EP 5 - Philip Nothard (Cox Automotive)

Hi everybody and welcome to the blind spot. The show that gets the real picture today is automotive markets from the people closest to the data. I'm Alison Campbell, the Chief Growth Officer at Market Check and today I'm joined by Philip Notard, Insight Director at Cox Automotive Europe and one of automotive industry's most respected voices on market insight, vehicle remarketing and strategic forecasting. With more than 37 years experience across retail, fleet, remarketing and automotive operations, Philip helps businesses, industry leaders and policy makers understand the trends, shaping the market and make better informed decisions. At Cox Automotive he leaves a development of insights and analysis that support customers in a rapidly involving automotive landscape. Alongside his role at Cox Automotive, Philip is chair of the Vehicle Remarketing Association and a member of the European Remarketing Advisory Board giving him a unique perspective on the challenges and opportunities facing the industry today. He is also a regular commentator speaker and a podcast host and he's known for turning complex market data into clear actionable insights. Philip, welcome to the show. Sounds like a good job that one. Yeah, I do, you'd start that way out there. Looking back at your career, Philip, what has been the most significant change in how automotive businesses use data to make decisions? I think in reality, I mean, if you were to look the market, let's say 30 years, you're all close to my start of my career. Data was largely retrospective. You're looking back at historical data. Today's business, you know, you expect data to be in real time. There's no historical data. It doesn't really in today's market tell you to anything. You know, we move from relying on experience, instinct, to combining expertise with vast amount of data and insight available in the market. And I mean, if you think about the biggest shift in, you know, it's not necessarily so volume, but it's accessibility of the data, you know, the data that is now available to more decision makers and accessible across the market than ever before or however, better data, better access doesn't automatically result in better decisions as we will certainly get onto. You know, I mean, the real challenge in today's market is, you know, really as you alluded to earlier is it's interpreting that data correctly and really understanding in its context. So I suppose in summary, you know, we've gone from what a shortage of data to an abundance of data, but the challenge today isn't necessarily finding the information. It's more about knowing which information matters and what to do with that information now. And it's in quite recent, isn't it? Because I remember 10 years ago, I would say I have all this data of what's for sale and the pricing and big deal of things to say, I know, we rely on Bob because Bob knows what sells and they've gone from Bob to all the data that could possibly have and now they can't decipher what the data's telling them. Yeah, exactly. I think that's that's the way I see it. Yeah, there's lots of data, lots of information, lots of insight, but it's, you know, it's what's the value of that data, what to do with it and what is that data actually telling you, you know, we talk a lot about the soul walks. What does that mean? What do we, what do I need to do? What do I make my strategy decisions on the back of that information? Yeah, and because you work across the entire industry, you know, you have a lot of access to a lot of different people, lenders, manufacturers, dealers, fleet providers, valuation providers, which part of the industry has become the most sophisticated and which you think still has the furthest to go. I think in reality, I mean lenders and valuation providers have made huge progress in terms of their offerings and accuracy of data. I mean, you know, the market now depends on accurate data. That's the key, whether it is, you know, particularly around that, that valuation market, I think giving respect to the larger dealer groups, you know, they are becoming increasingly more sophisticated around stock management and pricing of that stock. And, you know, as you alluded to before, you know, Bob is no longer good enough to make those decisions on scale with the speed of the market, the movement, and we've mentioned real time data, you know, that, yes, you need that experience, you need the understanding of the market, but you need that real time data. I mean, fleet operators, you know, they're starting to use predictive analysis more effectively than ever now. Probably the area with the greatest opportunity remains the sort of the end-to-end integration across the various sectors. I think too many organizations be it the lenders, the valuation providers, the fleet leasing sector, you know, they're still optimized their own part of the journey rather than looking at the total life cycle end-to-end, and that's the bit with it, you know, you need to understand what the OEM strategies are from a production perspective all the way through down to the end use. So I think, you know, to sum that up industry, you know, the does collect a large amount of data to say, but we're still not always really connecting the dots between, you know, acquisition, ownership, marketing, and disposal. I think all those areas of the mentioned, I think there's still quite a lot of work for each of those operators to get to that point of connecting the dots. And what concerns you most, like looking at the US and the UK, but what concerns you most about the decisions that are being made in the automotive ecosystem? I think if you look at it, you know, speed is increasingly being valued over the understanding, you know, it's getting that data to do something with