 Hello Global Supply Chain Designers, and welcome to our third and last life event of SC2X, Supply Chain Design. I'm Sergio Caballero, the coach lead of SC2X, and today with me is a person who doesn't need an introduction. He is a celebrity in our MicroMasters community. I'm a celebrity in my own head. Chris Caballero. As you might know, Chris is the director of the MITx MicroMasters Supply Chain Management, and he's also the main instructor in our course. So welcome, and thank you for being today with us, Chris. Great, thanks Sergio. I'm excited to be here, but I'm here with the course owner. Sergio owns the course, so I'm just a guest. Okay, thank you, Chris. So this is the plan for today. This is what we would like to cover in the next hour. I'm going to start by giving a brief recap of SC2X. Then Chris will be talking about the real application of one of the topics that we covered in the course. So freight transportation procurement. Then he will be giving more information about the next course in the series, SC3X Supply Chain, Supply Dynamics. And then I'm going to be giving additional information about the upcoming final exam. So let's get started. So we started this journey. We started this run of SC2X with more than 7,800 enrolled students in our course. So 734 students are pursuing the certificate, and our learners are coming from 159 different countries. So in this course we have a lot of interaction, so all of you are having very active in the course. So we receive more than 3,000 posts in our discussion forums and hundreds of emails every week. So in terms of context, in this course we covered the design of physical flow, also the design of information and financial flow. So that's a lot. And we might guess which are the most difficult parts that we would like to hear from you. So that's why I'm asking the following question. So I would like to have your input. So which was the most challenging week of SC2X? So you can go to your now on the slide or so you can answer the question. And as options we have the different weeks of SC2X. So again the questions and please provide us your input. Which was the most challenging week of SC2X? Okay, so we are to start receiving some answers. So I mean it's tied between supply chain network design. So our week two accounting and also a week nine discounted cash flows. I think it's more or less what we were expecting. Yes, someone new coming in. It's a totally new topic. The second half of SC2X totally changes. If you've done zero and one and the first half of two then hitting the second, what is it? Week eight and nine of the economics, the financial accounting. If you haven't done that before it can be very confusing. Okay, excellent. So the next question I would like to ask is about which one is your preferred week or what's your preferred week? Which week of SC2X did you like the most? Okay, please give us your input. Did we put the final exam in midterm on there as a week? No, that's not an option. That's not an option. Okay, so we're starting receiving a couple of answers. So procurement and discounted cash flows. So challenging and also they like this week very much. Yeah, because the discounted cash flows is something you're going to use whether you stay in supply chain or not. It's a great tool to justify investments. So you were a civil engineer too, weren't you? No, I'm an industrial engineer. You're industrial. Did you have to do, you must have done economic decision making with cash flows? Correct. I was a civil engineer and that's where I learned it initially. Just kind of cash flows and your just economic viability for any kind of investment. So it's a really practical tool. Excellent. So the majority say that this is their preferred week. Okay. Nothing that I taught. I'm very disappointed. So hopefully this is not related to the factor. We'll see. We'll see. Okay. Thank you. And the last question that I would like to ask you is general feedback about the course. Okay. So please describe in one single word your overall experience in exit 2x. Okay. So let's see. Challenging. Awesome. I think my family members might be typing this answer. Yeah. Exceptional. Amazing. Insightful. Okay. Great. Thank you for your input. And let me tell you that we are in the process of redesigning exit 2x. So please in the next few days, if you can send us an email to exit 2x help at MIT. Thank you. Just providing us with two or three things that we could improve in our courses and two or three things that we should keep doing. That's going to be awesome. Okay. So we value a lot your feedback. Okay. Without further ado, so let's talk about the main topic of this session, freight transportation procurement. Please. All right. Thanks, Sergio. So freight transportation is something I've been working on since the early 90s. My dissertation was based on it. I think Yossi talked about combinatorial optimization for truckload transportation. That was my dissertation. And so I practiced this for a number of years as well. So I wanted to dive in on the topic a little bit and kind of add some to it beyond just the mathematical models that you went through with Yossi. So what I want to do, just not working, there we go. I'm going to talk a little bit about what freight transportation looks like. How is it usually procured? What are some of the challenges? And I'm just going to talk about things that are kind of happening in the market now because it's kind of a changing practice and we'll get into that. So before we talk about how you procure, let's talk about transportation. So what we're talking about here is what's known as full truckload or TL, a truckload transportation that goes from one origin. You load it on the truck and that truck does a line haul move and you unload it at location B. It's point-to-point moves. It is not less than truckload that consolidates small shipments and put them together. It's a point-to-point move like a taxi cab as opposed to a bus. So this is the type of procurement we're looking at. So the way it operates in generally, this is the United States and so what you'll see a truckload driver would do, they might pick something up in one location, drive loaded, that's the solid line. In this case they go to South Dakota, they drive empty, that's that dash line down to Denver, a loaded trip to Chicago, unloaded down to Indianapolis, a load to Dallas and so forth. So what you see is it's a combination of loaded moves and unloaded or empty moves and the drivers usually only get paid and the carriers only get paid by the shippers for the loaded moves. So there's a lot of balance issues. So you don't want to deliver to a place where you have to drive a far distance to get a follow-on load. That's the hallmark of a truckload move or a direct transportation mode and that affects a lot of the activity and the economics of the mode itself. Just to give you some perspective, this is numbers from 2016 in the United States, the total transportation cost estimate