Innovative Uses of Driving Data for Safer Roads

Speaker 1: 0:08

Welcome to Auto Care On Air, a candid podcast for a curious industry. I'm Mike Chung, Senior Director of Market Intelligence at the Auto Care Association, and this is Indicators, where we identify and explore data that will help you monitor and forecast industry performance. This includes global economic data, industry indicators and new data sources. All right, Thanks everybody for joining. My name is Mike Chung, I'm the host of this session and I'm pleased to have Jeff Schlitt of Arity join us, and the reason why I asked Jeff to join us is because he and his team have a really great view on quantified driving data and behavior. So, Jeff, maybe you take a few moments to introduce yourself. Tell us about your role and what your company does.

Speaker 2: 0:58

Great, mike. Thanks, appreciate it and very happy to be here within the group here talking about mobility and telematics. As Michael said, jeff Schlitt from Arity For those that don't know, arity is a wholly owned subsidiary of the Allstate Corporation. We connect with about 40 million consumers today, collecting information on driving data, where they drive, how they drive, what risks they accumulate. And what we do at Arity is we analyze that.

Speaker 2: 1:24

We are, by trade, a data and analytics company and our goal is to make transportation safer, smarter and more useful for consumers. We take that information. We really understand things like the routes and routines that people do. We understand their behaviors like how hard do they break, how fast do they drive. We then work directly with the consumer, use that information to really make that consumer's life easier, whether it's providing safety.

Speaker 2: 1:47

Really quite an amazing company with the amount of data that we process, the insights and analytics we can use that data for and really impact society. So that's kind of the exciting part. My role as a solution engineer is I spend a lot of time with our customers really anything from insurance companies to mobile publishers, to the OEMs, to individual retailers and repair facilities that anything that has to do with mobility we work with and we try to find solutions that create tighter customer relationships, create a safer environment towards net zero power insurance to a more equitable price. So I work with all of those folks. My background is a data and analytics person, so I really live and thrive in this data world.

Speaker 1: 2:39

Fantastic, really appreciate that introduction and I think you'll be right at home in this new podcast. So kudos to our team for setting this up so we can dive a little bit more into data what it means, but also how it affects our industry. So one of the things you mentioned was collecting data, and I'm a little bit familiar with Arity, being a data partner of Auto Care Association. We get the VMT data from your group. Maybe, just for the benefit of listeners who may not be as familiar with your data, tell us a little bit about how that driving data is recorded.

Speaker 2: 3:13

It's a great question and today it's really important to recognize that the consumer owns this data right. It's coming from their vehicles, their interactions with mobility. So it's really important to start out with a very transparent way of working with the consumers, the way that we do that at Arity. We really get data from one of two sources. We get it directly from connected cars through relationships with OEMs, and we also get it from mobile phones, right, and those mobile phones are apps like GasBuddy, myradar, life360. These are apps that have a mobility background. One of the ways we do that is through things like family safety. Arity has analytics like crash detection from a mobile device. We can use the sensors on the phone to really understand whether or not there has potentially been an accident. We use machine learning, artificial intelligence, to really infer that those sensors and those readings may signify there's been an accident. And then what we do is we surface that data to the mobile publisher like like a Life360, and they can take action with it. They can get emergency response there. So, primary with us, when we work with a consumer, we're asking for consent to use the data from their phone both the sensors and location data to try to create an experience that helps them right. In the case of Life360, it's helping them stay safe and helping them understand where their family is.

Speaker 2: 4:25

I think there's a lot of discussion in the industry about transparency and how do you get that consent, and I think at Arity, we work very hard with that. Obviously, being a wholly owned subsidiary of Allstate, brand recognition is super critical, so our focus is to do it with the consumer first, right Meaning, if you were to log into any of the mobile apps that our technology is running in, you will be presented with consent forms that are basically clearly explaining why we're collecting the data, what we do with it. At any point, you can choose to opt in, opt out. Of course, some of that data can be used for advertising, and we try to do that, and again, you'll have the option to opt out of that.

Speaker 2: 5:01

But when used well, there's a lot of value for the consumers. We do a lot with things like fuel efficiency. We do a lot with just overall helping things become efficient. When you're driving a vehicle and maybe you're on your way home and a quick service restaurant wants to reach out, or a service repair facility knows that maybe it's time for an oil change and they offer you a discount because it's convenient for you. You'll be driving by their route. These are things that you can power with that data. But again, what's first and foremost is the consumer has to be comfortable to share that information and you've got to be very transparent when sharing it.

Speaker 1: 5:37

Oh, that's really helpful, and one of the things that I'm thinking about is as a consumer of your data and proponent of your data. I'm able to see it in aggregate total vehicle miles traveled for the United States and I've been sharing in presentations here at Connect that in 2023, we had maybe 3.15 trillion miles driven nationally, about 0.65% higher than 2022. And it gets interesting when you peel the onion layers back at a regional level, at a metro level time of day, and what I'm hearing is in terms of that data being aggregated. I'm not necessarily going to see Jeff Schlitt's mileage and so tell me a little bit about how sliceable that data is from an aggregated standpoint. Could you highlight that for us?