it, AI, you know, decisions are still being based on, you know, lagging indicators. I think that's the key and we get back to that real-time information. You know, the US, you know, they've got different business over there, you know, size, the scale, the dealership dominance in the US compared to the OEM dominance in UK and European markets is a distinct difference between those two ecosystems. I mean, in periods of transitions, such as electrification, historical trends, I mean, some of that becomes less reliable if you look at the transition to electrification, the US now looking at UK Europe as to what will come over there. If you think about the Chinese manufacturers, the US are looking at the UK and Europe as to what could face them down the line. But there's a lot of difference. It's no longer, you know, what happens over the water we'll get in whatever numbers of years. So I think that market is a big, big shift. So I think in reality, you know, markets have changed you faster than many models built, poorly predicted at the time. Yeah, I think you can see that from the live data of Chinese franchises. You know, there were a handful that they're a small part of another of somebody else's showroom. And then, you know, you and I've talked about this off camera, but there's this growth in these Chinese OEMs and you see who actually is the same Canada, but obviously you don't yet, you don't see it in the US. And you know, we were doing some work on how that would look at the US if that was applied, but before you only know if you saw them, it would be so slow to find out how they were growing. And now it's like they open and you know, next week. And so you can, you can almost do this in real time. I think that's the landscape into, I mean, it's moving quickly. And, you know, think from the set, so you're trying to manage AI data, transition to electrification, new entrance coming in and everything else that's hitting you. Where are the priorities? Where is the focus? And that's where you know, data is the element that adds that, you know, realism and real time information for you. Yeah, because I was thinking about this from the from the blind spot put of you, the name of the podcast. So let's explore some of these areas. Why also was it still operates with these blind spots? And what a better data ecosystem could look like in the UK and the US over the next five years. So if we start with the blind spot, when you look across automotive, where do you see the biggest blind spots in the data that people are using to make decisions? I think if we think about, you know, we are still in a supply and demand market play. We're still in our set vehicle in the middle of all this. You know, understanding true vehicle level demand is reality, you know, right? We can look at a high holistic level, but actually, you know, when you're the operator, what is the demand today? You and I talk about that European market and, you know, Germany, France, Italy, Spain, they all operate differently. They have different localised, you know, legislations and regulations and, you know, data protection and everything else. So it does. I think there's a element of understanding that true vehicle level demand. We've got to understand real world retail pricing versus advertised pricing. We've talked a lot about vehicle listing, but actually, what is that true transactional price when it gets to that level? EV ownership behaviors, as we use moving to that EV electrification marketplace, you know, what is real happening is a very quick moving environment. So I think there's a lot there, and really understanding those macroeconomic factors that influence vehicle choices, influence consumer decisions, influence buying decisions. So I think, you know, the biggest bias plot isn't what happened last month. It's understanding what's happening to the consumer demand today. It's the here and there is real time information, real time data that you need to make those informed decisions on. I suppose an obvious example that would be government changes the incentives to buy an EV. They change that, they announce the change, and all of a sudden, all these dealers have got all these EVs. Yeah, the incentive is slash by 50% and that changes the consumer behavior. Is that what you mean by those macroeconomic issues? The more recent with the Middle East conflicts and the changes at the point crisis, you know, it's whatever your stock profile was at the start of that conflict, changes in conflicts, potentially with the consumer demand and the consumer behavior is because they're looking at the market differently, they're looking [BLANK_AUDIO] ownership behaviour differently. So that's that's where we need to shift. So then you'd have dealers who, if they're fast moving, they're, they're clearing the stock and then they're bringing new stock in and they can react to that. But if you've got a dealer with a slower, slower to stock turn and behind that dealer is a lender who has to lend on what the dealer has, who's got the best view of them, of the market there? Well, I mean, that's again where, you know, we go back to the early point about data in those different sectors. I think, you know, at all level is you have to have that real-time lens of the market place, you know, because it's where is, from a lender perspective, where is your risk, where is your exposure, where is your opportunity, from a dealer perspective, you know, what do I need to move out of fairly quickly, what can I hold and take advantage in terms of getting a better rate to return up my investment? So there's a, there's got to look at it all the time and I think this, you know, you and I have been in the industry long enough, the 45 day, the 60 day age policies, the 90 day age policies, these daily. And sometimes it can be in the same day