is about $894 billion. Of trucking, motor carrier is about $600 billion of that, so the vast majority of transportation costs in the United States for logistics is truck. Truck can be divided into three components. Full truckload, which I described, less than truckload, those are the companies that pick up smaller pieces less than full truckload, usually pallets, 500 to 2000 pound type loads, consolidate them, move them through a set of terminals. And the last one, a private or dedicated, that's where a company whose primary business is not trucking has their own fleet. Think of a retailer like Walmart or Target where they have their own trucks. And so you can see that that segment is almost as big as the full truckload for hire segment. When we talk about truckload procurement, everything that Yossi talked about that I'm going to talk about is for that full truckload. So it's about $270 billion. It's a big business. To put this on perspective, it depends how you count, but there's over 150,000 truckload firms in the United States. The average size is less than three trucks. So they're very fragmented, it's a very competitive market. It's big. And so something that you'll learn in SC3X is how you analyze a market. What I show here is this is Porter's Five Forces. It's a framework for understanding how a market interacts. And if you take SC3X, you'll learn all about this because we use this a lot. And what it is, is where are the threats in the market? Is it new entrants to the market? Is it substitutes, producers of substitute products? Is it the buyers or the suppliers or is there industrial internal competition? And so it's really understanding where the competition comes from. Where is the power in the industry? And the big takeaway for truckload is that there are very low barriers to entry. To start being a truckload carrier, you need to have a truck. That's it. So it's very low barriers, low barriers to exit. So you're not trapped in the market. So you can go bankrupt very easily. So whenever you have low barriers to entry, low barriers to exit, a lot of internal competition, it creates an almost perfect competitive market, especially when you have a lot of suppliers here. So this is something that you'll see more if you take SC3X. We'll show you how you can apply this framework to other industries and other companies. But for truckload, what it says is that we have a very, very competitive market. Lots of carriers are competing for the same business. That's the big takeaway. Okay. So how do we do procurement? Let me see if I can... There we go. So now I'm going to go... That's what the market is. That's what we're trying to procure. Like a shipper, like Procter & Gamble, like Walmart, like Starbucks, anyone who's buying transportation services in the truckload market. The idea is that procurement is just one piece of the whole management side. You have the big network design, which Sergio led you through earlier. Where do I set up my facilities? How do I flow my product? Tactical modeling is more... Where should I set up my private fleet? What should my private fleet take care of? Procurement is, okay, now I've got my lanes where I think I'm going to be going. How do I get the carrier and assign which carrier to which lane or which part of my network? So this is all done ahead of the actual business happening. Then we get into execution tasks. So everything I talked before is all planning. That's what you've done mainly in this course. Then now you're in execution. So a load drops and the question is what do I do with this load? Which carrier do I select? How do I ship it? Do I consolidate it? And then how do I route or schedule all those kind of things? These are two very different tasks that are usually handled by very different parts of the organization. I don't know any company where this is done in the same group because it takes different skills. Generally, if you've taken these classes, everything that we have done so far is more of the planning side. To be honest, because that's where your optimization models are used. That's where you try to set up the plan. Execution is really using the plan or executing on the plan. And so it's a different set of skills. We'll talk about how this works together in a TMS, the Translation Management System, in SC4X. But this is just, I just want to give perspective of where the procurement falls here. So planning and execution. You can see how procurement kind of fits in this whole cycle. And I do this for supply chain, but this is really any activity for a company. You have some kind of plan. You set up how you're going to try to operate throughout the year. Then the next thing is the procurement. And this is an event-based activity. It might happen once a year, once every two years. You decide what carriers to use where. The execution, we talked about that. You're using the plan. So when a new load of shipment happens, you know who to assign that to, what carrier. And then hopefully after this, you do a review. You do some analysis and see what works, what didn't work. What carriers do I need to drop? And so I need to add things like that. So procurement is not an isolated event. It fits in the whole management cycle and the whole process here. So let's go deeper into procurement and see where it falls in. We talked, Yossi talked about in week four. Is that right? No, week seven. Week seven, I'm sorry. Week seven after the midterm. He talked about procurement and specifically had a lesson on transportation procurement using auctions. That's that strategic bid. And so this is pretty much commonly done now. It was new in the early 90s, but now it's the common practice within industry. And I'll talk more about that. There's a lot of interesting work here. If you want to get into academic sides, there's been a lot of work in combinatorial auctions over the last, gosh, 10, 15 years. There's been very little research on how effective it is in practice, to be honest. And I'll talk more about that in a second. But that's a widely accepted practice. The outcome of the strategic bid fits into the routing guide. The routing guide is essentially a catalog so that the person in execution can say, okay, I have a load going from Chicago to Miami. I look up what carrier have I assigned in my strategic bid for that. I award it to them. Now it's typically not done by a person looking through a catalog. It's done by a system, a TMS or transportation management system, and it matches. And then it also handles the problems of what if the carrier doesn't accept that load? Who do I go to next? So it tries to do that carrier selection and carrier matching process. And that's, again, something we talk about in more depth in SC4X. There's been very little research into routing guides because they're not that academically research interesting. Although we'll talk more about that. Most work on the research side, on academic side, has been in better ways to come up with a plan in your