Speaker 2: 6:31

Absolutely, and I think you raise a really important point, which is we collect data about individuals, but a lot of that data, when it's used out in the industry, it's used in aggregated formats, like you alluded to. So one of the first things that we start to do and I think there's some really exciting examples, you spoke about it earlier Pre-COVID, I think everybody kind of knew how people drove, right, you drove into work, you drove home. There was a very defined morning rush hour, a very defined afternoon rush hour. Covid hit and what we noticed were patterns changed and what's unique? We're past COVID now, but those patterns are still different. You know, as an example, you know in the morning, the morning rush hour still runs about three to 4% less than it did pre-COVID, right, which means obviously a lot of people work from home. You start to see a new rush hour around afternoon hours, right, you know people at lunch break, since they're working from home, start to go out and move around. So in aggregate, we can start to understand that behavior change, right, and that means a lot. You know, really, if you're in the quick service restaurant space or you're in an auto service area, right, where you choose real estate, where you put those brick and mortar footprints. That matters.

Speaker 2: 7:36

Now people are out working in different ways, they're commuting in different ways and so in aggregate we can start to understand where is that shift in mileage right, and we've seen that shift go from heavily urban, concentrated, you know, during the day, to now you're starting to see that spread out and there's benefits to that right.

Speaker 2: 7:55

Density's down in most of the urban areas. But there's also drawbacks from that. Speeds are way up right, because there's less density on the roads. Overall speeds are up and I'm sure everybody who's out there driving knows their insurance rates are up and that's somewhat of a problem because if speeds are up, the severity of crashes are up. When the severity of crashes are up, the cost to repair those crashes go up, and that obviously has to work in conjunction with how the insurance carriers work. So all of these things are grounded in data and I think what's really exciting in today's society and some of the work that Arity does, we're able to use that data to create insights that can power the next generation of change, help departments of transportation put in the appropriate infrastructure, help those retailers in the automobile industry really start to learn about the new behaviors, the new driving styles and then tailor products for those consumers.

Speaker 1: 8:49

Super helpful. So let's kind of double click, as we like to say in some of those areas. So, if I'm hearing you correctly, with the additional driving data you're able to aggregate that and tell a retail location at this time of a day, traffic volume is up X percent and that could help them perhaps with advertising. It could be a digital billboard, for instance, it could be mailings. So tell me a little bit about some of the use cases of that, as well as kind of getting to the how sliceable the data is. Is it simply this many vehicles are going by? How much more color are you able to give that with regard to it? Is this type of consumer, it is this type of vehicle, Because I know it's anonymous, but at what level can it be kind of stratified for those types of use cases?

Speaker 2: 9:46

It's a great question, michael, and I throw out a couple of things here and great, I love the use cases, so we'll give you two. One thing that we work on is a product called Retail Traffic Analytics, and what's really interested about Retail Traffic Analytics? All of this comes together with the ability to use cloud compute and have unlimited compute. But the way that product works is you know we work with certain retail locations, right? So take any brick and mortar footprint again could be you know an auto repair facility, you know that has a chain background, and so they have hundreds of stores in the United States. What we will do is we will understand the footprint of those stores. You know geographically we call them shape files, right and really what that is. If you go to any map and you're going to see an outline of the property, including parking lots, it's like drawing a polygon.

Speaker 2: 10:32

It's like drawing a polygon around the location, right. So if you've got a couple hundred of these you know arity with our 40 million consumers that have entrusted us with their data we can basically start to understand statistically. That's extremely relevant. There's about 270 million drivers in the US today, so we're sitting on a large, statistically relevant portion of that. So if you are that retailer, we can start to understand traffic trends. We can start to understand that on the streets adjacent to your location, you're getting 10% of that traffic will stop into the store over this duration. But you know, maybe just down the street, you know your competitive competitors, maybe they're getting 12% right, and so we can start to understand that. Now, what's maybe something that we can segment into as well? Is it a time of day difference, right? Are these people getting them in the afternoon been into as well? Is it a time of day difference, right? Are these people getting them in the afternoon? Are they getting a more traffic? Because it's a right turn into the location versus a left turn? You know, those are some of the things you can start to slice and dice to really understand. You know, how is that traffic interacting with that brick and mortar location? What's even better is. Then you can start to enact some changes. Right, you know, maybe you want to do some additional signage and hold that thought We'll get there with your digital out of home. You know, maybe you want to do some marketing campaigns to attract people.

