Will: Hello Scott. Welcome to the show.
Scott: Thank you. Good to be here.
Will: So, it occurs to me that an independent trucker is not that much different in some ways from independent consultants out there, because when they’re trying to fill their trucks, they go to A to B full; they want to come back from B to A full. We want to fill up our days. So, there’s some brokers in the middle there for truckers who are helping them fill those loads.
Scott: Yeah, that’s exactly right. I like to say in consulting, when I was doing consulting, that it’s basically only two problems, either too little work or too much.
Will: That’s right.
Scott: That’s true for the carriers, especially the owner operators as well or any small carrier. When you have a large scale, just as is the case with a consultant workforce, you’ve got a little bit more breathing room. But when you’re dealing with a small scale, again, an owner-operator, just one owner and driver of a truck or just a small fleet, maybe 20 or less power units. Then keeping those trucks moving and full of freight is the lifeblood of your company.
Scott: So, one of the major values that we offer is that we get those trucks full of freight. We have relationships with 14,000 clients, and we have relationships with about 40,000 truckload carriers, and we play matchmaker between those two sets of needs, the people who need to move freight and people who need freight to move.
Will: Awesome. Well, let’s start at the top. Let’s start at the beginning here. So, Scott, give me just an overview of the trucking industry, the types of full truckload, less than full truckload. Walk me through how the trucking industry works today.
Scott: I’m going to bring it through the lens of Echo’s business, because I actually think it’s a helpful lens just to understand the overall industry landscape. So, let’s start with what’s called less than truckload. Less than truckload is… First of all, we’re not talking about parcels. In this industry, the UPS or FedEx delivery person that’s showing up and dropping off packages at your house would be referred to as parcel business.
Scott: When we talk about freight, we’re talking about pallets or skids, generally loaded with forklifts or hand jacks and things like that. So, we’re dealing with pallets of goods, and less than truckload is anywhere from typically from about one to about eight pallets, and a standard 53-foot dry van trailer can probably fit about 26 pallets, depends on whether or not they’re stackable, depends on the nature of the freight, lots of other details.
Scott: But simply put, a less than truckload or LTL shipment is you’re shipping, say, between one and eight pallets would be typical. That industry operates a lot like airlines. It’s a capital-intensive business. They own a lot of trucks, and they own a lot of terminals. So, one of the ways it’s like the airlines is that you have a hub-and-spoke system.
Scott: If you were going to ship something from New York City at Queens, New York, and somewhere in that region to where I might have a warehouse in the western part of Chicago. Your shipment would get picked up. It would get delivered to a cross-docking facility and mixed together with other pallets of freight that are making the long haul to Chicago, just the way airline hubs work, and they gather people together. The freight gets gathered together.
Scott: Then that gets put on the long haul shipment to Chicago and broken down again into local deliveries that are sent out and back in the same day in that region. So that’s happening all over all the time with LTL shipments, and there’s about 130 of those carriers. The pricing is complex, but it’s known. It’s set. They’re actually referred to as tariffs, which hearkens back to before deregulation took place.
Scott: The pricing will depend on where it’s shipping from and to and how much it weighs and what kind of freight it is, but ultimately, it’s known pricing. So, for us, as a reseller of that capacity, we operate a lot like a travel agency, which is to say we’re going to put a markup on it, but part of our value is offering better pricing than a person can get on their own because we buy in bulk selection.
Scott: So we have contracts with all of those 130 carriers, some of which are national, some of which are regional, and service because one of the things that is a truism of this business is that things go wrong, and someone has to deal with it. So, we help manage that. That’s the LTL side of the world.
Will: There’s only 130 carriers. So, I guess independent trucks don’t really play in this space, I would guess.
Scott: That’s correct, yeah. It’s a very different model. It’s just like there’s only so many airlines, and the same thing just as airlines buy each other. I’ve only been in the industry now between four and a half, five years, and with some regularity, there are announcements about one of these LTL carriers buying another one or merging with another one. So, there’s consolidation in that part of the industry just like you hear about airlines merging, and for some of the same reasons, operating efficiencies and things like that.
Will: I guess it sounds like that’s more capital intensive because you have to have these cross-docking facilities and more scheduling and infrastructure. It’s not just send an empty truck to point A, drive it all the way to point B, but you have to have these cross-docking places.