strategic bid. Anyway, so we've set the bid that populates the routing guide. And then if the routing guide fails, let's say the primary carrier can't take it, the secondary can't take it. Then it goes to something known as a spot market. And a spot market is where you procure a load on a one transaction basis. As opposed to saying, hey, Carer, you win this business for the next year at this rate for all the loads on this lane. It's this load on this lane on this day. And it's a one off. And the spot market has had its ups and downs. We'll talk much more about this in a second. There's been a lot of interest now by companies, a lot of tech startups. Uber for freight is one of the big names that you start hearing there. Actually, Uber freight is a subsidiary of Uber. And where they're doing matching, just like for Uber for passenger travel, they're doing it for trucking as well. But there's companies like Convoy and other ones out there that are trying to do this as well. And the idea is forget the contracts that you do in the strategic bid, ride the spot bid, ride that spot market. And so there's some interest there and people are doing some interesting work and I'll talk more about that. But this is the general framework, strategic bid populates a routing guide. And then it goes into if the routing guide fails, you have a spot market mechanism to make sure that your load gets handled. Now, if I look at the bidding process, there's really three big steps, what you do before, what you do during and what you do after. And we talk about this in the lessons, but I want to focus in on two aspects that make transportation different. First, what is being bid? Because we talk about it being a reverse auction. In other words, I'm auctioning out the right for a company to do business with me. But there's something really specific about transportation. I'll talk about that in a second. And then who should win what? That winner carrier selection process. So let me talk about what is being bid. All right. There. So here's the situation. Let's say I've got some business going out of Chicago, Illinois, and I've got this one, two, three, four, five, six lanes. This is what happened last year. I think this is what's going to happen next year. So if this is what happened and saying two loads went to North Fort Wayne, four loads went to South Fort Wayne, one to Logan's Port, three to Anderson, one to Tipton and five to Indianapolis. So you see what this is. This is kind of what happened last year. And I think it's going to happen next year. So I'm going to bid this out. The question is, do I ask the carers to say, hey, here are six lanes, give me a price for each? Do I say, no, I'm going to, I really only want four lanes. I'm going to consolidate, right? So the Fort Wayne lanes, let me go back, those two, they're not that different. So just give me one rate. I don't need different rates for those. And I'll consider that one lane. So I go from six lanes to four. Or do I say, you know, Indianapolis is one lane and then anything else going to Northern Indiana, give me a cost per mile and I'll apply that to anything within that area. Or do I just say, forget it, give me some rate per mile for anything that comes out of Chicago. So if you can activate the poll for this, and I want you to decide which would be the best thing for a shipper to do. Should I represent this business as six lanes? That's option one. Four lanes, that's option two. Two lanes, that's option three. Or one lane, option four. So you have four options here. And this is something that seems trivial, but it really has impacts. And we'll talk about it as voting comes in. Is anyone voting? Not yet. We're very shy. So that's one of the things, usually when I'm doing an auction for anything, the question of what you're going to bid is, you know, it's obvious. It's the thing you have. Transportation can be represented in many different ways. So what do you bundle into a lane? What do you keep separate? Do you bid out regions? Do I bid out entire states? This is a question that seems trivial, but it has implications. And so there's a handful of people in, and it seems like the dominant one is that option two, people are saying this. All right, so let me just talk about the trade-offs of what you do. So if you did option one, one is if you did this for everything across your network, you might have 10,000 lanes you're asking the carrier to bid on. And if you have something that you had one load on, do you really want them to spend time on that? And so one of the things, it's almost exhausting. So it's something that you try to avoid. It also has another implication that I'll explain in a second. On the other side, this option, everything out with one price, that's too vague. And so what's going to happen is carriers will hedge. They'll increase because they don't know if everything's going to be a lousy load into middle of nowhere or something really close. So they're going to take an average, essentially, of what they expect it to be in worst case. So the best answers are probably one of these two, option two or three. And it's really a function of two things. One is specificity. And of course, this is the most specific. And then coverage. And this has the highest coverage. And it's really a trade-off between those two. Because this first one, the problem here is what happens if I have a load that's going to somewhere else that's in between, say, Logan's Port and Anderson. I now don't have a rate for it. If I just did these six lanes in option one, so that means I've got to go on the spot market to find that rate. So you have very bad coverage here, but it's very specific. So if you know there's a rate difference going to North Fort Wayne versus South, they'll definitely be able to pick that up. And so you have very specific rates that tend to be more aggressive. Hopefully that makes sense. On the other extreme, here I have coverage. So I know everything's going to fall for that. But since I'm so not specific, then my rates are probably going to be a little higher. They're like backup rates because there's a lot. You're forcing the carrier to assume the risk of the uncertainty of where the loads will go. So I would say, I don't know how the voting came out. Was it option two? Option two, definitely. That's the dominant. I think so. Even this will have some gaps, right? So I would probably do something like this, to be honest, because the other thing is these are relatively short haul lanes. And so even if you do option two, something could be between Kokomo. And this is by three-digit zip, which is just an aggregation of zip codes in the US. They're usually five. And the nice thing is if all of they fall in each other. So 469 is the first three digits. So that includes 46901, 46902, everything else. So they're all grouped together. The thing is, if something happens in the middle, you don't have a rate for it. So I would probably pick three, to be honest. But it's a trade-off. We've