Speaker 2: 11:51

This is where our predictive mobility and our predictive commuting comes in. What that's all about is if it's convenient for a consumer, right, we know we basically learn what we call our routes and routines, and a routine is basically an origin and a destination. The origin may be your work location and the destination may be where you work out and we know that you do that three to four days a week and then, after a period of time, we learn the routes you take. Now one of those auto service repair facilities may want to market to you proactively and say you know you may be driving right by our store. Stop in $3 off on our oil change, you know, stop in for a $5 coupon on retail parts. That's the type of convenience that consumers really value. We don't actually know who that consumer is. We just know that there is a mobile device or a car that tends to go this pattern right, and we cut off things like private locations, like home and anything of certain areas of interest that are considered very personal. We don't track or work with anything there, but what we'll end up doing is we'll enable you to engage that consumer. We'll learn those patterns. We can help you understand that during rush hour you don't see as much, so maybe you want to run specials in rush hour, maybe you want to put some additional signage out and then you can measure that.

Speaker 2: 13:07

Going to the signage, it's really interesting. We partner with some digital out-of-home sign companies and what's really interesting with digital signs is they can work with multiple advertisers to display a digital ad and usually those spots run from anywhere from 10 to 15 seconds. And what's critical is they want to know who was exposed and how many people were exposed. Not because we're targeting, like you said, we're not targeting that individual, but what we can start to understand is the profile of driving behavior that passes by a digital sign. Let me give you a few examples of that.

Speaker 2: 13:43

So you know, let's say that we have people that tend to drive high mileage, right, and we know that this particular sign in Northern Illinois right that runs digital ads for, you know, an auto care repair facility is seeing an inordinate amount of high mileage drivers. We can actually index each of those signs based on those behaviors, what that does for an advertiser. If you are an auto repair and you're interested in servicing high mileage vehicles, you may run a digital spot and pick the signs that see an inordinate amount of high mileage drivers. And that's some of the signal that we can do in aggregate. We're not targeting an individual, we are basically just grouping and understanding of the behavior that drives by the signs and letting the advertisers really tailor that message to try to engage that consumer approach to how your group is approaching it.

Speaker 1: 14:47

So I'm relieved because I think there's always going to be that, or not necessarily always, but I can see where people might think Big Brother is watching right and, if I'm hearing you correctly, that data is being collected and aggregated so that that data is reported to organizations that can use it. But it's really at the overall level and you highlight and I'm hearing that correct yeah, that is correct. So one thing you mentioned was high mileage drivers, and I remember in previous discussions resident mileage, where the trips are originating from. So is that part of the calculus here where you're able to say, oh, we know where the trips are originating from because the signal is being captured through that app and we know that that driver is driving from, let's say, 40 miles away, 10 miles away. So tell us a little bit about that aspect of things, how you're able to separate residents, non-residents, high mileage, if you could.

Speaker 2: 15:45

Yeah, absolutely. And so when you think about these again, the way Airdy works is when our technology goes inside of a mobile app. It is all anonymous, right? You know, we know that there is a certain vehicle that is driving. You know, we know the patterns. So, for instance, to your point of the origin and the destination, we only wake up. The technology wakes up on the phone when it senses that somebody is driving a vehicle.

Speaker 2: 16:09

Right and again, like everything in today's world, this is artificial intelligence, machine learning, algorithms that take signals such as speed and take signals such as location Are you on a road? Because if you're going 20 miles an hour, you could be on a bike. So we have to understand you're on a physical road, somewhere that traffic may be. All of those signals get put into a model that basically make an inference that, yep, I think and feel pretty good that this is somebody in a vehicle driving. That's when our technology wakes up, starts recording information about those behaviors, and those are anything from you know how fast are you driving right, what routes are you taking and what roads are you on. And when you think about that again, it's just tracking a user ID. It's not an actual individual. I have zero personal information on that individual right, but what I do have is I have kind of the trails they take and how they move, and then what we're able to do is we can start to associate that you know these trips keep starting in this geographic area and so we can infer that that must be an important point of interest for you, such as an area where you live or an area where you work, and so we actually look at that about monthly. Every month we'll update because consider the you know individual individual that maybe vacations up in Northern Michigan in the summer. We may not see you if they're in Illinois, we may not see them in Illinois, but then we see them over and over again originating trips up in Michigan somewhere. So for that month they may be tagged as a resident of Michigan and what that allows us to do. Then, when you aggregate all that data up, we can see those shifting patterns, right, and we can even mark it. As you know these may be vacationers, right, you know summer vacationers at a cottage, and so we can start to understand you know how many mileage in Northern Michigan is due to people summering there versus people that live there year round. And that's important, because if you're a business and you're looking at inventory or out of stock or forecasting sales, those are all signals you could use. Again, very privacy safe. It's not about an individual. It's about individual behavior rolling up into geographic and temporal patterns that you can analyze. You would ask the question around, like where they come from.

