Auto Care ON AIR
"Auto Care ON AIR" is a candid podcast dedicated to exploring the most relevant topics within the auto care industry. Each episode features insightful discussions with leading experts and prominent industry figures. Our content is thoughtfully divided into four distinct shows to cover four different categories of topics, ensuring collective professional growth and a comprehensive understanding of the auto care industry.
The Driver's Seat: Navigating Business and the Journey of Leadership
To understand organizations, you need to understand their operators. Join Behzad Rassuli, as he sits down for in-depth, one-on-one conversations with leaders that are shaping the future. This show is a "must listen" for how top executives navigate growth, success, and setbacks that come with the terrain of business.
Carpool Conversations: Collaborative Reflections on the Road to Success
Hosted by Jacki Lutz, this series invites a vibrant and strategic mix of guests to debate and discuss the power skills that define success today. Each episode is an entertaining, multi-voice view of a professional development topic and a platform for our members to learn about our industry's most promising professionals.
Indicators: Discussing Data that Drives Business
This show explores data relevant to the automotive aftermarket. Join Mike Chung as he engages with thought leaders in identifying data that will help you monitor and forecast industry performance. Whether global economic data, industry indicators, or new data sources, listen in as we push the envelope in identifying and shaping the metrics that matter.
Traction Control: Reacting with Precision to the Road Ahead
Every single day, events happen, technologies are introduced, and the base assumptions to our best laid plans can change. Join Stacey Miller for a show focused on recent news from the global to the local level and what it may mean for auto care industry businesses. Get hot takes on current events, stay in the know with timely discussions and hear from guests on the frontlines of these developments.
Auto Care ON AIR
Driving Change with AI in the Aftermarket
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
What if you could transform the challenges of the automotive aftermarket into opportunities for growth? Tune in as Mike Chung, sits down with the visionary, Lauren McCullough from Tromml. Lauren shares her fascinating journey from software startups to revolutionizing data analytics in the automotive sector. They dissect the transformative waves of technology and e-commerce that have reshaped distribution and supply chains, especially post-COVID-19, and how mid-sized businesses can rise above the "race to the bottom" in pricing wars. Discover Tromml's innovative use of AI to navigate the complexities of extensive SKUs and fulfillment networks while enhancing bottom-line margins.
Dive into the world of data analytics and machine learning as Mike and Lauren explore their pivotal role in optimizing supply chains and deriving powerful business insights. We discuss how data-driven decisions significantly impact inventory management and profitability, using real-world examples from platforms like Amazon. Our conversation underscores the necessity of maintaining high-quality, centralized data sources, and how tailored AI tools can offer specialized solutions for industries like automotive sales. Lauren emphasizes the importance of focusing on depth rather than breadth in software development, ensuring that businesses can effectively harness their data to drive meaningful change.
The episode takes a strategic turn as we examine the implementation of AI in business, advocating for a structured approach starting with clear business questions. We consider the potential evolution of e-commerce and its future impact on the industry, alongside practical advice for recent graduates entering the workforce. From embracing technology and nurturing curiosity to striking a healthy work-life balance, this episode is packed with insights for both industry veterans and newcomers. Whether you're navigating the intricacies of data sharing in your sector or simply want to glean advice from the forefront of auto care industry innovation, join us for a compelling discussion that promises to enlighten and inspire.
To learn more about the Auto Care Association visit autocare.org.
To learn more about our show and suggest future topics and guests, visit autocare.org/podcast
Future Trends in Automotive Aftermarket
Speaker 1Welcome to AutoCare OnAir , a candid podcast for a curious industry . I'm Mike Chung , senior Director of Market Intelligence at the AutoCare 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 that will help you monitor and forecast industry performance . This includes global economic data , industry indicators and new data sources . I'm Mike Chung , director of Market Intelligence at Auto Care Association , and I'm really thrilled to have Lauren McCulloch of Trommel Incorporated join us for today's episode . So , lauren , welcome to the show .
Speaker 2Yeah , thanks for having me , Mike .
Speaker 1And tell us a little bit about Trommel and a little bit about how you got there , please .
Speaker 2Yeah , so Trommel is a business insight solution that's really been designed for the automotive aftermarket industry , and a lot of this came to be because my background has been most of my career working in software startups and that's really where I came from is , how do you develop new technology into solve big industry problems , and one of the industries that I really got attracted to was places where other people aren't paying attention , like it wasn't interesting to me to have you know , some consumer app around better ways to invest . Right . I wanted to have this , this really tangible impact , and , from being from the Midwest and being from a blue collar background , I think that it really fit culturally with me and also the community that I saw . There was just something that was really vibrant . But working in the industry in 2019 , when I started , there was this rapid change that happened within the way people distributed byproducts , the supply chain , just as a result of COVID , and I think with COVID and the emergence of e-commerce and again all these different shifting demands around the way parts are sold , we started to see this huge shift in where money was actually being made , and what happened a lot in the e-commerce space is we saw this kind of race to the bottom , and so you'd see all these parts that were getting sold on Amazon , getting sold on eBay and all these mid-sized businesses emerging , because it was easy to just push up 100,000 listings , but then , as the race to the bottom was happening , companies were just like going bankrupt , and this was really hurting them because of , I think , the supply chain disruptions and even places like marketplaces like Amazon that are generating traffic but taking 13% of those sales , and there really wasn't much money left .