Scott: That’s exactly right, yeah. One of the things that happens that’s interesting is that last year, there was a huge spike in demand relative to supply. Trucking rates went up, and it pushed a lot of truckload shipments into the LTL category. What happened then is the LTL carriers got overwhelmed and actually ended up shutting down new business to some of their facilities.
Scott: They actually what’s called embargoed their facilities. They would say, “We’re not taking any more shipments out of the Minneapolis region,” because their cross dock just couldn’t handle any more volume.
Will: Wow.
Scott: So they basically ran out of capacity to handle that hub-and-spoke model, and that was a knock-on effect of what happened with the broader truckload portion of the industry. So that moves us nicely into the truckload. Truckload is very different. I mentioned 40,000 carriers earlier when we only opened. If you own a truck, you could be Bachman Trucking, and you’ve got a truck, and you keep it rolling, and it’s your business, and you’re a one-man show.
Scott: Or that’s at one end of the scale. The end of the scale, you’ve got massive truckload companies: Schneider and JB Hunt and Werner and things like that, and what’s different about that business is that it’s mostly operates… At least a spot market operates like a commodity market. So, unlike LTL where there’s set pricing, the simplest answer to how much would it cost to get a shipment that the full truckload to go from Queens to Chicago, the answer to that is simply however much it takes to find someone who’s willing to do it for that rate.
Scott: So it’s a classic market supply/demand situation. What’s fascinating about that, what’s been fun for me in the analytic space is that means that pricing is moving around all the time both in geography and time along both those vectors, and it’s absolutely fascinating, and it’s all driven on supply and demand. Last year was a banner year for a lot of folks in our industry because truckload rates went very high, which meant there was lots of profit going around, and that happened because of economic strength.
Scott: So, strong demand for trucking, for moving a freight, and a constraint of supply which happened as a result of mandate on electronic logging devices that control how many hours truck drivers can operate. What happened prior to the electronic logging devices was that the drivers would cheat on their log books and drive a little extra beyond what they were supposed to and the electronic logging devices have made that much more difficult to do.
Scott: The effect of that was constraining capacity. So, the market responded to that with increased rates. So, long story short. It’s a classic market. In that space, we serve as a market creator, matching carriers that need loads, and shippers that need full trucks. Then there is an arbitrage model in that in which we are trying to buy low and sell high. Sometimes, that’s successful, and some percentage of the time we’re underwater, we actually sell it for less than it cost us to move it. But that’s the nature of the truckload brokerage business.
Will: Yeah, so in terms of some competitors that are coming into this… I mean, most listeners are familiar with ordering a ride-sharing on their phone, an Uber or Lyft. What’s going on with that in the trucking space? Can you call up an app and just say, “Bring me a tractor-trailer truck to pick up my load”?
Scott: That is the argument that some firms are making including Uber. So, Uber has a division that’s called Uber Freight, and they have an app, and they are trying to allow owner-operators to find freight via the app. They’re trying to allow shippers to post. There’s a couple of challenges to that. First of all, we take that competitive threat very seriously.
Scott: But what I can say is that it shouldn’t be underestimated, the complexity, and the difference between hopping in a car to go downtown to go to dinner with some friends versus moving for a cross-country. Those things are not equivalently complex. There’s different kinds of equipment types. There’s different kinds of delivery appointment needs and scheduling needs. There’s refrigeration requirements in some cases and lots of different elements.
Scott: As I said earlier, things go wrong, and someone has to deal with solving problems in real time. But the argument that Uber’s making, and another firm that’s out of the Pacific Northwest called Convoy, is that there can and should be a lot more automation. So they entered the market saying, “We don’t really need human beings to do this. The technology can just do it.” What’s been apparent over the last several years as firms have tried to do this is they start out with this theoretical ideal of, “We just have all of this fully automated and no people involved.”
Scott: Then it turns out things go wrong, and customers get upset, and carriers get upset. They can’t get anybody on the phone to talk to help solve the problem. They get very frustrated. Then sure enough, these companies are hiring lots of operations folks to deal with those problems. So, in my mind, the future is a blend. There is no doubt. I mean, Echo has been a technology-enabled company since its founding over 10 years ago. That’s going to continue to accelerate.