Speaker 2: 18:14

And this is another very important thing for brick and mortar retailers is understanding the reach. If you want to put a new footprint of a brick and mortar store, what you want to understand is, let's say you have 100, 150 stores already, you'll know that the average distance that you get people to come to your store is maybe 10, 12 miles right. But when you see people getting further out, right, and you're seeing enough that you know maybe your competition is starting to pull that business away. Those are some of the signals that may say there is enough population to drive the sales you need and there's a convenience in that you can shorten that window, which in a lot of cases leads to higher conversion rates in those retail footprints. So all of that data, all of that information is driven by the data that we collect, right, and it really gives you that opportunity. There are opportunities to use our data individually.

Speaker 2: 19:03

And again, that usually gets into things around insurance pricing.

Speaker 2: 19:05

But again, it's extremely important for any consumer to realize that none of that happens without their consent.

Speaker 2: 19:11

Right, and even to the point where, if you're, let's say, getting an insurance quote, you would be asked exactly. We are what you call a credit reporting agency, just like TransUnion and Equifax for your credit scores. So, being a credit reporting agency, you know we are what you call a credit reporting agency, just like TransUnion and Equifax for your credit scores. So, being a credit reporting agency, we are subject to all of the fair credit requirements. So if you go and you are asked do you consent to use the data to price and insurance, you can say no with no impact. You can say yes, you consent to that, and we are required to tell you why we got these scores, what we understood. So really, to me, that is the right way to do this. Right it is if you don't engage the consumer, not only will they be suspect of you, but they also want to understand the value in it. Right? So we take a lot of pride in that engagement with the consumers.

Speaker 1: 19:57

I think about. When I signed up for a safe driving app to track my hard braking, my sudden accelerations and late night driving, I opted in. The default was not that you have to opt out. So it sounds like if an individual wants to have a direct impact on their individual bill, they physically have to opt in. Correct, but if I may ask this as an insurance organization, I would imagine that's just a. Really the data that you're getting is very helpful to understand how the population is driving, to assess risk and perhaps how that can kind of sharpen the actuarial tables for writing policies. Is that fair to say?

Speaker 2: 20:41

It's absolutely fair to say. And, again, I think one of the most important things is equity and pricing. Insurance is a really interesting industry. I've worked in and around it for many years and insurance is one of those. It's a highly regulated industry, right, and the general sense of insurance is to create an equitable pricing framework. Right Now, everybody who pays an insurance bill probably has a hard time believing that, but it truly is that way. Every state has a department of insurance. Every department of insurance is trying to make sure that if you're going to raise your rates right, there's justification for doing so. If you're going to price one person differently than another person, it has to be done in a balanced, equitable manner.

Speaker 2: 21:20

So things like credit scores have been used in insurance for years and it is very predictive of loss. Now, why is it predictive of loss? That's really a hypothesis that you've got to kind of unpack to understand. It could be predictive of loss because of the fact that people with higher credit scores may live in the suburbs, right, and suburbs tend to have less density of traffic, and if you have less density of traffic, therefore, you may have less losses Still equitable. But if you're an individual who lives in a city, maybe in a very urban area, and you're a very safe driver, you're actually, to a certain extent, being penalized for where you live there.

Speaker 2: 21:56

Where driving data becomes super important is it's rating you as an individual based on your driving behavior. You're in control of that right and that's why we became a credit reporting agency because you get to dispute, you get to understand your credit. That is your right. But it is a fair way. If you choose to drive in risky manners, you should pay equitably more for that right. If you want to drive 100 miles an hour the expressway right there, is a higher chance that you will get in an accident, and when you do, those costs have to be passed on to policyholders. So it's much better to have that cost passed on based on risk than it is spreading that cost across everybody. So, truly, telematics is a very equitable way of doing things. But you're right, it does mean we have to understand how you're driving, and that is a scary proposition.

Speaker 3: 22:43

This is DTP IT Director and Sustainability Committee Staff Liaison at Auto Care Association. Are you passionate about shaping our industry's future? Join an Auto Care Association Advisory Committee and make a real impact as a volunteer. You will drive innovation, tackle key challenges and collaborate on cutting-edge solutions for the entire supply chain. Don't miss out. Join us at our upcoming Leadership Days event to start making a difference. Learn more at autocareorg slash leadership days and find information on current committees at autocareorg slash committees.

Speaker 1: 23:30

And I think from if I go back to Economics 101, because we were talking about sons in college taking core classes it's the information asymmetry problem. Right, and with this technology available to fill in those data gaps for lack of a better word it allows a company to price its product better. It allows a company to price its product better and in a field that has many competitors, it's kind of just upping the state of technology and the state of the art so as to have a stronger foundation for making those pricing decisions.