Speaker 2And when we would see companies , even those who were relatively technologically sophisticated there's just too much data for them to try to understand how to use it and so it was a little bit crippling and it was really starting to pull them back , and a lot of that , I think , is because , if you think about being an e-commerce seller , there's tons of analytics tools out there , but they're designed for CPG companies Doesn't really understand the concept of you having 100,000 SKUs and kits in this robust fulfillment network , and so we really wanted to have a solution that was a bottom line , margin focused company that could support these customers , and I don't think what I expected when we launched Trumbull , though , was the interest we would get across the supply chain . We're like oh , this is a retail tool , we'll focus on retail , maybe we'll go to other e-commerce industries . And was pretty well received as a fact that we realized that suppliers distributes everyone same problems , just slightly different manifestations .
Speaker 2And with the emergence of AI getting a lot of interest . Well , ai is really good at asking a lot of different questions , a lot of different ways , but you have to harness it for a specific use case . It has to be taught how to do what it needs to do , and so you can't expect Microsoft to understand how to be effective within the automotive aftermarket industry . You have to reshape these AI agents , and that's really what we're doing as a company .
Speaker 1Really a fascinating background and information and so much to dive into right there . So , lauren , thanks for that introduction . You said we in terms of co-founding , and Trommel , who is we ?
Speaker 2Yeah , so I formed this with my co-founder , harry Park . So he is , you know , here at the conference where we're at Fall Leadership Days and he actually spoke this morning for the category management committee . And Harry's been a great co-founder to me because he has a background in software development , did a lot of contracting for the Air Force , so he understands data security and robust projects , and we worked together at another startup and I think one of the hardest things to do when you're starting a company is find someone to do it with . That you can still get through the hard times because it's going to be hard and you're going to have disagreements . Well , where can you find someone that can appreciate those and use those challenges as opportunities to grow ? And we already had been tested in that way and so we knew that and we knew that appreciation for each other and so , yeah , he's been an amazing partner as we've been building this company together .
Speaker 1I appreciate that and you mentioned race to the bottom . So when I hear race to the bottom I think lowest price , and am I interpreting that correctly ?
Speaker 2Yeah , that's a lot what it was , because if you think about I'll use the examples we saw a lot like selling on eBay . Right , I'm selling spark plugs on eBay AC Delta spark plugs on eBay and it used to be all you had to do to sell those spark plugs is you had to get an account with a big WD , let's call it , say it's parts of the word . You're someone , you push it over to a catalog manager , you list those parts up , you sell it . Well , your listing looks exactly like everyone else's listing on eBay . Okay , well , how do you win on price ?
Speaker 2And so you go down and down and down and down until there's no margin left . But once you account for the fees that you have and the shipping that you've got and the returns that you're receiving , there was so many companies that were just literally not making money at the end of the day , there was no margin left to even have a viable business , because unless you had the buying power to get really competitive pricing like , you couldn't compete , and so looking at your bank account is not the best measurement for your company running a sustainable business . But that was kind of what was happening , especially for anyone you know , sub $10 million .
Speaker 1And what I'm hearing is in that race to the bottom it catches up to you because the cost of doing business ends up just being insurmountable and you can only run like that for so long . And I'm thinking about perhaps you can do that , have like a loss leader and perhaps make it up in other things that you sell to perhaps keep new competitors from joining . But if price is the differentiating factor and the consumer is really just looking for price , a lot of what you said makes sense .
Speaker 2Exactly , and that's the case often , I think , on these third-party marketplaces is it is competing on that and sometimes it's competing on brand if you're a supplier . But if you're a supplier , in that situation you also have more margin to play around with . But a lot of times people are just looking for for the cheapest price with the fulfillment times that they need , and so there's just not , there's not a lot of wiggle room , and what happens is if you're running on these razor thin margins , you have no room for mistakes , and mistakes will happen . One of our earliest customers they had a pricing error and overnight lost 60 grand . What do they do ? Because if it was an Amazon , so if they were to have canceled all those orders , they would have lost their Amazon account . So they just eat that whole loss . Well , that ate up all their profits for maybe the next , you know , two , three weeks .
Speaker 2So how do you balance that ? And that's where I think , as an industry , we needed to come together and have this kind of awakening moment to say like , okay , we can't keep doing business this way . And I think I've seen , really , the last three years , things start to shift and that , being conscious of exactly what you said , is . It's okay to have a loss leader if you know it's a loss leader . And it's okay to have a loss leader if you know that that can generate future traffic , that people add on products to it , that they're gonna come back again because you own the customer's data . But if you're not in an environment where that's possible , where that's the culture of that purchasing behavior , it's not gonna work that way . You're now just selling something at a two-point margin and that's it .
Speaker 1And you mentioned , mistakes happen , and not just from the company , but from a consumer as well . You mentioned returns , and if I'm buying that AC Delco spark plug , perhaps I ordered the wrong one and I need to return it . And there's that friction element for additional costs .
Speaker 2Absolutely , and it's . And even you know , sometimes these returns happen because people have bad Fitment data right . Your Fitment data right , your Fitment data is incorrect . Sometimes I see it happen on the accessory side where , you know , we called it blue , someone else thought it was black and we had a disagreement . They sent it back . But the platforms and even most e-commerce policy sites , like it's pretty liberal returns , like you can return something for any reason and most of the time if you've got a lot of volume , you don't have time to fight with someone about oh well , you returned this because you decided you couldn't install it yourself . Like that happens all the time . Like we've had people we've worked with who've gotten things back with gas in them . You know and you just can't . You can't change the fact that the consumer is going to behave that way , and so you have to instead figure out how can you manage that risk , both through your margins and then also just through early indicating factors that might indicate that there's a potential red flag here .