Scott: We are making very large investments in our IT space, and there is going to be a little bit of a race to what I call bionic reps because I don’t think… There’s going to be automation that’s truly fully automated and then there’s going to be partial automation where you basically have humans being assisted with lift machine intelligence and things like that, which is again one of the things that I love about my role is that I’m involved in developing some of those intelligence sources, and it’s really fun.
Scott: So I feel like on the one hand, we take those competitive threats seriously. On the other hand, we feel well positioned to be responsive and to evolve as the industry is evolving.
Will: Any other major, new entrance in this space, in this brokerage space? Amazon doing anything?
Scott: Yeah, so Amazon… What’s interesting about Amazon is it depends on how you count them. There’s some hubbub recently in our space based on an article that got published about Amazon’s freight brokerage, and there’s not a ton of detail. We don’t know all the details and exactly how that’s operating, but we know that Amazon has an enormous freight spend, we estimate somewhere in the ballpark of about $20 billion.
Scott: So, what we believe is most likely happening is once they’re pushing a lot of freight from location A to location B, then it’s reasonable for them to help the carriers they work with getting their trucks positioned back from location B back to location A. If they can find freight to put on those trucks, that’s incremental profit for the carriers, which then probably translates to lower shipping rates for Amazon.
Scott: So, it’s a situation where we take it seriously because those back hauls could be getting very, very low costs, which would put pricing pressure potentially on us. That’s the bad news. The good news is we don’t have any reason to believe that they want to be in this business beyond supporting their own supply chain network. So, our hope is that there’s some limit to how much they plan to grow this particular business beyond the footprint of their existing supply chain.
Will: Now, I know you can’t talk any anything internal to Echo, anything that you’re working on. But from your role as leading analytics, could you talk just broadly in the industry of the types of analytical challenges and the opportunities for companies in this space to use data and analytics to really drive the business performance?
Scott: Definitely, I think a couple things. First, there’s been a lot of analytic activity in this industry for a long time, but it’s been… At least from my chair, it seems to have been almost universally oriented on asset optimization, or I should say utilization of assets. So what I mean by that is if you own assets, you want to make sure they are full and working as often as possible, and that you are filling the trucks optimally, that you’re running them optimally, and that makes all the sense in the world.
Scott: There’s lots of operations, research, math that large shippers use to optimize their loads, their locations, their routes, their timing, and similarly on the LTL side, on the truckload side for the big players, they’re run lots of analytics to optimize their what’s called operating ratio. It’s a funny way that our industry looks at profitability. Essentially, one minus the operating income rate is called operating ratio, and that’s the standard term in our industry.
Scott: So instead of saying that a company operates at a 2.5% operating income rate, they would say, “They operate at a 97.5 operating ratio.” But anyway, there’s been a lot of focus on that. Now, I come from a retail background predominantly. I spent most of my career at Best Buy and then Ulta Beauty, the cosmetics retailer. So, I come in with a pretty different lens, and I was really surprised at the lack of some of the things that coming from retail would be pretty fundamental really detailed understanding of price elasticity, for example.
Scott: Really detailed understanding of the matching between client needs and client services. These are things that retail has been obsessed about for a long time now and feels like it’s still very early in this industry. Again, that’s the thing that’s exciting to be a part of.
Will: Let’s talk about each one of those. So, client needs and client services, so how would client needs differ? Tell me. I don’t even know what that means.
Scott: In the LTL space, for example, some folks might view it as a relatively commoditized industry, but there are definitely service and performance differences between the different carriers. So, what is service? Well, on-time pickup, on-time delivery, the frequency with which goods get damaged, the speed at which things are delivered. So, there are different service components, and as is the case with any given consumer, everyone has a sense of value that’s a relationship between the price they’re paying and the benefits they get.
Scott: Same with B2B clients. Some are more concerned about price. Some are more concerned about service. So, one of the things that we’ve done for LTL for example is produced a value score on every carrier to help our reps make smarter choices about which carriers they recommend to our customers based on a combination of price and quality scoring.