done projects with like CH Robinson trying to find the optimal region side you bid out. And it's kind of an interesting exercise. But it's just one of those things that make transportation a little different. Okay, now let me move to the other end of the process. Not what to bid, but how do you assign it? Who gets what? This is the carrier assignment problem. And this is one formulation using package bids. Yossi showed you some different ones, but this should be relatively familiar to you. I've got an objective function where I'm minimizing cost. In this case, all my bids are in packages, and you can have a package of one. And your decision variables are the number of loads on lane I to J assigned to carrier C under their case package. And then you have to have a binary variable that YCK, that is a one if it's assigned and a zero otherwise. So essentially what you're seeing here is the objective function that catches the cost and other factors. And then a series of constraints that make sure all the loads are covered. And that you have a linking constraint there to make sure that if you give a carrier some volume on a lane, that you actually know that Y variable gets triggered, some upper and lower bounds as well as the limits on the number of carriers. So there's a bunch of different formulations you can do here. The two key things I want you to remember or recall is that in the objective function, I can capture tradeoffs. And in the constraints, I capture hard constraints, things that must or must not occur. So what this is doing, and the reason why optimization based procurement is so popular, it gives you so much flexibility. And let me give you a framework to think about it because we use this not just in truck transportation, but in procurement of anything, because when you want to procure any service or products, there's things that are easy to measure like cost, that's usually pretty easy. But then there's the softer things. And so you have to figure out how do I handle those. And so there's a hand, there's only really three big ways you can handle things. One is you can ignore it. And that's what I'm doing here, things that are neutral. So in this formulation, for example, I don't care if the size of the trailer is 48 feet or 53 foot. I'm neutral on that topic. So my decision making doesn't tie into that at all. Now you can have other things on the extremes. I might have things that I ban. And so I could have something in here that carriers that I don't allow carriers in that haven't been financially successful for the last three years. In other words, a carrier who's been declared bankruptcy in the last three, five years, I don't let them bid, right? Or I might have a carrier on a certain lane that I'm not permitting to happen. And I simply need to add a constraint or that will keep them out of being considered. On the other extreme, there's some things that I will require to happen. Things have to happen. Like here, that first constraint is making sure that all the loads get covered. So I can force something to happen. But then in the middle, there's all these soft things where I can penalize or reward to disfavor or favor. So think of the idea of an incumbent carrier. So I've got a carrier I've been working with for years and years. I would like to continue working with them. And so I could favor them by having their bid that they submit be reduced by, say, 5% in my objective function. I'm still going to pay them what they bid, but I'm going to make them a little more attractive. I'm going to favor them in the objective function. Why would I do that? Well, because there's soft costs and there's a reason why I maybe want to have that incumbent carrier. A lot of the new carriers who've never done business with me before, they don't know my business and it's a risk. So I'm willing to give a little bit of a favor to them to reward them. Not a ton. And so you have to make a trade-off. But the nice thing about this optimization-based framework, I can do hard constraints and I can do penalize or reward on these softer constraints. And essentially, you're treating the objective function as a utility function. And so it's really letting you make those trade-offs between the things you can measure, the price, usually the costs, and those other things that are softer. And so that's the real benefit of optimization-based procurement. And it lets you make those two trade-offs. You use this by making the winners' assignment. And what I have on this chart is the horizontal axis is the bid amount that the carrier is submitting. And on the vertical axis, the total cost of the assignment. And so I can look at the direct cost very easily. So if a carrier bids $1,000 versus $700, that's very obvious which one is going to be more expensive as a direct cost. But then there's the soft costs. And the idea is generally, if the bid is a lower-cost carrier, their service won't be as good. Now, we've done some research onto this, but as a general rule, we can think of this that better service costs more. And the problem is those costs are not readily apparent. You don't know what their acceptance ratio is going to be ahead of time or their service level or their damage. So you have to use a proxy. And so what you're trying to do is not necessarily find the lowest direct cost. You want to find the lowest total costs. And so you're including those things you know, the direct costs, what they bid, and an idea of those soft costs. And you want to come up with this business optimal somewhere in that sweet spot, which is a tough thing to do. But the optimization model lets you make those trade-off and include some of those soft costs in your objective function. And that's why we use it. And so just like we do for network design, the way you would do this is by doing scenarios. You do a series of different runs. And so for example, on the left side, I might do a run where I favor incumbents by 20%. So I reduce all of their bids in the objective function just for the purpose of the model by 20% and see how they do. I might want to require that if you win a lane, you get at least a load a week so that you get at least a minimum volume. You want to select a carrier group that you might want to increase and make sure they get at least 10% more business than they had last year. You can do something where you say, I only want 10 carriers assigned. So you have all these different scenarios and you're essentially doing a decision tree. And so I've done bids like this where you've done over 100 of these different scenarios or profiles. And what you're doing is just exploring the solution space and trying to find out what you want to end up with. What do you want your business optimal assignment to be? The base case cost where you don't add any of these extra conditions is always going to be the lowest cost, always. Because as you know, hopefully by now, if I add a constraint to an optimization, it cannot improve my solution. It just