Speaker 1And going back to some of the introductory , as you were looking at a business solution , you landed on aftermarket and the things we just talked about the cost of doing business , the race to the bottom , errors are made , the ability for a company to absorb these . Compare aftermarket to other industries .
Speaker 2for me , yeah , I mean , I think it functions in a similar way to what I kind of call the parts and supplies world . I think that anytime you're talking about something that fits something , there's some uh , there's some overlap . The fulfillment networks are similar . Um , you're , you know . Again , you're buying something that fits another product . But the aftermarket industry , right , we have these distinctive tools that we've developed , and there's , those tools have been developed for some of the reasons that I said . Right , Even if you think about a PIM , most PIMs would not have a fitment or application level .
Speaker 1PIM stands for what ?
Speaker 2again , product Information Management System . So like , if I'm listing products , I need to have Fitment data tied to it . That's not something that most industries are going to do , because even other people and , like we talked to , sometimes companies in ag they don't have ACES standards .
Speaker 2And so you end up on these kind of really specific tools that don't integrate with other tools , and so it really creates this data silo , and I think part of it is to just recognizing , like , what are the business challenges that we have , and I think every industry has their own . Yeah , they have their own challenge . I guess for lack of a better word their own specific nuances . But given the size of this industry and the scope of this industry and the complexity of the fact that we're you , we've got 5 million parts running through the supply chain on a regular basis you need some specific trainings and tools for that . It's just the same reason that we try to keep people in the aftermarket industry , because we know the importance of institutional knowledge , and that matters with people , but that matters with technology too .
Speaker 1Sure , and so perhaps a layperson example might be I'm ordering shirts from a vendor and there might be small , medium , large . Maybe it gets a little bit more complicated with athletic fit , slim fit , extra slim fit , and then perhaps you go to shoes . There's US , uk , european sizings and then we come to aftermarket and it gets exponentially more challenging .
Speaker 2Exactly , exactly , exponentially more challenging . Exactly exactly . But what we have in our benefit is we do have standards , because the fact that we do have some some data governance allows a lot more to be possible where without those the standards and that that's for product data , but also in just the way that we exchange information it does , it does allow us to better leverage technology . I think then you know , maybe someone who's I don't know some other manufacturing type of industries where they just don't have that level of community . That's created governance and it continues to hold them back .
Speaker 1So , thinking about how Trommel fits into this ecosystem and the data that you're gathering , can you tell me a little bit about how Trommel helps to who ? Are you helping the third party seller ? Can you just tell me a little bit more about that ?
Speaker 2Yeah , so Trommel started off . I think I might have mentioned working with the e-commerce retailer is really where we started . So anyone who was selling typically into the DIY market you know technicians will still come sometimes buy these platforms but that was mostly who we supported because we saw those were the people who one we knew them from our past experience . But we also saw those were the people who were really really managing on thin margins . Right , there was just no play left .
Speaker 2But what we're now starting to get into is just how to evaluate the sale of products in general , because for us , what we want to be able to say is how much money are you making on every order ? Where is there a discrepancy or an abnormality ? How is the velocity of the products that you're selling changing ? How does that mesh up with your inventory volume ? And so it's kind of for us it's starting at that retail and the e commerce
Leveraging Data for Business Insights
Speaker 2end . And what was beneficial about that ? It gives us a rich data set , like you know , if you use Amazon example , if you're selling on Amazon .
Speaker 2A rich data set . Like you know , if you I'll use Amazon example , if you're selling on Amazon . You have a lot of data . Can you do anything with that data . Right , that's where we have to kind of get into that work .
Speaker 2But now , as you go back into the distribution network and going to suppliers , particularly those , you maybe decide I want to sell on Amazon through a 1p relationship , which is basically where Amazon becomes their distributor . They're still the same questions that remain , and so what we want to be able to do is understand how those nuances change across the supply chain , but also recognizing that by supporting each part of that supply chain using machine learning , our system can become more and more intelligent , where , again , if we were just building this as a generic analytics tool , we'd never be able to train those AI agents . In the same way , we'd never be able to provide insights . That said , hey , we noticed that . You know , we pulled in some market intelligence data and , based on the VIO information we have , you are going to run out of inventory in Florida for everything that fits a Jeep Wrangler . You know , we want to be able to have better insights into that data and it's , I think , the future of software companies is depth , not breadth .
Speaker 1And so I think that really going deep into companies is how you start to allow people to flip the right levers at the right time . So , just as a hypothetical example , mike Chung Industries produces a variety of parts and I sell them online whether through Amazon , ebay , my own website or perhaps other e-commerce vendors and a company like yours will be able to take my sales and consumer behavior data , analyze it and sell it and presumably let me know which channels are working well for me , how my consumers or potential consumers are behaving , if I have the ability . You mentioned the Jeep Wrangler in Florida where I might be in need of more inventory to meet customer demand for a particular part , a particular geography .
Speaker 2Am I getting that ? Yeah , exactly . Or did the promotion that I used on one of my channels work ? Like , did that actually make me more money when I ran this promotion ? When I changed shipping providers ? Did that work ? Or , you know , mike Chung Industries is selling on their website , but they're also selling to Lauren Industries and Lauren Industries is reselling their products with five other brands . What is that relationship Like ? Why did Lauren stop buying SKU AB1234 six months ago ? You know , why did they stop buying yesterday ? Like , what are these ? Because there's so many different assets that you have to leverage just by looking within , within your own data . You have to know how to harness those and how to deploy them at the right time .