Scott: So, just as a quick analogy, if you ever go to StubHub, one of the places I like to get tickets sometimes, you can sort the available tickets based on price, or you can sort them based on best available, or you can sort them based on a third column that’s called value, and what they’ve done is they’re looking for the biggest differential between the amount that’s being asked and the quality of the seats.
Scott: So, we’ve essentially created a similar concept for LTL carriers to assist our reps in making the right recommendations for our clients.
Will: Okay.
Scott: What we don’t yet have is that on an individual client-by-client basis. That’s the next step of where we’d like to get to.
Will: Yeah, so you’d help then advise your clients and say, “Look, we can get you the cheapest price from Boston to LA, but if it’s really important that the speed is important or if this shipper has a really good reputation for low damage or for the better on-time delivery performance or some other metrics, then you could help them optimize for something else.” They might pay a little bit more, but they get what they really care about.
Scott: Exactly, right. Yep.
Will: Then in terms of price elasticity, so that would help you what? Figure out with some clients, like you might be losing some. So, you might say, “Well, in some cases, we want to… ” Some clients are maybe more sensitive than others in terms of rates and so forth.
Scott: That’s exactly right. So there’s two different sides of the price elasticity. One is continuously studying and advising where our margins are and how that’s relating to our win rate and understanding as detailed as we can the balance between volume and profit. Then the other side of it is providing insight back to our carrier partners on how their pricing is impacting their volume because as a large volume channel for the LTL carriers, for example, we move a lot of volume on their behalf, and they will come to the folks that manage those relationships here and say, “If I move my pricing by this much, how much more or less volumes do you expect that they’ll get?”
Scott: We can actually provide them with some really great visibility on shifts in their volume based on how much they move their price, and that’s something that helps us be a stronger partner to them and continue to get favorable terms and have a strong relationship with those folks.
Will: Could you give us any ideas for someone who… the listeners who are doing analytics or think about taking an analytics role. What does it take? You’ve done this now at multiple places, Best Buy. Like you mentioned, Ulta, your current role. What’s it take to go from having the analytic insights, then to actually transform the organization and take advantage of those insights?
Scott: That’s a really good question. A couple of thoughts. First, I believe that there’s a sales and marketing aspect to essentially every discipline, and analytics is not actually different than that, and I think that many members of the community would like to believe that because analytics is always looking on a fact base and trying to make demonstrably superior decisions, that that is self-evident and good enough for everyone to just listen to it.
Scott: But that’s not actually true. So, one of the things I tell my team is we’re actually in the business of credibility. Maybe what we’re trying to create is superior decision-making at our company whether it’s the CEO of the company or the sales rep on the floor and every layer in between. That’s our driving mission. But the way that we can actually affect that mission successfully is to create credibility with our audiences because if we produce fantastic answers, and they’re not listened to, then we haven’t actually accomplished what we’re trying to accomplish.
Scott: So that requires a couple of different things. One, it requires the business respecting that we understand their perspective well enough to take into consideration. It means that we are showing empathy to the pressures that they are under that we are respectful of certain elements of the conventional wisdom, and we’re listening to the experience, perspectives, and that when we come with an insight that challenges conventional wisdom that we do so with the right attitude, that we show up with the right combination of humility and conviction in our own skills and capabilities, but also the humility to engage in a genuine dialogue with folks that maybe don’t have the analytic skills, but may have been in the business and running it on their gut for a very long time.
Scott: So that’s thematically some of the things I think it takes to create that influence. The other thing I’d say is I like to say the change management is a hand-to-hand combat sport. I think that sometimes folks believe that they can write a white paper or deliver one PowerPoint presentation, and then everyone should just change their mind, and that’s just not been my experience anywhere.
Scott: What I find is, yes, you can publish your paper, and you can give your presentation. But then meeting after meeting, coffee break after coffee break, and lunch after lunch with lots of individuals, you have to continue to explain and answer questions and drive the change in mentality and that every time I have tried to take a shortcut, or I’ve watched people try to take a shortcut, it’s almost always come up short in terms of driving the kind of transformation that they want to drive.
Will: I can imagine for the kind of work you do, it’s not just coming up with the insight, but then all sorts of processes and systems have to change, and people have to incorporate it into their role. So, it might require IT changes and new tools for people, just training people, but also new algorithms or new company policies. So, a lot of things have to happen to go actually capture the business impact.