Speaker 2: 23:53

That's 100% accurate and I think what the consumer has to realize in this equation and regardless if it's insurance or if it's how you engage with your repair facility, the more information that you can provide somebody, the better you know. Imagine if you go to your doctor and you're not feeling well. The first thing they ask you is what are your symptoms? How do you feel? They want you to share data with them and insight so they can give you a diagnosis. So what we have to really focus on is how to build trust with the consumer such that they're willing to share that information. And that's why, when we work with these mobile apps, when we work with the OEMs, we want to make sure the experiences that we put in front of the consumer right our experiences they value, they are in control not us on how that data is used right and, quite honestly, we give feedback right.

Speaker 2: 24:38

If you are driving using risky behaviors, you know we want to give that feedback back to you. We want to give you the opportunity to improve those behaviors, to lower the risk. So a big part of what we do, you know, whether that's through the departments of transportation and in our public sector areas, whether it's the coaching experiences that we put inside of these apps. It is really to educate the consumer that they are in control. Right, it's not big brother watching. It is your data that you can control how it's used, and you get to change your behavior to impact what you pay for insurance Makes a lot of sense.

Speaker 1: 25:11

One thing that you alluded to were things like high-speed driving is up. The use cases of people vacationing in northern Michigan Broad open-ended question. But what are some of the changes that you're seeing?

Speaker 2: 25:26

Yeah, you know the speeding and again, a lot of these are hypotheses to be proven out over time with the data but the speeding, I do believe and this is Jeff's opinion is that because driving spread out more and people are working from home, you're on roads that are less populated, driving spread out more and people are working from home you're on roads that are less populated, you know, and if you're on roads that are less populated, you probably feel like you know you can push that accelerator a little bit harder, right, nobody's around you. The risk with that is we're also seeing continually higher distracted driving rates. Distracted driving is up over 32% from pre-COVID, right, which again I think emboldens these behaviors that nobody else is maybe immediately right in front of you, so you're just going to peek right at that phone, you're going to look, take one more look, and so we take that very seriously because, again, that is one of the largest risk factors. And so, instead of just calling it out, really what we want to do is start to change that behavior. Right, and you know whether it's the way. Predictive commuting is a great opportunity, right, you know we don't want to advertise to you while you're driving. Right, we don't want to make notice of that. That's why we built predictive commuting, because what we want to do is say I'm pretty sure you're going to be on this route, so why don't I go ahead and notify you before you get in the car? I'm going to make it more convenient and also safer to have that conversation with people.

Speaker 2: 26:47

You know, as people change, I think one of the other behaviors that we're seeing is you know the length of trips and where people are going and how they're doing it and when human behavior changes. Right, you get to adapt as a business to. You know, meet the consumer where they are Cliche but it actually is true. Right, you know when people decide hey, you know, I'm bored of working out of the house, I'm going to go to the Starbucks. Right, you know, maybe you can influence that behavior. Right, people do like to socialize, they like to get out, and I think that's where this data, you know, with some of the changes post-COVID, is valuable. Right, we can actually engage customers in different areas, we can create new situations, and that mobility data really helps you understand what that consumer's preference is. Right, and it's not all one right, different consumers value different things, but I think that's where this driving data can really help you understand some of those changes, and then again you can meet the consumer where they are, with goods and services that they value.

Speaker 1: 27:39

Thanks for that, jeff. Two things came up as you were talking. The first one was the distracted driving going up 32%. If I can just go down a little bit of a rabbit hole here, can you describe, define distracted driving? And part of the reason I ask is this If I'm driving my 2012 Nissan, which has Bluetooth, but that's my wife's car, I don't connect my phone to it. If I pick up my phone and look at it, maybe it unlocks. To me that's an obvious distracted driving. If I had a rental car that was maybe a 2021 and I paired my phone with it and I could see message coming in from so-and-so, you could hit a button to ignore it. Or I'm driving or not engage it otherwise. Or if the phone rings and I hit the button on the steering wheel to answer it, which of those might count as distracted driving, and can you just tell us a little bit more about what constitutes it please?

Speaker 2: 28:38

Yeah, it's a great question. And again, you can take data and display it anywhere you want. So, at Arity, the way we describe distracted driving when initiated from a mobile device is kind of what you alluded to. It's an unlock of the phone, okay, it's followed by movement. So phones have accelerometers inside of them, so we sense that change in acceleration of the phone as it moves through the air. So the unlock event, followed by the accelerometer right, is what we typically will call distraction, right, and we have to sense it for a short period of time, right, you know, we kind of give everybody the opportunity.

Speaker 2: 29:13

We also look at the speed of the vehicle moving, right, if you're stopped at a light, right, we won't necessarily call that distracted because the vehicle isn't moving okay.