Speaker 1So tell me a little bit about that process how to figure out what data to capture , how to mine that for insights . Can you tell me a little bit about that ?
Speaker 2Yeah . So you know , a lot of this starts with getting the data into a central source , right ? The hardest part about analytics often for folks is just getting the data into a normalized format . And so there's this statistic that 80% of an analyst's time is spent collecting and preparing data and only 20% is spent in the analysis itself . And if you think about an analysis situation , an analytic situation that needs to happen in this industry , you're getting hit with like oh we have a board meeting this week , we have a conference this week , I need to do updates in pricing . You've got all these different queries you're running . You , I need to update some pricing . You've got all these different queries you're running . You don't have time to do any deep thinking , you don't have time to have that depth . And so what we want to be able to do is make that process much easier by getting that data already , like basically eliminating the friction points of that collection , normalization of the data and giving people a security warehouse that's complete with the data that they need . And then , when we talk about having that granular data , that's where that's the only way you can extract some of these insights and that's where machine learning models and AI models can be really helpful because they can find clusters in the data .
Speaker 2Could a human have done that ? Of course they could have if they had enough time , because really all AI is doing right now is replicating human behavior , so we have to teach it . But if we say I want to find you know what is the patterns in promotions that have increased profitability versus those who haven't , what have been the skews that have been the what's a common factor of skews that have increased in velocity ? What's been the impact of this pricing change ? And then being able to take that , and if you give someone an insight , the real key of this and this is what we're working towards to become a really intelligent tool is okay .
Speaker 2We real key of this and this is what we're working towards to become a really intelligent tool is okay . We told you something . Now we saw you implemented it , we measured the outcome of that , we fed it back into the model and now it becomes smarter over time and it continues to learn . In this , machine learning models can really really become sophisticated because they develop institutional knowledge within your own company , but they also develop institutional knowledge of what works for the industry , and that's all we're doing . I mean , it's nothing that analytics and statistics haven't been doing . It's just leveraging the best technology available to do more faster , better and more impactfully .
Speaker 1That makes a lot of sense . So just to recap a little bit of that . It's really critical from what I'm hearing is to come up with a data structure and cataloging system so that it's reliable , it's repeatable and hence searchable , queryable , mineable for those questions that could evolve over time .
Speaker 2Exactly . And yeah , and we're still at the point of our company where we keep human in the loop . We call human in the loop with anything that we do . That's AI-driven , because you need to keep quality control , and quality control is on the data itself , but there's also quality control on the output , because you have to make sure , if we're going to generate these insights to make a big decision for our company , is this data correct , and so that does become a big attribute of it , and we always say don't let AI do your math .
Speaker 2It's very bad at doing math , and so the calculations we run have nothing to do with AI . The way we automate data has nothing to do with AI . What it does is it continues to ask questions and it can generate that data in a format , especially through large language models like ChatGPT , if you host those in a private database . Well , now I can take the data that's already in my data warehouse , fill in the blanks for the board report that I need to generate , and it's done Versus something that might have taken an analyst a week to do , and that week they're not going deep on the data of your company , and so that's the thing is .
Speaker 1I think people get a little threatened by oh , it's coming for my job , but what it's really doing is giving you time back to do more hard , creative , human tasks , because the analyst might be searching around for that week and if the data is messy and not clean , it makes it . So let me ask you this then In terms of you've got the company launched , you've got your data set launched and you have your structure and everything there , but now you realize , oh well , you mentioned something about promotion . So let's just say , now I want to do promotions analysis and I need to add more fields to my data . Tell me a little bit about that process what works well , what doesn't work well ? Because I think in the beginning stages of a project , starting a company , you have so much time to define your sort of layout and it's kind of like the 80-20 rule , right , because you want to get it done , get it out to market and you can't necessarily think of everything or anticipate everything . But tell me a little bit about when you need to add more .
Speaker 2Yeah , I mean , for the way we treat it right is we're collecting everything that we have available in a data lake and so this data is already there . It's just a matter of what do we pull into our data warehouse and what gets analyzed . And so , for example , if we were saying , oh , we weren't pulling in promotions or we weren't pulling in credit card transaction fees or whatever it's going to be , we would just add that as another field within our database . And that's the same
Optimizing Data for Business Insights
Speaker 2type of data structure . And this is something , as a young company , that happens all the time . Our clients will be like , hey , I'm really really curious about X , y , z . Do you have this data available ? Oh , we don't have it Right . So now we'll go back and structure it .
Speaker 2And I mean we , we still at this point we don't put the , we don't put the onus on our clients to do this for them . And we think that's really important because it also gives us that consultative time to say , well , what are we actually trying to find out here ? Because , depending on what they're trying to measure for the impact of that promotion , the profitability of the promotion we may decide to structure this differently , but we can have that kind of consultative feel to it to make sure that we're actually delivering value , because it's not about necessarily just having the data . It's about what things will move the needle and and this will never not be a learning process for us we're always going to continue to be adding new fields , adding new ways to slice and dice the data and trying to do our best to be mindful of you know , diminishing returns , like you said of that right .