where you've got to react, you've got to react to real-time and be responsive to move that vehicle on and get that asset paid. So why do you think the industry became comfortable historically with making decisions with averages and periodic reporting and historic trends, rather than real-time signals? I think you and I see all the time, I mean average years are easy to communicate, easy to understand, to tell potentially, you know, an unrealistic story and, you know, we can see that in today's market, you know, where we're seeing double-digit growth in new-car registrations, we're seeing single-digit growth in some of the markets across Europe, in what is a challenging political and economic environment? And it doesn't make sense why we should be seeing this significant growth in new-car registration numbers on the backdrop of a very challenging, yes, cost of living economic climate. So and it's because the averages and the headline figures don't tell the whole story. So, you know, automotive, we've developed with a, you know, that's physical supply change, slower information cycles, you know, the banking and insurance sector, you know, the digitiser earlier and built in data infrastructures around transactional data, but automotive is only now probably in the last five years now starting to think about real-time capability, seeing elsewhere in other sectors. So, I think, you know, average is a useful, but no vehicle is average, no customer behavior is, is not anymore, or customers in average, so everything's unique, but we always look at averages. Yeah, I guess that started with when you had, what was it, many said, there's a billion combination of options, try pricing that one. Well, yeah, the Volkswagen was the same with their, their, and many are the same, yeah, it's everything's unique. Yeah. So, if you were sitting inside a lender or a dealer group or, I know we am, which blame spot would you prioritize to fix, sorry? I think the, you know, visibility of actual retail demand and transactional activities is key, because you can respond to that, you can, you can make pricing decisions once you understand in real time what the retail market is, is truly looking at, I think, everything ultimately starts with consumer demand, because that's the end market place for, for product in, in our sector. So, you know, if you don't fully understand demand, then everything downstream decision becomes more difficult. So, you've got to understand what the buyer market, what the consumer, be it business or private, what their behavior is at the minute, and how do we service that demand? And that's probably trending months before it gets to a forecourt. Yeah, you can, you can identify early, some's quicker than others, but you can see those early trends and, you know, I mean, conversations now around, okay, that's today from the Chinese New Veterans and some of their behaviors, but then go back to China or something in China. Is that what the long-term goal is in the UK and European markets from the Chinese New Zealanders? That might be five years away, okay? But even if you think about the pandemic period, we know it's way to 26 and the shift that we've seen since the pandemic has been huge for the sector. So, five years isn't that long in our sector at the minute? No, it does feel like it was five minutes ago, but actually a lot has happened in that time, just even just through the stock profiles and the behavior of consumers afterwards. So, do you have like, if we made it a bit practical, have you got an example where like a small day-to-gap to anyone in the space has made a significant difference? I think when you start to think about particularly in the fleet finance sector, you know, anybody that is holding on to assets or any sort of exposure around residual values, you know, you're trying to set a value on a vehicle that might be 12, 24, 36, 48 months out. You're making decisions based on today's intelligence for something that you're taking, the financial exposure and risk on in 12 to 36 months away or even longer. So, I think, you know, that's probably the bit that there are gaps in that data. I think even now, you know, we've mentioned the Chinese and your entrance, we've mentioned electrification, we've mentioned, you know, localised legislation and regulations. We walk it into a potential change of following the consultation to the Zeb Monday. Whatever happens on the 23rd of October, whatever shift that has, if there is a change, that will change a lot of decisions on residual values that are being set for product coming back into the market in the next 12, 24, 36 months because everything was based on this transition to a more electrified market. So, there's things like that. I think all the estimating demand can lead to, you know, excess inventory and margin pressures, underestimating demand can lead to, you know, missails opportunities. So, I think this leads back to that last point about that demand. So, I think a 1% forecast scenario on a single vehicle, you know, may seem fairly significant in its own isolation, but you put that on an entire portfolio of product. That's a big financial decision that you're making on that product just to be that 1% out on that residual value. Some of this has to be almost impossible. I mean, correct me if I'm wrong. If you look at a government change and then you say, if you're a lender, here's your book, if you're a big dealer group, here's what the stock you've got on the ground, and that you don't know what this government change is because it might be trailed a few days out. If you're lucky, maybe it'll be, they'll sneak out, you know, leaks a month before, but you can't, you can't change your whole stock profile and you can't change your book within 7 days. Well, and that's where, you know, there's an element and