Speaker 2: 29:21

So you know, all of those signals are processed and again, we also look at them in the context of society, right, in the context of insurance loss, and when we tie that data together we can start to understand that this type of movement on the phone is really what we consider distracting, because it's highly predictive of loss severity, loss frequency, you know, accidents happening, and so it's never just one loss severity, loss frequency, accidents happening, and so it's never just one set of rules that stay forever right. The really important thing with analytics and statistics and machine learning and artificial intelligence is it's constantly learning and it's constantly getting smarter right at to what the distraction means, and so I think today that's how it's defined. Tomorrow it may be different and again, as society changes, as technology changes, we'll adapt those models. But again, really what we're looking at is that signal of risk and in our world that is an insurance claim or a loss on a vehicle. We look at that and we say what behaviors led up to that, so we can help coach those behaviors and become safer.

Speaker 1: 30:20

Is there an industry standard definition? Because if I'm National Highway Transportation Safety Administration, I think that's what NHTSA stands for I'm thinking about tell me the difference between somebody physically scrolling versus texting, versus initiating a call, versus X, y, z. So is there a standard in the industry for the different types of distracted driving and are you able to detect that?

Speaker 2: 30:44

Right? Yes, as far as I know, there is no standard yet and I think it's incumbent on companies like Arity and maybe the insurance sectors. There are conversations starting, right, if you've heard of the net zero initiatives and Tell us a little more about net zero. Yeah, net zero is one of those things in the Department of Transportation and within municipalities and the public sector that are trying to say how do we get to the point where there's no accidents, no deaths, right? You know, we have really intelligent technology with these cars, whether it's the advanced driver assist systems they call that ADAS right. Whether it is, you know, the self-driving vehicles, whether it's the material we're using in the car, all of these things have the potential to lead to, you know, zero accidents, zero deaths right. But again, the consumer has to balance that with the desire and the fun of driving a vehicle, right, like everybody loves to drive a car. You know you turn 16, you can't wait to get your license. But I think you can have both, right. You can have a safe environment, and that's the promise of technology, right, and? But I think you can have both, right. You can have a safe environment, and that's the promise of technology, right. And so I think what you're going to see is you will probably see the industry standardized on this.

Speaker 2: 31:56

The government has to say location data in general is a very hot topic right now. It's a hot topic in Washington, it's a hot topic in insurance, a hot topic in advertising. I don't think it goes away because it's extremely valuable. I think we need to make a contract with consumers that they value right. So and again, you need companies like an Allstate Corporation, and that's what I love about working for Arity. Being owned by Allstate gives us the investment and the opportunity to do it the right way, right. And again, I think, if you look at the leadership at Allstate, right, they're heavily vested in safety. Right, you're in good hands. That's the nature of their motto and we live into that, and I'd like to believe that that's where we need to take society length of trips and earlier you talked about.

Speaker 1: 32:51

We can detect because I'm connected to this phone and you can see for this device, at least if I'm hearing it correctly, I have, my patterns have changed and now I'm driving in Florida. Maybe he's not resident, right, so you can tie it to an individual, but you don't know that it's Mike Chung, age X, gender, et cetera, et cetera. So I'm seeing you nodding your head. It sounds like I'm on the right path and tell me a little bit more about the length of trips. You mentioned earlier that there could be combined purposes. Could you just tell me a little bit about what you've seen broadly over the last few years?

Speaker 2: 33:29

Yeah, I think what you see with the length of trips is they tend to become longer, right, more mileage is being accumulated, and I think that has to do with, again and this is really important in the retail space, right, because we're starting to understand changes Everybody has to go get groceries. You've got to go charge your car or fill up your car, um, you've got to go get, you know, material and clothes and go to home depot and do the things that you do on a daily basis. But when you're working from home, right, people tend to try to go to a convenient method of of purchase. Anyways, right, if I'm on the way home, I'm going to hit the fuel station, the grocery store in, you know, my neighborhood uh drugstore to pick up a prescription. So, now, what we're getting to, though, is when people venture out, they're learning new patterns themselves, right, so we're seeing longer trips. We are seeing people venture out at lunch hours and run errands that they typically would do on the way home, right, and so, you know, the length of trips are increasing, right, and, again, I think that has a function to do with, just, if you live in the suburbs, right, houses aren't usually right next to you know retail locations, but when you're at work, there's retail locations all around that, because you drove to work, put your car away, then you walk to lunch. You walked over to get a haircut. So we're seeing that change in behavior right Anywhere from. If you start to look at you know parking lots. When's the last time you've gone to a parking lot and it's full right. Those are things you're starting to see.

Speaker 2: 34:51

We even and I'll give a great example when the Key Bridge incident happened in Baltimore. What's fun is almost immediately we can start to analyze what happens in that scenario. And I think earlier, about a year earlier, on I-95, they had to shut down I-95 for four or five days and I-95, major thoroughfare and we were able to look that was just outside of Philadelphia, correct, outside of Philadelphia. We were able to look at the impact to the infrastructure. Where did that traffic go? You can't go on I-95. Where did you go? And what we saw were times ballooned by 20, 30% when I-95 happened. What typically would be maybe a 10-minute ride is now 15, 18 minutes, right, just ballooning times. And you can see the density on the side routes that weren't designed to handle that.