Speaker 1So I appreciate that , and maybe another example is this that'll lead into another topic is your client says to you tell me what our competitors were doing and that could be a bolt-on data append to your data lake . And then it's a matter of sorting through importing that variable , adding it to your structure , of sorting through importing that variable , adding it to your structure . But to your point , well , what are you really trying to learn ? And there might be several follow-on questions , and then it becomes that consultative . Okay , what are we trying to find ? Now let's go find the data to do it , whether or not we have it , and let's say we need the purchase data you mentioned , credit card data , competitor data . Tell me what goes into your and your team's mind when you're evaluating other data sources . What are you looking for ?
Speaker 2Yeah , so one . We don't want to put it in if it's not of manageable quality , right , it's that garbage in , garbage out problem , and so that's always going to become some level of a qualifier for anything we do is can we get this data into a usable shape before we actually put in our data warehouse ? Because the moment we kind of contaminate it with bad data , you start to structure the integrity of everything you're doing . But what we also recognize is that and I think this is kind of what you're alluding to , mike is that we don't need to be the creators of the data . So if we think about a competitive analysis , for example , we don't necessarily need to be the one who's doing that competitive analysis If there's another service company that's doing it . What's really helpful is that I can take that data and put it into context of your own data , because I think a lot if you do it correctly , if you're mindful about it . Data is very much a one plus one equals three , because if you look , I could have the best competitor data , I get the best demand forecasting data and I could have the best business intelligence data . But if none of those data sources talk to each other , then I don't actually maximize the utility of all .
Speaker 2There's a company we've been starting to partner with called content status that I know they've been involved in auto care and they do auditing of of online listings so they can tell you what your and they do auditing of online listings so they can tell you what your . You know , is this an A , b , c , d listing ? Whatever ? We don't want to scrape that data . We don't want to read that data . What we want to do with them is figure out how do I match that up with your own sales data ? Hey , we noticed that you are losing velocity on the SKUs that are performing at F .
Speaker 2This is probably a really good place to put some effort behind versus , like you know what other things that we can read and I think that's the , that's the big you know I mentioned earlier is , I think software companies need to focus on depth , not breadth , and that's exactly how we can do that is , we can say it's specialized in this thing , but let's figure out how to wire it together , cause , at the end of the day , if we're not providing value to the supplier , the distributor , the retailer , then , like , what are we doing here as companies ? We're just sitting here fighting for budget items . No , we can maximize value . We just have to figure out how to collaborate . And it's the same thing if someone's working with that company doing the competitive analysis and you know our companies , it's their data , you want to send a feed to them we can have that conversation .
Speaker 2But you just have to figure out how do we be practical ? And we really want to be the best . We want to be the best not just in the industry , but like one of the best in the globe about figuring out how these insights need to be delivered . And so , for me , I'm really excited about , you know . We start to get into the kind of behavioral , economic side of it and like how I can say this a hundred different ways , but when do I say it , how do I say it and in what structure do I say it . And I think that mastering that human side of analytics is a really important aspect of it that is going to become almost a requirement for anyone who wants to have their technology adopted by a product organization .
Speaker 3Hi , I'm Jonathan Larson , Vice President of Digital Products and Standards of the Auto Care Association . Are you getting the right part to the right place at the right time ? If you're not utilizing the AutoCare Association's ACES and PIEs data standards , you're not only wasting money , but you're probably duplicating your efforts . Schedule a consultation at autocareorg forward slash standards to learn how to lower your supply chain costs , increase your speed to market and reduce your returns with ACES and PIEs . We look forward to hearing from you .
Data Depth Versus Breadth
Speaker 1You mentioned depth versus breadth , so depth is to me . Can you give me some examples of deeper questions you've had with clients in regard to data and how you're , you and your team are able to sort of dig deeper and engineer a solution ?
Speaker 2Yeah , I mean it was a little surprising to me at how generic a lot of other analytics tools that they were using were . You know , on the e-commerce side there's a lot of these like oh , I signed up for a hundred bucks a month or something and got a dashboard . That dashboard doesn't tell them anything except for maybe aggregate a couple of sales channels together , and it's that . The tracking of changes is what's been a big part of the , I would say , depth that we've gone into and I mentioned earlier I think I was picking on Microsoft or something they don't have a stake in training tools to ask the specific challenges of an automotive seller . Right , they're not going to necessarily evaluate the link between sales velocity and fitment data and part failure rates and how predictive maintenance can actually inform where we put what products . But that's something that when you specialize that , you can do , and I think that you know . For us right now , the big focus has been on those changes in margins , those changes in sales . You can kind of those velocity movements . But then also being able to come in and say this is what we're starting to get into now is say again we saw you did this thing , we tracked it as an event . We saw you updated your Fitment files on November 1st 2024 . Well , what's happened since then ? Okay , well , we ran a regression and we now found that every time you do X , y happens . We think that might be a good recommendation we can provide in the future , and it's having that track everything mentality Track everything and only report on the most impactful things and making sure that people get that data faster .
Speaker 2Because I was just talking to someone who was saying like oh , I'm up late at night trying to figure out all this data and trying to sort it together and I want to be data-driven , but it takes a lot of time , and , yeah , it doesn't . It shouldn't want to be data-driven , but it takes a lot of time , and , yeah , it doesn't . It shouldn't have to be right , we have the tools and the resources to get this in front of you , but we have to figure out again how does that delivery look ? And again , just as much of this is a human problem as a technology problem , and I think that sometimes software companies can forget that that at the end of the day , like you might be selling technology , but you're really just selling better lives . You're selling a better workday , better bottom line and if you can remind yourself of that as much as possible , it can really impact those user experience metrics .
