM:  Great, my pleasure, Joe. Thanks for having me.

JF:  You bet! Thanks again for joining us on another On Point with Korem, and don’t forget to subscribe where this podcast is posted, whether that’s Apple podcast, Google podcast, Spotify or YouTube. And, if you like today’s podcast please leave a comment in the comment box. Join us again for another episode of On Point with Korem, where we’ll get On Point.