Announcer: Welcome to MIT Supply Chain Frontiers from the MIT Center for Transportation and Logistics. Each episode features center researchers and staff, or experts from the field, for in-depth conversations about business education and beyond. Today, center researcher and instructor David Correll speaks with Dr. Angela Acocella about ghost shipping and freight transportation. The two explore cost implications for businesses and consumers, and the psychology that may lead to these so-called ghost lanes. Take it away, Dave. David Correll: Today we're delighted to be joined by a CTL family member, so to speak. Angie Acocella, formally a PhD student here with us, and now doing really interesting and amazing things in another part of the world. Angie, thank you for coming, and could you maybe describe the position you're holding now and what brings you back? Angela Acocella: Yeah, absolutely. So, as you alluded to, I did my PhD here at CTL, all in freight transportation, procurement, and now I've moved across the Atlantic to Europe, so I'm a post-doc now at Tilburg University in the School of Economics and Management. Doing a lot of similar, how do we improve procurement and transportation services, but now learning more about the European market, trying to make sure that we can do some interesting work and understand the differences between the US focus and the European markets as well. David Correll: Excellent. Gosh, well, thank you for coming back and sharing some time with us to let us know what you've been working on in your latest projects. I think our theme for this discussion is spooky things and Halloween, and your recent work on ghost lanes. So, as I was getting ready to talk to you, I read one of your recent interviews, and you had a sentence that really jumped out at me. "Ghost lanes happen at a frightening rate." So, in the spirit of our Halloween theme, what are ghost lanes and what's so spooky about them? Angela Acocella: Yeah, so, short answer is that ghost lanes are freight lanes that are contracted to motor carriers, which are never used over the course of the contract. So, maybe, a little bit of background for our audience members that aren't thinking about freight procurement as often as I do. So, when companies call those shippers, they procure their freight transportation services with motor carriers. These are the trucking companies that move goods over the highway. These shippers spend a lot of effort and time finding the right providers for their distribution needs at the right price, and that's because transportation services are an incredibly important piece, for shippers, to get their products to their customers, or to their retailers, to get products on shelves on time and undamaged. And so, transportation is going to directly impact customer service, the shipper's reputation in the market, and ultimately sales for the shipper. The problem is that these procurement processes can take months, sometimes six months, when contracts are only in effect for about 12 months. So, you're really doing this process just to start it over again halfway through the contract length of that year. And so, this procurement process, it involves shippers working with dozens or hundreds of carriers. Those carriers submit bid prices for hundreds or thousands of lanes within that shipper's network, and then the shipper needs to sift through all these bid prices and decide on a carrier for each lane. The problem here is that that process, as I said, takes so much time. The size, the cost, the complexity of that process is directly driven by the number of lanes that that shipper is procuring, and that's why it takes so much time and effort. So, getting back to these ghost lanes and what's so scary about them, we have these lanes that are included in the procurement process, but we don't end up actually using them. So, essentially, we're wasting all these efforts. And to give a little stat on why it's happening at such a frightening rate. So, we looked at a large data set of many shippers' procurement outcomes. We looked year to year, industry to industry, and found that across the board, about 70% of lanes that are procured in a given year end up as ghost lanes in that year. So, that's 70% of the procurement process that didn't actually need to happen. And that's why it's a particularly big problem that a lot of shippers, and carriers, in fact, have not put a whole lot of thought into, as we found in talking with a lot of industry folks. David Correll: Wow, gosh. Until I was learned about your work on this project, I had no idea or wouldn't have expected that it would be such a high percentage of ghost lanes. I wonder, for people who aren't familiar with the lane terminology, could you describe, does that mean Boston, Chicago is something that doesn't get freight from a shipper? What would a ghost lane look like in an example? Angela Acocella: Yeah, great question. So, essentially, say I'm a shipper that, I have product to move and it's going to go from Boston to New York. If I expect, I forecast my demand for the next year and I think I'm going to have, say, 10 loads per week, and I said to a carrier, "I'd like to contract you at a particular price." Then that carrier's going to expect to get 10 loads per week. What's going to happen in a ghost lane situation is that carrier expects 10 loads per week, and then for some reason, that demand doesn't materialize. Say, that endpoint in New York didn't end up having customer orders that we expected, and so that carrier's never going to be called upon to actually use its capacity. And so, this is what we're calling a ghost lane, a lane that just, we don't have volume on it. David Correll: So, thinking about the impact of that, and keeping with our spooky Halloween theme, if we think about what makes movie monsters scary, to