Speaker 1So going to the breadth question then , like , give me some examples of these are kind of getting away from the main thrust , or these could be distractions or perhaps not answerable by the data that we have . Can you give me some examples that you've come across ? So you're saying like , what , like specific tools that are going too wide I think when thinking about the depth versus breath , right , what questions could be considered ? I guess going into the breath direction in the wrong way .
Speaker 2Yeah , I think that I think I'm answering this
Strategic Implementation of AI in Business
Speaker 2correctly .
Speaker 2So I mean , one of the things I'll say , like from people's attention spaces , right , a breadth question is , like we want to have an AI strategy at my company .
Speaker 2Like I'll kind of , I'll use that so , and maybe you even go one step further and say , okay , we want to learn how to incorporate large language models at our company and people create entire initiatives around just how are we going to leverage open ai ? But that's not really the question . I think , um , I think that what happens is you just end up going like two inches deep on everything , um , because you're just kind of like playing this like whack-a-mole , and meanwhile , potentially , your competitors or the rest of the world is getting focused . And I think that you know , if you just use large language models again as the example to say , like a breadth problem would just be saying , oh , we're going to , you know , incorporate open AI into our entire company and everything we do , instead of saying , oh , wow , this is a really good way to , you know , generate board reports ? Right , but what does it take to generate a board report ?
Speaker 2Well , it means you have to have an AI agent trained on what makes a good board report for you know my tongue industries Right and you have to make sure that you can give it feedback so next time it gets smarter and smarter over time .
Speaker 1That's part of that machine language and adaptation .
Speaker 2Yeah , yeah , it adapts . I mean it's just . It's just like a person , right you ? But it's a person without consequences , like my co-founder , harry , was joking earlier at a presentation that you know it's we call it . We call it the intern and not to . You know , talk down to interns . But it really just does what you tell it to do and it doesn't care about getting fired . It just like you told him to do the thing and it just knows you want results , so it just sends you the result and sort of hallucinate . So it's something called AI hallucinations , where it'll just make up things because the last thing it wants to say is I don't know , it just gives you data and so if you don't know those specific ways , you can deploy it . More dangerous than what I just mentioned is like sporadic deployment is relying on a tool that hasn't been trained for the function of what you need , and now you're literally using bad information to make data driven decisions .
Speaker 1I mean , you mentioned garbage in , garbage out , and now this is just garbage .
Speaker 2Yeah , and you can't know because you , you , you trust it , and so I think it's really important that you know . If you're going to use , you know , a software product , for example , that doesn't focus on the automotive industry , well , what does that mean ? What does that mean ? It could be getting wrong , and sometimes the answer might be nothing . Sometimes it might not matter at all , because what it actually matters is the fact that you're a remanufacturer and this is really good at remanufacturing , cool , okay . Well , this is a situation where , maybe for pricing , for distributors , where it really needs to understand the nuances of the industry , and maybe it doesn't . And what is that risk for us ? And is it a bad data problem ? Is it a depth problem ? It's just , you have to really be mindful of these tools , because we trust software really fast . I've noticed there's kind of this . I think it's happened with every kind of new wave of innovation .
Speaker 2You see this kind of like trust really really take a while to kick up and then it just really trusted Right . And I think that what's happening is if , during that part where you're starting to earn trust in things like AI tools , there's mistakes , now all of a sudden , you're just again flying blind , making decisions on bad data . So you just there has to be this consciousness of of yeah , what is my goal here ?
Speaker 1Yeah . So I think , to recap a little bit of that Be focused , yeah . See where you can apply an AI solution directly and not just necessarily . Oh , we're going to use AI period , but think of a process . Think of a process in your company that could benefit from it Exactly .
Speaker 2I think about it as kind of starting with a business question . I'm working with some folks over on the BTC for MIMA and we're doing a paper on AI for the industry and the business technology council for them , yeah , and so one of the ways we structure the paper on AI is to , you know , here's some background , but let's start with business questions . And so we pick two specific business questions that someone might have that gets brought up at a board meeting . Okay , how could AI be deployed here ? And each one of them has like 15 different ways , right , some you can start tomorrow , some might take , you know , two years to implement , but it's knowing what those opportunities are out there .
Speaker 2I think that we have , and we also have to help each other to understand where , if we want to not get caught on our heels and make sure that we are , you know , seen as one of the most innovative business industries , which I think we can be that's a big goal for me is to not just promote trauma but to promote innovation within the industry , and for us to be able to do that means that I want to , like 15 years , 20 years from now , I want to see automotive aftermarket as one of the most forward-thinking industries for business technology . And how do we do that ? Well , it's because we were willing to knowledge share on what's working and what's not with each other 10 to 15 years .
Speaker 1I was thinking what can we expect in five years , 10 years , 15 years ? What do you see in the roadmap ahead ?
Speaker 2Oh man , there's a lot , I think , that commerce . So we kind of look at e-commerce right now and I think it usually gets this label of this is the DIY market . This is the people who are , you know , either wanting to use discretionary income to upfit their car or to maybe fix it at home to save money , but the B2B e-commerce , which kind of just means that commerce , all commerce , is e-commerce . I think we're starting to get to that day , like I think the phone call orders are going to be starting to come to its end pretty quickly and we're going to see this mass digitization and then , you know , that's a mass amount of data , and I think that what we're going to have to be really mindful of is how do we act really fast ?