Scott: That’s exactly right, and actually, you know what? I had apologized when we got on the phone today. I was a few minutes late because I was literally in a meeting with our CIO and our head of engineering for IT, and we’re talking about instituting a couple of the algorithms that we have developed into operational systems and how we’re going to do that and what the optimal way to do that is, and it’s interesting.
Scott: I actually think of my function in many respects like an R&D research laboratory. I was trained in the natural sciences. My undergraduate degree is in biology. I believe that there’s a whole nother topic we can get into, but I believe there’s too much data and not enough science in data science today. But we look for new insights. We test them in a digital laboratory environment. Then we try to pilot them in vivo in the real world, and if we find that those things are proving out value in a pilot mode, then it comes time to industrialize those things and put them in production.
Scott: That process can take quite a while, and we have several things on the shelf that we’ve developed that we think we can add a ton of business value, but there’s lots of products and elements on the development roadmap for IT, and we have to get our place in line to get those things deployed. So, 2019’s going to be a particularly good year. I’ve got three major elements at least that will get deployed this year that I think are going to produce a lot of value, so that’s exciting.
Scott: But yeah, it’s definitely a huge component of this stuff, seeing the light of day and actually delivering value.
Will: Talk to me about the technical skills of doing data science that are required for someone in a role like yours who’s leading an analytics group.
Scott: I’m probably a bad example, personally. A lot of folks who have my kind of job came up through computer science or have advanced degrees and mathematics or physics or things like that, and what I like to say about my career that’s strange is most careers go from a very specific specialization and become more generalized, as they advance through career. My career has done almost the opposite. It’s basically started general and has become more and more and more specific, which was not really intentional. It just happened.
Scott: But I would say that as a leader, the number one thing is being able to have the respect of a team. Analytic professionals tend to be pretty skeptical, so you have to be able to handle that skepticism and focus it for the benefit of the company challenges where necessary, and hone it and then continue to develop their skills. The other thing about a group of analysts is that they tend to be hungry learners. So, if they’re not getting development opportunities, if they’re not working on interesting problems, they can get bored very easily, and it’s a real threat to retention in terms of the team.
Scott: In terms of folks that are in an earlier phase, the sort of top skill sets that I’m looking for usually involve SQL coding, R coding, which is a statistical language that’s become very popular. Python is a growing option, and the other thing I’m really looking for… I hinted at this earlier. I’ve come across quite a few people over the last several years that can manipulate a database successfully without actually producing much material benefit, and I don’t typically have a lot of folks like that on my team.
Scott: But that comes back to the question of the scientific method and the ability to look at a problem and break it down to its component parts and translate business challenges into analytic approaches, and that is in rarer supply right now in the marketplace.
Will: Well, on that, tell me a little about how you screen and hire members of your team. Are you going to a practice database and you give them a practical exam, say, “Take a look at this and tell me some insights”? Or is it purely face-to-face, interview-based? How do you screen and interview candidates?
Scott: We do actually have some tests that we deploy. If we’re going to hire someone, we require that they are adequately skillful with SQL. We do issue a test as part of our hiring procedure. Then I have some questions. In fact, I’ve stolen some of the classic MacKenzie mindset on interview style. What I’ll ask the person to estimate some values for me, and I’m really interested in how they break down that problem.
Scott: I find that some of the things I’m… There’s technical aspects like the SQL test, and generally, like my director of data science, I’m having him kick the tires on their technical proficiency. The stuff I’m usually looking for has more to do with mindset. It has to do with curiosity, has to do with tenacity, has to do with whether or not I think this person can show up with the right mindset and be a real partner to the business.
Scott: When I opened with talking a little bit about what it takes to actually affect change, those are things I take every bit as seriously as the technical aspects. So, yeah, those are some of the things we’re looking for.
Will: Yeah, I mean, do you have an opinion for the technical skills on whether someone learned that at a… got a computer science degree at a fancy college, famous college, or if they’ve gone to one of these boot camps that is now coming up with like Metis and other Flatiron School and other boot camp schools, or if they’ve done self-study and learned it by MOOCs. Do one of those matters a lot more to you or…
Scott: No, I don’t care at all. My attitude is you can either bring it, or you can’t bring it. Whatever means by which you acquire the skills and the mindset, that’s a person’s individual path, and I feel pretty strongly about staying open-minded about, but there are many… I’ve had very successful people that have come from very diverse backgrounds, and I think that’s great. I think it makes for a great team.