can't. The best it can do is not have an effect. It'll usually worsen the solution. So the idea is you add all of these scenarios on these other conditions, you're going to lose savings. And I talked about this a little bit. The idea is that you really don't want to give out the base case solution to too many people because they'll hold you to that. When in fact, that's just the direct cost minimum solution. It's ignoring all those soft costs that you're now adding in to get to that business optimal. So this is the practice that's typically done. And it's done in practice. It was kind of unique in the early 90s, but now every company uses this for transportation. And so that's where we stand. So the dominant practice today is something like this. You have that reverse auction usually once a year, sometimes every other year. And for the amount you're putting on to bid out is the average cost per average volume per week. You invite your incumbent carers and select a few new ones. They submit their contract rates for a year because their rates will hold for a year. They're asking them, say, for the next 12 months, how much will you charge me on this lane? The primary carers are selected using that optimization and price fed into a routing guide. And they have kind of a gentleman's agreement or a soft volume commitment expectations. The alternates or the losers get fed into the TMS as backups. And tendering is made in practice and in execution to the primary. And then if they reject to the alternates and then only as a last resort to the spot market and contract rates are paid that amount through the bid that the bid determined throughout the year, irregardless of how the price changes. The only mechanism that most companies have in the United States is something called a fuel surcharge program. So if the price of fuel goes up and down, they take care of that. They fluctuate that and share the risk. But if the market gets tight or loose, the contract rate still applies. And so the question is, what's wrong with this dominant practice? Like I said, this is done by every shipper that I know. This is a dominant practice. Well, this is where things get a little tricky. And this is kind of an introduction into some of the stuff you'll do in SC3X. Because what we've done so far is focus on the model. We did this really cool optimization, which is always fun and came up with an answer. The challenge is, what does that mean in the larger scheme of things? And so what we have here is what's known as a dominant design. It's where a practice or a product dominates the market. And we usually think of this in terms of products or physical things. And it's a concept that's been around for about 20, 20, 30 years. The idea of a dominant design is that you have a product and everyone in the market uses it. And it might not be the best for everyone, but it satisfices them. Not satisfy, it's a satisficer, which means it's good enough for most people. And so when you see a new product coming into the market, it's usually very fluid. You see all these different variations. You don't know what it's going to end up. At some point, though, a dominant design starts emerging. And at that point, all companies focus in on how do I make it faster, cheaper, better. But the essence of the design of the product doesn't change because the dominant design has been selected. And then you enter into what's known as this specific phase or the complacency phase where there's this dominant design. And why does that matter? Because this is where things get disrupted because we're kind of being complacent. We're falling in this common practice. Nothing is changing and you just kind of do it because that's the way it's always been done. Let me give you a handful of examples so you understand what dominant design means. The classic example is cell phones. When did you have your first cell phone? Ten years ago. Ten years ago, so that was what? 2008? My first one was in probably 99? Maybe? I had to page your first when I was in a software company. But if you look at how they come out, think about the first phone you had. Mine was a flip phone. Then you had the little Nokia bricks, all these different things. But now if you look at your phone, and if I guarantee, not guarantee, but I'm pretty confident, if you pull out your phone, it's going to be about the size of your palm, maybe a little bigger, glass on front, touchscreen, you can't take the battery out. It's going to look like an iPhone or a Samsung. They really have come down and looked the same. And if anyone had a BlackBerry, those are kind of gone now where there's a separate keyboard. The vast majority of cell phones all look the same. Are they perfect? No. But they're good enough for most people. It's a dominant design. Let me give you another example. Bicycles. When bicycles first came out in the 1860s and earlier, no one knew what they're going to look like. People experimented with a big wheel up front, two wheels on the side, all these different experimentations. But eventually you hit a dominant design. And the dominant design is two wheels, one front, one back, same size, chain connecting power to the back. It doesn't mean it had to be that way, but it's kind of hit that dominant design. Let me give you one more because this one's fun. Typewriters. The first typewriter really started coming out after the telegraph, where you could now start doing letters instead of just dots and dashes with a Morse code. And there was all these different experimentations. The first one on the left in 1865 was based off a telegraph machine. Then you had this Hanson's ball, which was, interestingly, this ended up being what they used in IBM Selectric later. And then in 1873, you had the first QWERTY keyboard. And what's a QWERTY keyboard? If you look at the top left of your keyboard of your laptop, it's the letters in the top left corner in order. Introduced in 1873 for a bunch of reasons. No one's really sure whether it's to slow down because of the speed of the telegraph or not. But the interesting thing is as the keyboard's evolved, it really hasn't changed much. In fact, I bet your phone still has a QWERTY keyboard as does your laptop. And the question is why? It's not the most efficient because there's other layouts that are much more efficient that only use your primary fingers, but they're not being adopted. Why? Because you have a dominant design. So we give all these examples. And so the question is, what does this have to do with procurement? This has nothing to do with transportation procurement, right? But it does because something has happened. And something in what's happened in the U.S. is this. This shows from 2008 to 2018, the average rate per mile indexed off of 2009, I want to say, what the price is. So as it goes above 100, like for now 142, you're now 1.42 times the price it was on average in 2009. So it's an index. And so from 2010 to 2017, the average annual increase was about inflation, 3 to 3.5%. You know, your ups and downs and things like that, but generally it's been pretty stable. The last 18 months, however, have been very different. 