be scary, they have to be threatening someone. So, the shark from Jaws is threatening the town from Jaws, or Mike Myers in the Halloween franchise is threatening the family, and everyone's scared of the specific threat that the monster poses to a specific community. Could you describe, if these ghost lanes are haunting our system, who are they a threat to? Who are they hurting? Angela Acocella: Yeah, so essentially, they're hurting everyone. For shippers, I mentioned, this procurement process is very cumbersome. It's complex. And so, this piece of it, those administrative costs are hurting shippers. More than that, for the shippers, we actually modeled carriers' behaviors year to year, and found that carriers that are receiving higher frequency of ghost lanes in one year end up having higher contract prices for that shipper the following year. So actually, shippers end up having to pay more as a result of ghosting these carriers. And to put more numbers around it, for every 10 percentage point increase in ghost lanes that a carrier sees in one year, that's going to result in a 1% higher contract price from the carrier to that shipper the next year. So, it's real numbers here. And that's in particular on high volume lanes. So, these important lanes for shippers. They're seeing higher contract prices on their important lanes within their distribution network. So, that's the shipper's perspective. If we take the carrier's perspective, as I said, it's hurting everyone. So, not only have we wasted carriers' time and effort in bidding on these lanes during that procurement process, and expecting to see business on those lanes, they've also expected to see revenue generated from those lanes. So, that's going to not only hit their expected revenues for the year, but also, very often, carriers will plan their networks around contracted lanes. So, if they're expecting to make capacity available from that, say, Boston to New York Lane, and then they have a New York to Chicago leg for another shipper, if there's no volume materializing from Boston to New York, now they're stuck with trucks in Boston that they need to get to New York in order to then get to their next leg in Chicago. But they're going to be moving empty. They're not going to get paid for that movement. And so, carriers actually will have to either have much higher what we call deadhead, so we have much higher empty trucks on the road, which is, of course, a problem in terms of sustainability and things like that. But it also is going to ultimately interrupt their network balance and their operations. So, we have these compounding issues for both shippers and for carriers. So, short answer, we should all really be scared of ghost lanes. David Correll: So, I hadn't thought about how much it hurts the shippers, but that is so interesting how your research points out that they pay higher rates for having ghost lanes in their system. And then, carriers, I certainly understand how they don't want to set up for business they don't get. Do you think there's a potential that consumers are also haunted by ghost lanes? Angela Acocella: Yeah, I think so. I think, in a lot of senses, that cost is going to have to be trickled down to the consumer. We see that a lot with, as transportation costs go up, we see shippers often noting how their transportation cost escalations impact their underlining budgets. And so, as a result, we will end up seeing probably higher prices for consumers as well. David Correll: Gosh. So now that we know to be afraid of them, I wonder, why do they occur? Is it a glitch or is it a ghost in the machine, if you will, that's causing this malfunction to happen? Or is it deliberate? Is it part of shippers' strategy? Angela Acocella: Yeah. I don't think anyone's being malicious, if that's an overarching story, I don't think anyone's being particularly malicious. I do think, though, that it is a strategic choice for the shipper, and this comes from what I'm calling a coverage strategy. It's in response to demand uncertainty. Shippers don't necessarily know when their customer orders are going to come in. They have to plan their networks a year in advance. So they need to be able to figure out how to cover the risk that's going to be incurred because of uncertainty. And especially because shippers want to make sure that they have contract rates on file for any demand that may materialize over the next year. So, they end up setting up these contracts on lanes, even if they're not sure if they're going to materialize, because they don't want to risk being exposed to more volatile spot markets later on if that demand does materialize. They also use this procurement process to vet their suppliers. So, they want to make sure that if demand does end up materializing, that they're going to be working with a carrier they know. That that carrier understands their business, they understand their facilities. They have those payment processes set up. So, essentially, they're using this strategy to make sure that if there's potentially volume that's going to materialize, we're going to have it covered. And so, for shippers, as I said, it's not malicious, but it is a strategic choice to make sure they have some availability and some certainty to reduce their risk going forward. David Correll: That makes sense, and it's really an interesting example of how problems arise from rational actions. No one's really behaving irrationally, but they're introducing a problem that affects all of us. Now, cue the music. Let's be Ghostbusters. Let's try to find the ghosts. Does your research tell you anything about which lanes are likely to be ghost lanes? Angela Acocella: Yeah. Yeah, I think once we realized how frequently this was happening, we wanted to say, "Okay, can we identify them at that procurement event before they even