Speaker 2: 35:43

So that was a real difficult point for consumers. With the Key Bridge, the infrastructure there really kind of absorbed that change quite a bit. Now that could be the type of traffic that went there across the Key Bridge, but what we did end up seeing is that the infrastructure in that area absorbed that change pretty easily and again that's a success story for the Departments of Transportation and that's kind of the promise of what this data can do. I mean, consumers have a very short fuse right. If you're sitting in your car and you're stuck in a red light, your anxiety goes up, you're frustrated and if we can plan around that and we can plan with a little bit of leeway in some of this infrastructure, that really goes a long way in society. So this is some of the real positive benefits of having access to this kind of data and, again, making that contract with the consumer that you will benefit from the sharing Terrific.

Speaker 1: 36:37

One thing I thought of as we were talking was possible noise you talked about you might be going 20 miles an hour. It might be a bicycle. I remember signing up for that safe driving app and it would say it would thought I was driving but I was taking the subway and I had an opportunity to say I was not driving. How about if you're taking a car service, you get picked up in a Lyft or Uber. I remember I think you had shared the acceleration is different in the front seat versus the back seat. Can you tell us a little bit about that noise reduction process?

Speaker 2: 37:09

Absolutely. So you'll start to sense really quick how mathematic and how hard it is to actually glean some of the signal. But with the advancement of AI and everything else, these are positive opportunities for it. So we have something called vehicle mode, right, and vehicle mode is a methodology where we try to look at and determine, you know, are you on a bus, are you on a bike, are you on an airplane? Some of these things are easy, right? You know, if you're at 10,000 feet, going 300 miles an hour, I'm pretty sure you're not in a car anymore. Okay, so those are pretty easy inferences to make. But take rideshare like you said, that's actually pretty difficult, except if you look at.

Speaker 2: 37:46

And what's really interesting if you look at, and what's really interesting if you visualize it, if you look at that vehicle, that rideshare vehicle, and you look at what it does, the pattern on a map is really, really gives it away. I mean, it goes somewhere, it stops, it goes back somewhere, it stops. It just looks like it's driving in circles, right, so that pattern can be recognized by machine learning. And when you recognize that as a consumer, if you get inside of a vehicle, you know, and that vehicle is zigzagging in a pattern that looks like a ride share. We can market as that.

Speaker 2: 38:15

Now we always give you the opportunity to override it right, meaning that a lot of the experiences you know we can. We call it tagging the trip. You know the consumer can go in and say no, no, no, I was on a bus. You, no, no, no, I was on a bus, you know they can click that. That is great too, because it actually helps us get better at these analytics, right. That's that continual training and learning we do, but we're pretty good at it right. Our accuracy at figuring it out pretty good, and it's again, it's the positive aspect of data is extremely important in understanding behavior, right, and when you understand behavior, you know when used for the appropriate manners, it really does make people's lives better, and that's what we do.

Speaker 1: 38:50

Terrific. One thing that I'm thinking about here is earlier you talked about advertising. One way, two way, so tell me what that means. Does that mean my data cannot be used for advertising purposes? Does it mean advertisements that are shown to me? Can you just distinguish between the two for us?

Speaker 2: 39:10

And the advertising industry is highly regulated. So I'll start with that Meaning that, again, as a consumer, you have a ton of choices, right. It's also a extremely complex technical environment, but it's super interesting. So we'll go back to that predictive commutes and I'll even give you some statistics again. We do a lot of what we call in-app advertising, right, and in-app advertising is when you sign up with an app, you know, pick any of the apps that our technology is running in. You'll be asked do you consent to sharing your data for uses in advertising? If you say yes, right, it means that we will display ads in that app. You know, that's one way most mobile publishers make their revenue. Right, in order to use that app, whether it's a weather app, whether it is a family safety app, a lot of those free versions are powered by advertising, right, so there's a benefit to the consumer right there, which is you're getting free software and in return, you've got to look at the advertisement. I mean, the whole purpose of mass media and television is built on that model. Now, what's really unique, though, is it's not just about, you know, what we call pray and spray, right. It's not just put the ad up and hope that the people looking at it are interested in the product. Now, when we start to understand those behaviors, we can actually tailor these products and again, the consumer will value this.

Speaker 2: 40:19

I'll give you an example. We worked with a fuel retailer where we were doing our predictive commutes, which again is us making a prediction that you may be driving by one of those locations in the next hour, two hours, three hours, and again the purpose of doing that is we know what routes that phone or that device is taking and so, within the app that you've entrusted to advertise to, that advertiser can basically say I want to contact people who are going to be driving by one of my locations this afternoon. So you may be just be using a weather app and you may get a coupon for food or auto repair or whatever it may be. You may look at that and go. You know what. I'm going to drive right by that location and I'm going to get $5 off. This is a good deal. I'm happy with that.