15% on average annual increase. So we've had this rapid increase. We've had all this failure of transportation. It's an exceptionally tight market. And the way of doing contracts is kind of failing. This is a shot from a company chain of litics where they look at the spot versus the contract. Remember we talked contract or those rates you've gotten that strategic bid fed in your routing guide set for a year. Spots are when the routing guide fails, you go out to the market. Look at the price differentials between this. Again, this is an index off of 2016. But you can see that it can be up to 60% higher for a spot rate on average. So this is an average for each of those periods. So the individuals can be very high. So you see when the market gets volatile, you have more spot rates and you have higher spot rates. So you have all this activity going on and it's causing a problem. So you might ask yourself, say, OK, so maybe this just is a tight market or something. So why does this mean that auctions don't work anymore? And it's not that they don't work. In fact, auctions work perfectly for the situation where they were introduced. So in the United States, when people in the 90s started using optimization based auctions or any kind of auctions, we were shifting from a type of transportation where transportation carriers had monopolies on lanes. And so if I was a carrier and I had the lane from, say, Chicago to Atlanta, kind of in the middle of the country, that if another carrier wanted to come in, they had to petition the government. So I owned that lane so I could set my prices. Shippers were beholden to me. That changed in the 80s with deregulation. And so suddenly now companies needed to have a mechanism to find what carrier can haul for me. What should the market price be on each lane? And the best way to find a market price is an auction. It's the best mechanism when you have a bunch of different providers to make a price. What is the price going to be? Run an auction, you find out what the market price is. So that's why auctions became dominant because we were trying to search. We had a competitive, perfectly competitive market trying to find the best carrier to haul. What is the market price on that lane? That's why we ended up using auctions. But the problem is as we get more evolved, it causes a problem because as you look through, and this is another survey, here's a lane and the whole idea of auctions is to find a market price. But there's a question, what is the market price? So here's a lane and this is from data that I've had, 450 mile lane, over 10 shippers are on this, 3,000 loads per year and the median raised $2.93. I wanted to turn the poll on and I asked you to say, what is the market price? If I was going to run an auction on this, what would the market price be? And a market price you can think of as being something that if I just go out, this is roughly what the price I should be. And I say less than $2.85 a mile, $2.85 to $2.90, $2.90 to $2.95, $2.95 to $3 or above $3. And so we're getting some of you, well, we're getting a response. But as you go through and you look at this, an auction seems like the perfect way. There's a lot of noise around this, but generally you can say, oh maybe there is a market price somewhere in there. Maybe if you squint, you can see a downward trend or something. Just looking at this data. So I want you guys to vote and it looks like everyone's voting the same. That's okay. All right. So what I'm going to do is show you a view that shippers don't usually see. So a shipper might see this or if they go to a company that reports on the industry and they show these kind of rates. But these are average for a day. And so what if I break it down to, instead of an average per day, each load and each load would look like this. Notice that the horizontal axis changed. Let me go back. The median was what I say $2.93 a mile. So $2.93, you can kind of see where that is. But look at you have loads as low as $1.50 a mile and as high as $4.50 there in August. So you have all these different loads. So the question now is, what's the market? What's the market rate here? Is it really that band of the solid lines is about a dollar difference? I mean, that's a huge range. When you're talking a $2.25 to $3.25 per mile. So the question is, what is the market rate? It's even more interesting if you color code it by shipper. That's what this is. So this chart shows a lot of interesting things. Every shipper is a different color. So you look at some shippers like that red one that's right around $2.60, $2.70. That's the routing guide that you see. They're constantly going out at a contract price. It's the same. Same thing with the orange below it. The blue above it, the black and the aquamarine, I guess it would be. But the question is, where did those rates come from? Why is there such a differential? What happened to the shipper, the black shipper? Where suddenly in April, the rates went down by about $0.10 a mile. You can see these different things here when you dive a little deeper. And the takeaway from this picture is, we're no longer dealing with a pure market rate. Transportation is more complex. It's not as simple as it used to be, like a taxi, you call a cab up and you go. You're adding extra services. So you look at what the aquamarine shipper, they're paying about $3.25 versus the, go with the orange one that's paying about $2.25. What's the difference here? How come the one that's higher, they're doing more? You see there's more dots than the orange one. But why are they paying a dollar more? Are they adding more services? Do they have some other activity there that we don't know about? There isn't a single market rate. So therefore, an auction might not be the perfect mechanism for everything. So I bring this up because transportation is undergoing a big change in the United States, procurement especially, because we're finally recognizing that it's not a single thing. It's not just a truck anymore. There's a continuum of relationships that you can have with transportation, with any provider, to be honest. And you think of the continuum on one end, it's a spot market, transactional. On the other, it's private fleet where you own it. And in between there, there's all different gradations of your relationship. And you can use these different types of relationships differently across your network. So for procurement for these auctions, you can still use those for your large carrier, big partners for large sections of your network. But maybe you want to plan on using some different kind of contracts for different parts of your network, or maybe have dedicated fleet for other pieces. We've done a number of projects. We did one with Walmart a couple years back that tried to determine, where should I use my dedicated fleet? And then