it'll be get back to that point of data tells you part of the story and it's interpretation of that data. And that's where, you know, conceptualization of that data is key. And, you know, we're talking about some of the complexes today, some of the challenges that the sector's facing. Those are the challenges, you know, we go back to early parts of mind and your career, Alistair. You know, that Ford Fiesta would get a facelift and it would attract a five, found upon premium on the previous one, it would still get 43% of its cost near a three year, 60. And you knew that the next Ford Fiesta and that came out in terms of two or three years, whether it was a more extreme facelift or a minor facelift would get another slight premium and it would sit and compare us into all the models. You could predict a lot of that and it was the same cycle of continued. Those days are gone, the shifts to your point, the shredulation changes, the legislation changes, changes in consumer behavior. But that's changing the market dynamics significantly in minutes. You know, those predictions, the impact, financial impact of those predictions being incorrect, are increasing in terms of their financial exposure, prediction, predicting them is getting harder and harder as well because of the shifts in the marketplace. So I'm going to bring, I'll be remiss, but didn't bring AI into this conversation. So when you're trying to do this and if one had very good data, one could argue that AI would be excellent, but is there a risk here that the data isn't good enough and AI won't fix the bad data, it will actually just create bad assumptions for people. Completely. I think you're right. I think that's the, you know, AI, can't identify patterns, but it doesn't necessarily correct flawed assumptions. You know, poor quality, input creates poor quality output. You and I know that. You know, AI, you know, increases potential to the speed of decision making, but it also then increases the importance of the data quality because you know, you get, you know, taking the time to consider the output of that data. It's making very quick decisions. You align on it, you're making decisions based on that output, which could be wrong. You know, in AI, it doesn't necessarily and doesn't remove the blind spots. And you've had any concerns of what you've seen about how it's implemented currently. I mean, that's that's the thing. I mean, you know, it often shines, you know, may shine the light on the blind spots a little bit more, but it doesn't remove them. And, you know, that's your early point of blind spots, you know, it doesn't necessarily identify a lot. You speak to a lot of senior, really senior people across continents and across different parts of the sector. How do you think they should, they should look at the governance and the transparency, transparency, sorry, and the, and the explainability of the decisions they may make using AI? Because this is, and this would be especially true, I guess, on lending and buying. I think the, you know, again, decision should remain explainable. So that relies on the AI import and out port, you've got to be able to explain that out port. If not, it's got no context, it's got no value. - Some more about the logic that goes in. - So the logic, yeah, you know. - Accountability, you know, remains with the people. Not the algorithms, you know, at the end of the day. It can give you the information, but then the book lies with that senior leader whoever they are, whoever's put that residual value on that product, if it's wrong, that lies with them. There's no blaming AI for the cloud set. - Clouds, clouds full, our book is going down five percent. - Exactly, you know, I think we've got to be careful we don't become too reliant on the AI. The AI is there to serve as a purpose. It's there to back to that previous question, it's there to identify the blind spots. But it's again, you know, that that leader's got to make those decisions. I mean, the bottom line is, if you can't explain a decision, you probably shouldn't really automate it in the reality. - So how do organizations balance, like we've got better data, we've got AI, but we need human judgment on the input because of the, because what you're saying makes to us the logical sense. The AI can be used to analyze that data better and give you a help of making a decision. But if the inputs role, the decisions you're making on the back of it aren't actually ever good. - It couldn't be correct. - Yeah. - And that's where you need that quality control or the understanding and, you know, I think for you and I, data can tell you whatever story you want to tell you. That's the reality. Pre-AI, you can say, okay, I've got those data. What do you want? Well, what's the data telling you? What do you want it to tell you? I can give you something that supports your decision. I can give you something that challenges your decision, depending on how you interpret the data. AI isn't necessarily a different because if you put the bad commands in or you put the wrong data in, it's going to give you the wrong decision. So if you're not there to check that you're giving up the right intelligence, the right data, the right commands, but at the same time, if you don't understand what the output is, based on how do you, how do you make that, how do you make any decisions and that's the clarity around you. And you can see that with, you know, everyone has seen that when they're using AI. You could literally ask it for something simple and you get something you didn't ask for. And therefore, when I've run really complex AI builds, it is quite interesting. Even with the latest models, you still have to fix them or you still have to review and you have to work through the process for the output. Otherwise, I mean, I've done this by analyzing