Speaker 2Because I know we've seen a lot of folks that we work with that are getting pressed on price now from overseas competitors . So if you're selling parts and there's a Chinese company that comes in , who knows getting their costs subsidized , I don't know what's happening , but you're now trying to compete , and so it's like how do we leverage what we have ? Are we doubling down on brand and trust and authority ? Are we just figuring out ways to save margin here that we didn't have before , and I think that we're just going to have to be really nimble and fast , because EVs are kind of a question mark . We don't know what's happening there . The life of a mechanic is going to start changing very fast to being more of an engineer in a technology-focused industry , so we have to , on the part side , be able to adapt to whatever changes end up happening .
Speaker 1Some of the questions that perhaps could be answered through AI technologies and we talked a little bit about , I guess , efficiency , profitability of a company . Could it perhaps be in the future where , with the data that you're collecting , the algorithms , the large language models , the machine learning is going to be able to say , to do an analysis of you're losing money because of bad fitment data that results in returns on these SKUs , and here are solutions to improve it 100% , and I think that's where this willingness of you know .
Speaker 2when we work with our clients right now , their data is like isolated right , like it is in their own tenant . But I think the next step to this is okay . Well , what are we willing to share , like ? What can I share ? To give something back , and people don't like to be told what to do Again .
Speaker 2This is the human element of being a service provider . But , like you might be able to say hey , we noticed that when your Fitment files are more than one year old , your return rate goes up by 20% . And then you , what we could say then as response is we've noticed that suppliers like you typically find this rate goes down significantly when updating every three months . Anything more has diminishing returns to it , and that's how you can deliver that right . You don't have to say you should update it every three months . You can say , based on people like you , this is where you know weekly doesn't matter , don't make yourself . Do it weekly Every three months is good , or whatever that metric is going to be for that category . And I think that's where , again , these are all things a human could theoretically answer , but we don't have enough resources to hire the people to do that much analysis on a regular basis . It's just too much .
Speaker 1And you mentioned regression earlier and it could be and you gave some great examples and it could be that the product weighs this much , or it's in a box this big , or it's in this driveline component , or X , y , z . That's where the power of the big data and the algorithms and the analysis to do that number crunching quickly , versus having a team of analysts who may spend many more person hours , for example .
Speaker 2Exactly . We're just getting them , and sometimes , if you're focused on it like what they actually need for the analyst is just to point where to look at you can just sometimes say something's happening here
Navigating Data Sharing in Industry
Speaker 2. Because what you realize with a lot of folks whether they're an operator , an analyst or an executive is that when you point something out with some details , they know what's going on . These are very smart people and so you don't always need to be prescriptive . Sometimes they can figure that out with the data sets . But what you need to do is say , hey , there's something going on here with your return rates on this category of products , and it looks like it started on April 1st and it's been going up ever since .
Speaker 2So now , instead of them spending all their time , why are our margins down this year on breaks ? Well , we've isolated it to this much smaller issue and now they can be human . And now they may need to pull in a couple other data sets . Right , that they're only needed for this specific situation . But look at you like you're doing what you're trained to do . You're using that brainpower you're having . Well , let us carry the load of the easier stuff .
Speaker 1Right and it's a feedback loop between what executives , managers , line managers , sales . They might notice some patterns and then harness the power of the AI or whatever analytical solution .
Speaker 2Exactly , exactly and again that loop . Keep learning , because data-driven is not a linear thing , it's not a single direction , it's a loop , like to be . Data-driven doesn't just mean you make a decision based off data . It means you made a decision based off data , measured the impact and took that into account again . Right .
Speaker 1And earlier you said something about how much am I willing to share , and I'm interpreting that in two ways . One could be I'm Mike Chung Industries and I'm looking at Mike Chung industry data , but you could also being at Auto Care Association thinking about things like demand index across other companies . So I guess as a basically a data consultant right , I don't know that you would necessarily have liberty to pool your data across different clients . I know that's usually there are usually firewalls along that , but that could be a very interesting avenue to see what patterns are emerging across an entire industry .
Speaker 2Yeah , and we've seen this in other industries , so we've paid attention to you know how this works across , like grocery , if a grocery has done some of some of this and , um , you know there's a little bit of usually this negotiating like I'll give something to get something back , sure , um , but like , how do you anonymize it ? Right , is it ? Okay , we can give margin date on category if we strip off brands and we've all agreed to do that , okay , that means we can give that benchmark back and and that's something that we're we're not I don't know what the threshold is going to be , and it also it's not a blanket threshold is , you know , there might be a distributor who's willing to share a lot of data with their buying group , but they're not willing to share it with another distributor . So they're saying , okay , you can keep it , but you have to anonymize it before you share it out , and those all just have to be data exchange agreements that we have , but I think we're going to have to .
Speaker 2My personal opinion on that well , my opinion is that we , you know , work with our clients to to make sure we because we recognize the sensitivity of this data . Of course , however , I do think that we do have to have this conversation around like what are you really giving up compared to the gains you're getting back , and can we push ourselves just a little bit more because we're so confident in the like ? Are we competing with each other as much as we're competing with bad fill rates , with waste ? Sometimes not , and that's kind of my . Something I'll always say is that I think the industry's biggest competitor is just waste , and so it's like if we think about ourselves as competing against waste , well we know we have a slightly different lens to look at this through .
Speaker 1Right and , of course , coming from Auto Care Association , I think about what Daniel Zenko does with demand index and what a trade association could do and , depending on the context , I guess the world is our oyster right .
Speaker 2Exactly and I think we don't have answers . And the reason I can't give distinctive answers for how I think this is going to play out is because we've never really had the opportunity before distinctive answers for how I think this is going to play out is because we've never really had the opportunity before Like it's really just been recently that we've been able to have enough data worth aggregating .