that determines what other lanes I send out to for hire. So you're kind of making this decision and treating transportation procurement, not just as an auction of all the same things, but as something you use the portfolio of different solutions for. You can use your dedicated fleet for some core contracts for other, alternates for others, and spot market for others. And you can even have finer grained. Because what's happening now is that procurement is changing. It's evolved in the last five years. A lot of it is technologically driven, and a lot of it is driven by just the change in the company's acceptance of analytics, because that's really what's driving it. Now, optimization-based auctions and bids will continue to be done, but it's not the one-size solution. You're not going to put everything into an auction. You need to find the right design, and you need expertise to identify, think back to SC1X here, it's segmentation. Look at my network, how are they different? How do I segment my network? How do I assign different types of relationships to different parts of my network? And then how do I manage that portfolio? And the rough rule of thumb is the more stable the volume is on a portion of your network, you want to use your own assets. The very volatile ones, you want to use more spot. So you want to identify those and assign your assets accordingly. So the call to action, we're going to see more automation. And there's a lot of companies that are automating this care selection process, and it's going to happen. It's happening more with Uber Freight, with Convoy, with Coyote, CH Robinson, all the brokers are starting to become much more automated. And the irony is it's not going to replace people because it takes more expertise. So you guys are going to be in more demand because you just don't use these blindly. You're going to have to understand how they can be used. So in this fluid phase again, you know, we're not in this dominant design. We're seeing a lot of shippers change the way that they procure. And there's a lot of experimentation going on. We're doing different thesis projects right now with different companies exploring index based rates. What if I have a rate where it will float a little bit based on an index of the industry? So if the market gets tighter, I'll pay them a little more. If it gets looser, I'll pay them a little less with the idea of keeping them more secure for that capacity. Maybe I give my volume based on the tiers of the volume in that lane. Maybe I have spot rates fit directly into my TMS. Maybe I have what's known as synchromodal freight assignment where I look for steady state flow and I assign carriers to that volume. And if the load doesn't materialize, I pay them a penalty. I pay them for equipment not used. So you have all of these different things coming in to augment and make the transportation procurement decision even more interesting. So auctions are awesome. It's still the backbone of procurement, but it's just the starting point now. And that's what I wanted to get across in this. So everything you've learned in this course so far is the starting point. And there's so much more that you can add on to it. And with that, I'll turn it back over to you. Okay. Thank you. Thank you, Chris. Thank you for the insightful presentation. So we have some room for questions. So guys, please use the slide to answer questions to Chris. So we have a couple of questions here. SK Shetty is asking, what is the maximum size of the post-bid MLP mill formulation? Oh, gosh, maximum size. Usually it's not... Because what you're assigning is lanes, not individual loads. So it's not quite as big as you think. And so the biggest one I've seen is probably 10,000 lanes across 100 carriers. It only gets complicated with package bids. And the interesting thing is package bids are not as widely used for a bunch of different reasons. So usually the solving the solution time is never an issue for this. The MIPS, the MIPS are big, but the trick is a lot of companies use network design optimization tools to solve the procurement problem because they're usually pretty fast engines. The thing that's a challenge is the size of the bid, you can overwhelm the carriers. And so really it's not a technical limitation. Those models usually run very quickly now. The issue is having too many lanes for a carrier to focus on. Okay, excellent. And there's another question. And Ivan, he's asking if there's a way to normalize the bid prices so that they can be compared with each other. You do comparisons, not in the bid practice. I understand what you're getting at, but you can do that when you're assigning this, you're comparing apples to apples. So I'm doing a bid on this lane, 10 carriers on this lane or this lane or this lane. So they're bidding on the same thing. If you want to normalize them, you can do that for competitive purposes ahead of the bid. So you want to see how am I doing against the market. So there's a bunch of companies that do indices. And so you can compare how each rate compares to the indices to see, look at the residual there and see how well you're doing, each carrier is doing compared to this market. And so you can look at it as a percent above or below market. But I haven't seen that too much in the procurement event. I've seen that used as like an index. After the fact to see should I look and re-bid some lanes or some of my lanes creeping up too high. So indices, and you can normalize just by comparing it to market in percent above or below market. But that's used to identify trouble lanes in my opinion. Okay. Thank you. Thank you, Chris. There's another presentation. So again, Esquichetti is asking, how big is the freight exchange markets and how its impact transport broken business? So freight exchange market is another word for spot market and the brokers, all of that. Honestly, it's usually about 10% in the United States. It's growing a little bit the last 18 months it's grown and I've seen some shippers do as much as 20 to 30% in that market. And I have ongoing debates with different people. We have graduates of the program here at MIT who've started freight brokerage companies and exchanges like this. It's a growing area right now. What's interesting, a lot of these companies are going out and they're saying they set the price constant that they offered to the shipper. Right? So they set a price just like a carrier, but they use their own, they figure how they're going to get a carrier to go to that price. And so what they do, they fix it on the sell side and on the buy side is where they do their magic and try to find, they make their money on the margin, on the differential. So what that means is they have to be very clever in how they set that price. And so you see a lot of really nice advanced modeling using machine learning that you'll learn in SC4X about how to set that price so that you maximize your potential profit based on what you know about what's going to happen