our, like, you know, CRM. And you get a whole bunch of numbers that meet your people and say, "But that can't be right." You look at the logic and it was making it a mistake and the data it was bringing in. So in the end, these senior people and these big companies are gonna have to take on trust at some point that people lower down, have actually given the AI the correct prompts and very clean data. Or they have controls or checks in place to ensure that it's all correct because, you know, that, you know, AI is looking at stuff that people have fed into AI and that's what it's accessing. It's not so saying that that is the correct information that's being fed. It's just information that it is accessing. - Yeah. - Okay, so I've got two more questions for you. One's really easy and one I think is a bit of a challenge. So we'll start with a challenge so we can end on an easy one. If we were having this discussion in five years' time, how would a generally modern automotive data ecosystem look? - Would a modern, how would it look? If the data was clean, all the AI prompts were excellent and people were using it properly. - I think we'd have probably greater use of real-time data as we've sort of drawn before. I think we'll remove all that historical information is relevant and shouldn't be dismissed, but it's not there to make decisions today. I think we'd see probably better integration between retail, finance, fleet, and remarketing as we've cited before. Away from that, you know, there are those areas in there that swing lanes a little bit. We'd probably see enhanced vehicle level forecasting. I think that's where you'd hope to see the system change increase use of predictive analysis. I think there's opportunities there to think about what's happening in the market. You know, we were in the month of September. You can look at September today, but actually you should be looking at what September means to the next 12, 24, 3, 6 months, 'cause that's when those September vehicles enter the use car ecosystem. And I don't think we do enough of that. So I think the future isn't necessarily about having more data. It's probably more about having a more connected data ecosystem that creates, you know, the complete picture of that fall end to end. I think we could see that, couldn't we? 'Cause I remember if you go back to 2018 when I had this tranche of UK data, people always disappointed I couldn't go back any further. They could only buy a couple of years, right? So they were very interested unless I could go back like everybody else did all these decades. COVID hits a year later. And maybe two years after that. And nobody cares pre-COVID. - Nobody cares. - Because COVID had completely thrown the rule book out, changed all the pricing. There was also some stock issues. And what I think you're basically saying there is COVID is a massive example of that. But a government changed the policy, the Chinese arriving. All those things are kind of many effects of the same thing. So you need to be more up to date with your data as opposed to going three years ago, something that looked a bit like this happened. 'Cause it doesn't impact you in the same way anymore. - You know, we're now moving into a market that we probably, and I won't say returning necessarily to a pre-COVID environment, but the COVID environment changed the landscape. - Yeah, forever. - But the extremes of the COVID environment are now gone. - It's the aftermath of that. It's the clear up of the COVID impact still going on. - Yeah, 'cause you can't ignore the 3 million vehicles. You can't ignore the market share of the established brands to the new entrance coming in. And we're not for this, but you start to think about it. Everyone talks about the new car market. Everyone looks at new car registrations. Nobody's addressing the used vehicle park and the after sales impact of all of this, because it was just still current, which you still could. I guess we'll be for it. - Quite fine. - It does because of gone. - But those vehicles from the COVID production shortage are gone. They've not come back, they're not delayed, they can't. They've gone out of the vehicle park and that age profile continues to move through time. - Yeah. - So that's a limited shift. - And then all the little micro shifts we're seeing all as together to having better real-time data than it does having that historical piece. So a nice last easy question. - If you could give one piece of advice to an automotive executive who's looking at the next five years and you could feel free to answer this from your perspective and insights and data strategy, what would you tell them to prepare for first? - I think they've really got to invest in data quality. I think there's lots of data out there. There's lots of interpretation of information and data. I think the quality of that data is going to be crucial. Back to our earlier point of poor data, poor management of data, creates poor decisions. Now I think key is this is making sure that you, wherever you go for that data to, I'll see your challenges to identify those blind spots. You know, I think that is about the quality data. So I think you know, focus on that adaptability rather than prediction of stuff that's coming. So I think, you know, the winners of the next five years won't be organizations with the most data. It involves organizations that understand their data, the best in reality. So it's data interpretation, they use the data, not how much data they have. - There you go. So if you're out there and you're listening and you look at the next five years, that was your master plan. - It's been a pleasure, Philip. Thank you so much for joining us on the blind spot. We've really enjoyed this and there is loads of great insights in there unsurprisingly from you for everyone who's watching.