Speaker 1I guess you can say .
Speaker 2And so it's like we're pretty much on a new frontier right now , and so I don't think that we've even as an industry , figured out what we're willing to do and not , and what's actually proprietary and what's not , and so , and again , where can we say I'll show movement but not sales numbers ? Okay , cool , because once we know what can be released , we can function within that . But it's going to be , I think , yeah , a lot of conversations , a lot of conversations with our clients , especially as we get more and more people who are selling the same things , competing with each other . But again , just recognizing , like , what is the greater goal and is there ways that we can have a little bit of flexibility to move the industry forward ?
Speaker 1Great , and then I think you know we're going to try and wrap this up in the next five minutes and thinking about that uncharted territory with all the data that is now available and the new technologies and analytical techniques
Embracing Technology and Entrepreneurial Spirit
Speaker 1for it . What advice would you give to somebody , say , that's coming out of college , maybe entering a retail , a manufacturing , a sales-oriented position ? What are some of the things that you might advise him or her ?
Speaker 2Yeah , I think that if you're hungry for technology and and this can include things like ai , because it's , you know , it's the the buzz right now , or it can include any type of other you know , data warehousing systems , like whatever you can be if you can be the expert which the expert now is actually a pretty low bar , right , like you can know more .
Speaker 2Someone was I care exactly . The quote was like . Basically it was saying that like you can learn , if you're a student and you're studying and you start using AI , like large language models , if you use it for three months , you're going to learn more probably than anyone who was using these models for the two years before because they progressed so fast , and so you have so much opportunity to become an expert on the way you , within your organization , you know , within your job , function , can sales room , they go back to the phone , whatever it's going to be you can start to understand how could I create efficiencies and process with technology , sometimes with just like better policies too . Like I said , not all these are technology problems , because I think that it's there's there's so much opportunity to be a thought leader here If you're just willing to kind of have that curiosity for a curious industry right .
Speaker 1Right , and I was just going to say be curious , yeah , and I was having this conversation with somebody in terms of our reflexes , how much can you train those fast , twitch reflexes ? But curiosity could arguably be practiced and improved upon 100% .
Speaker 2Yeah , and getting yourself that hunger and right . It's the same thing . It's that feedback loop that you have too , and I think , as managers with young hires , you know , if we think about the way we talked about the machine learning model , it's the same thing . If you have a young hire who is , you know out , hungry and asking questions , don't put them down , don't say like you don't understand , like it's not done this way , right that's always like the whole .
Speaker 2Like you know , it shuts people down just like oh no , yeah , you don't have enough background , like it can't be done this way , and I think that the more you can encourage these , people would even say like that's a really great idea . It's probably not something feasible right now , but keep learning , keep asking hard questions right and keep them to be hungry , because the longer you are in an industry , the more bias you become , whether you like it or not , and so we need these that's the way we've always done it .
Speaker 1It's always been this way .
Speaker 2Exactly so . You need these kind of these fresh faces . And there's a guy that lives in North Carolina with us and he named John Walport and he wrote a book called the Two Butt Rule and it's a . I was reading through it and it's a pretty good one , because it's like saying , well , we can't do this , but and then ? But what if we could ? And I think that developing that , that mindset can be really useful because it challenges that maybe the status quo doesn't have to be the status quo anymore , and I don't think without new voices , new brains , new faces , and you can have that happen as quickly .
Speaker 1Good advice , no matter where we are in our tenure , it sounds like . So my last question is this , and it harkens back to something you opened with you mentioned blue collar Midwest . Tell me more .
Speaker 2Yeah , so I grew up in Southeast Ohio , central Southeast Ohio . My dad was a refinished furniture for a living . My grandpa ran a small HVAC company . So I feel like I just kind of was born in that entrepreneurial spirit and it's interesting , kind of working Like I feel right that I found myself in this mix between a more industrial industry with technology .
Speaker 2It just kind of culturally fit with me because I got out of college I fell in love with software and what it can do to make impacts . But I I'm very driven by the human side of building technology and like I I say sometimes , like you know , our goal as a company is just as much as anything to make sure that that manager doesn't miss his kid's soccer game because he's trying to figure out what went wrong . You know , in the fulfillment data , like like if we can help people not have to miss their kid's soccer game and take a vacation , not feel like they have to keep their phone on them all the time , like that's a win . And I think that's perspective that you get from being a child of entrepreneurs . And my co-founder , harry too , he was , his parents were entrepreneurs too and he's from Alabama and so I think we both kind of have that spirit of we understand . We love technology and we understand it , but we love deploying in places that have a real human outcome .
Speaker 1Thank you for sharing that , and my last question for you is this you're hosting a weekend party , you're having people come over . What type of food might you serve ?
Speaker 2Oh man , what type of food do I serve ? It depends on how much time I have . I feel like barbecue . I'm in North Carolina , so you got to have people for it . It's either barbecue or like a taco bar , just because a taco bar is like one of the most versatile things that you can do for a large group . You're just like make a bunch of things , put them out , let's go .
Speaker 1And it's interactive .
Speaker 2And it's interactive . Yeah , exactly .
Closing Remarks on Auto Care
Speaker 1Excellent . Well , on that note , we're going to close up this session of Indicators . My thanks to you , lauren , for a really engaging , informative discussion , and my thanks for our listeners . We hope you found this informative . We look forward to your feedback and , until the next time , have a great time . 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 . Autocare OnAir 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 .