in the industry. So the freight market exchange, it's getting bigger, but it's still, even in this tightest of times, less than 25% of the total volume. That's high. Okay. And the last, last question. So Chris is asking, the segmenting of requirements and assignment decision to use on-flit versus spot sound similar to the production planning problem on in-house versus outsource. Yeah. It's a make-buy. It's a make-buy or rent decision, really. And so, yeah, and you can think of it in those terms. It gets complicated for transportation because you're operating on a network, right? And so you have network effects. And so what's interesting for the project we did for Walmart, you're trying to find like tours for your private fleet to do, because if you are able to get your, a truck to take a load and then get them back home and those are both pretty consistent, that's great. You want to keep that. You want to give the loans to the four higher carers that go somewhere and you have nothing coming back because then they absorb that cost. And you assume that they can find another follow-on load for some other shipper, easier than you can find a load within your own network. So it is like a make-buy decision, a little more complicated because it's on a dynamic network. Okay. Thank you. So let's now talk about ACC-TEX. All right. So everything you've done so far, yep, everything you've done, there we go. So far has been about the models, which is awesome. So 0x, if you took it, you learned all about optimization. You learned all about some stochastic systems, regression, probability, statistics, some simulation, all those things. Then in 1x, you learned the trade-offs between inventory, between cost, fixed cost and variable cost, transportation and inventory, shortage and excess. And in 2x, you learned about the network flow, and you spent a lot of time here. Now, in 3x, we're going to kind of introduce reality because the real world's complex, you have to think about a little differently beyond the initial models that you set up. And we broke it up. The format's the same. It's 12 total weeks where you have a midterm exam in week six and a final in week 12. But we broke the weeks one through four and seven through 10 into four big two-week blocks. The first block is all about complexity. And you'll learn why supply chains are complex and understand what that means and the implications. And we'll introduce you to system dynamics, which is a really interesting way to model and manage complex systems, not just supply chains, other systems. So you learn all about system dynamics. We'll go into the bullwhip effect a little deeper. So that's all about complexity. And each of these modules end up with an in-depth case study. So we give you a framework for how to think. We'll give you tools on how to analyze it, and then we'll apply it in a more extended case study. So each of these modules consists of a very involved practical case study as well. So after you've done system complexity, we're getting to strategy because we're not doing supply chain in a vacuum. It's therefore a company that has a strategy. So it'll help you understand some of these strategic frameworks that are taught to include the one we introduced this talk with, Porter's Five Forces, very common framework. And then we'll talk about supply chain alignment because supply chain, you can talk about supply chain strategy, but that really has to align with the company strategy. So we'll talk about how you do that, how it fits, thinking about a supply chain as a process and how to improve them. And again, we'll finish that module up with a case. After the midterm, we'll go to global supply chains. And this is where things get weird, right? As soon as you cross a border, you have to, there's so many other considerations you have to consider, the currency differentials, the trade restriction, taxes are different. You have all of these different things you have to consider when you have global trade. And so how you design your supply chain, where you put your manufacturing, what markets you serve, we'll go into all of that with Dr. Bruce Arnson. We'll dive deep into this and we'll finish it up with a more involved case study again that help you understand and kind of cement all those tools that you've learned. Then in the last two weeks, we say, okay, there's all these exogenous factors. And these are factors that you have no control over, right? These are things that happen that you, you're out of your control, out of your ability to change. So think about disruptions, whether it's a manmade disruption or a natural disaster, these things affect supply chains a lot, especially if you're a global supply chain where you're all over the place. Professor Yossi Shafi will talk about how you handle disruptions and how you build a resilient supply chain. And then we'll talk in myself and Dr. Alexis Bateman will talk more about how you handle social, technological, economic, environmental, and political events or trends that impact your supply chain. Because one of the big things that's happening now, and you'll see this in the global supply chain, regulations have more effect on your supply chain than the cost of you doing business. And it's one of those things you just need to recognize. You need to be aware of all these restrictions that are coming in and these other factors that affect how you run supply chains. It's not just low cost. So all four of these modules kind of fit together. And what we're doing is introducing the idea of complexity and the real world into supply chains. So when you finish this course, then you hopefully will be very cognizant and we're giving you some tools of how to handle real life supply chains. And this fits you in to take SC4X afterwards that deal with how you use these models in a large scale real world situation as well. So that's what you can do for SC3X. What you've done in SC2X sets you up perfectly to go into 3X. Okay, thank you. Thank you, Gris. So let's now talk about the final exam. So the final exam of SC2X will be released next week, December 12th at 500 UTC. The exam will be available for one week. So it closes on December 19th at 500 UTC. So it will be a time exam. So this means that once you start the exam, you will have only four hours to complete it. So the exam will cover all the topics that were covered in the course. So that means the four weeks before the meter exam from week one to week four and the four weeks after the meter exam. So that's from week seven and week ten. So the final exam is worth 45% of the final grade. So please remember that your org might be your own. So I do not post any questions or any comments in the discussion forum. And if you need any clarification about a particular problem, please send us an email to SC2XHealth at MITEDU. Chris, thanks again for being today with us. And thank you all for attending to this live event. Good luck with the final exam. See you next time. Good luck. Take care.