Speaker 2: 41:04

That's what our advertising can power and, amazingly enough, when we did that with this fuel retailer, we ran this for 60 days, I think, on about 100 locations and over that 60 days. When you compare it to the 60 days when you weren't running that predictive ad you were just doing the pray and spray we got a 39% higher conversion. What does that mean? It means that the people that saw the ad that was convenient converted bought a product 39% more times than they did when it was just putting a generic ad up. Right, and again, what that tells me is the consumer valued that experience and that advertisement. Right. Again, you get to choose whether or not you want to see those ads.

Speaker 1: 41:47

Right. And then that's on the mobile phone and, similarly, on an outdoor billboard you have stats on how many people are driving by it, Correct? And that can just fuel the targeting as well. That's right, Yep, Fascinating, so I think.

Speaker 1: 42:06

One last question I have for you as we close up, and I'll say this it's been fascinating talking to you. I feel like we could talk for hours. We should probably do this again. So much thanks for joining us here. What do you see in three to five years with regard to data that's going to become available? And I'll jam in one more question here, because when I gave a presentation here at Connect, I showed the inferred high impact collisions month by month, and it was a slightly positively sloping curve right. And the question came up does that account for ADOS? Because this man conjectured if there is ADOS, I would expect that to go down. So I was thinking about that. Is there a way to mash those data sets together? So kind of a two-part question Is ADOS helping? Is there a way to prove that? And then what's the future?

Speaker 2: 43:11

Yeah. So it's a really good question, by the way, and I think time will tell. What I will tell you is, again, the average age of vehicles on the road, I think today is around 12, 13 years, and so when you look at that, there's a lot of vehicles on the road that do not have 8S, right? The second thing is a lot of people turn 8S off, right? I have a Honda 2019 that has 8S. I love the system, but you know, it was also kind of like Gen 1. And you know, at times it's like, okay, I'm just going to turn it off right. So you've got those two things at play. So what is the real number of cars doing that? It's actually, I'd say, a scary proposition, right, because if things are still rising and we have safety features preventing it, it's probably rising at a higher slope or even more exponential than you think, because ADAS, I guarantee you, is working, but we're still seeing them rising in high severity collisions, right? So parlaying that into the three to five-year plan net zero you just can't complain about that, right? There's nothing negative about trying to save lives, right? You know, I have four children. Two of them are driving, two are going to get their permits this year. Yeah, I want them to be safe, right. So anything around family safety to me is a net positive.

Speaker 2: 44:19

So, using data, building that contract with the consumer to be able to build a safer world to live in, you know? Yes, three to five years, I do believe. And there's a lot of conversation around the use of that data for insurance and pricing. I do believe, from an equitable position, it's a more equitable way to price. It's way more equitable than credit, than your gender, your age, those things today are being used. I think the way you drive you control it. If you want to drive riskier, you should pay more. Used. I think the way you drive, you control it right. If you want to drive riskier, you should pay more.

Speaker 2: 44:50

So I think in three to five years, the insurance industry believes in this. The numbers don't lie. I think what we've got to do is figure out how to get the consumer confident with it. So I think over the next three to five years you're going to see a continued effort to bring that consumer along. Whether that's the way that we set up the businesses, whether that's the contract we make with them, whether that's the education, I would say that's important.

Speaker 2: 45:13

And then I think outside of that. Quite honestly, I think ads that are tailored towards convenience, that are meaningful to a consumer, are way more valuable than the pray and spray. So what you may. Actually, in a perfect world, maybe you're going to see less ads, but the ads you see are going to be more worthwhile. Right now the advertiser is going to pay more for that, but in general, their spending will stay flat because they're not going to spend as much on the pray and spray and they're going to use data like this to really target the consumer with what they want. So those are maybe utopian views of the world, but I do think that's the promise of this data and analytics. And again, it's just important that we do that collaboratively with the consumer so they're coming along with the journey. They're not being forced upon.

Speaker 1: 45:56

That makes a lot of sense from an efficiency standpoint, as well as a customized content and higher return on investment, because you have to do the cost benefit analysis as you're alluding to. That's right. Well, jeff, so great to have you. Always a pleasure to talk. I hope we can. Thanks for tuning in to another episode of Auto Care On Air. Make sure to subscribe to our podcast so that you never miss an episode. Don't forget to leave us a rating and review. It helps others discover our show. Auto Care On Air is proud to be a production of the AutoCare Association, dedicated to advancing the auto care industry and supporting professionals like you. To learn more about the association and its initiatives, visit autocareorg. Thank you.