Will: So somebody who studied English or philosophy in college and then said, “Hey, I really care about data science,” and just done self-study or boot camp and gotten super sharp in that?
Scott: Yeah.
Will: Yeah?
Scott: Yeah, exactly right. Yeah. I actually think philosophy is a great preparation for data science. I know it may sound weird and a little countercultural, but if someone actually studied real logic, then they have probably had lots of practice in breaking down problems that are quite complex into subcomponent parts and understanding the relationship of hypotheses and theorems. If they can do that, and they have the mental… the cognitive flexibility to apply that in a different space, that can be really effective.
Will: How do you continue your own professional development other than just doing the job? Do you go to industry conferences or just talk with peers at other firms or continue studying stuff, like reading industry trade journals. How do you stay current with all the developments that are going on in the data analytic space?
Scott: Yeah, that’s a good question. I guess first, I accept the reality that I can’t possibly keep up on everything. There’s just too much. But I definitely attend conferences. I was just at the Spring Symposium for the International Institute for Analytics. Portland, the spring. So I definitely attend events like that. I talk to colleagues, former folks that I’ve worked with. I read stuff. I’m a relatively voracious learner, but not necessarily… I’m not sitting around reading textbooks.
Scott: In fact, probably the most common place I go is Wikipedia because it’s fast and easy and available, and it’s right almost always, and I learned early on in my studies as an undergraduate that I’m not afraid of reading technical journal articles. For example, I’ve been studying behavioral economics a lot lately. I’ve been reading books like Thinking, Fast and Slow and The Undoing Project and things like that, and then those have references to actual papers that were written by Danny Kahneman and Amos Tversky.
Scott: So I go find those papers from 1979, and I read those papers that are seminal works in behavioral economics. So, sometimes, if it’s an interesting enough topic, you just keep pulling the string. I’ve become a huge fan of endnotes and bibliographies that if I’m reading an interesting topic, sometimes before I’m even done with the article, I’ll go to the end and look at what it’s referencing, and I look for some of the most influential papers or topics, and then I go straight to those.
Will: For someone who’s trying to… an ordinary, intelligent person who doesn’t know a ton about analytics, but is trying to get smarter, are there some articles or websites or books or any resources that you think are some good places to get started? Courses, anything that you think are some good places to get a foundation?
Scott: Yeah, let’s see. So, a couple things. First, a couple of my favorite books that I think are really nice would be How Not to Be Wrong: The Power of Mathematical Thinking by Jordan Ellenberg. He is a math professor at the University of Wisconsin. That book is terrific. Nate Silver’s The Signal and the Noise, that got very popular when he was successful with the election predictions, is a really well-done book that I think covers a lot of topics.
Scott: Let’s see what else. Some of my favorite podcasts. I know it’s funny because again it’s… I don’t spend a lot of time on highly technical stuff. I mean, yes, I’m reading some of the academic papers and those get technical, but by volume that’s not usually what I’m consuming. My favorite podcasts are Freakonomics and Hidden Brain by Shankar Vedantam from NPR does fantastic job.
Scott: So those are all places where… Again, so for example, there was a piece on Hidden Brain called the Vegetable Lamb. He interviewed a philosopher of science named Cailin O’Connor from University of California. I was so impressed by the interview that I ran right out and bought her a book called The Misinformation Age. So, I’ve been reading that.
Scott: I always stumble a little bit on this question because I don’t have these seminal references other than those first couple books I named. For me, it’s so organic that it never feels like I’m getting things from just one source. Like I said, I got a couple of favorite podcasts. Beyond that, it’s just always chasing the next reference.
Will: I love that. I love that. Going to the endnotes. Well, Scott, I have really enjoyed hearing about the freight industry and all the cool stuff that’s going on there when you see those trucks whizzing down the highway, and who knew that Uber was getting into that space. I guess not surprising. They’re trying to elbow their way in and really cool stuff what you’re doing. So, thank you so much for being on the show.
Scott: You’re very welcome. It’s great to talk to you.