end up as ghost lanes?" So, we built this predictive model that took in characteristics of lanes before they're procured and said, "Can we identify if they're likely to become a ghost lane that year?" The first and the strongest indicator for if a lane is going to be a ghost lane in a given year is if it's new that year, meaning it was not included in the procurement event the previous year. And we actually found that 85% of ghost lanes were new that year. That's a huge amount. And that's, back to my example earlier as maybe that Boston to New York Lane, that New York customer was new in the distribution network, or it was a new warehouse that didn't end up having the orders that we expected. And we procured it just in case so that we had a rate on file, we had a carrier that we could count on or we think we can count on, but those orders just didn't end up coming in. So, first, the highest, the strongest indicator is if a lane is new. The second one was if a lane was procured the previous year, most of those lanes were actually ghost lanes the previous year as well, meaning we procured them, but no volume materialized on them. So, we have these ghost lanes recurring year to year, and we're not maybe looking at what happened last year in last year's procurement event and incorporating that into this year's procurement event. We're just seeing what we had in our network last year and including it this year. And then, of the rest of the lanes that were ghost lanes, most of them were very low volume, which also indicates maybe we don't need to be procuring lanes that are very low volume. Maybe we can be more strategic about the way that we procure them. So, I think the main thing is whether it's new or ghost lanes the previous year. And then, the next thing was about, how is a lane defined? So, we also found that lanes that are defined at higher geographic aggregation levels, so, region to region or zip code to zip code, as opposed to the actual facility address to facility address, those higher geographic aggregation lanes end up being ghost lanes also more often. And what's happening here is that often when we aggregate uncertain demand across larger geographic areas, we are aggregating multiple facilities to help smooth out potentially volatile demand anyway. But what we're really seeing is that, say a region that we define in the procurement event has three warehouses that we're trying to smooth the demand over. We're also defining lanes at those warehouse addresses as well. So, we're actually duplicating lanes. We have address to address, but then we also have higher, more aggregated geographic regions also defined as different lanes within that procurement event. So, if and when demand does materialize, the shipper's going to those more granularly-defined lanes and not using those more aggregated geographic area lanes as well. So, really, what's happening here is duplication of lanes, and those higher aggregated demand volumes don't end up being used as often, so they end up as ghost lanes. David Correll: Gosh, interesting. And I don't mean to cross the streams here, but to make another Ghostbusters reference, the classic line, "Who you going to call?" So, we know what the problem is and we know it's a big one. It's a big marshmallow man coming for our city of consumers. How do we solve it? What does your research point towards potential solutions in this space? Angela Acocella: Yeah. Well, to answer your question, who you're going to call, I think you should call your local freight transportation researcher and ask them for some help on what to do. But in all seriousness, if there's one thing that we can do, I think, we identified what types of lanes end up as ghost lanes. Those new lanes. I suggest removing those from the procurement event, right? I said it was 85% of ghost lanes are new that year. If you flip that statistic around, 95% of new lanes in a given year end up as ghost lanes. So, that means only 5% of new lanes actually materialize. So, I'm saying let's throw all of those new lanes out of the procurement event, let's either procure them later on in a mini bid if and when that volume does materialize, or just cover them on spot if they end up being low volume. But then we asked, "Okay, well, if we do that, what's the effect? How effective is this coverage strategy anyway, in covering these 5% of lanes that do actually materialize?" It turns out they're not performing well anyway. So, this coverage approach isn't really covering those 5% of new lanes that materialize well. So, we looked at carrier acceptance rates of those 5% of new lanes that materialize. They end up being about 73%. So, meaning 73% of loads that are tendered to or offered to our contracted carriers end up being accepted. And so, a large percent end up being rejected. And often, shippers are stating in their service level expectations, "We want somewhere around 95, 99% acceptance." So 73 is way below what we want. And as a result, shippers then end up paying for higher priced backup or spot options when carriers are rejecting their loads. So we looked at, on top of that, so we know carriers are rejecting loads on these new lanes. Well, of the carriers that are accepting them, what was the contract price for those lanes? Do we at least have something competitive in place? And we actually found that no. In fact, the contract prices for these new lanes are far too high. In fact, they're 13 to 40% higher than spot prices for those lanes at the time those loads are materializing. So, we're way overpaying for those loads that materialize on new lanes anyway. So, as I said, I suggest taking those new lanes out of the procurement process, but every shipper's business is going to be a little bit different. So, as I was saying, I think we need to take a better look at what happened last year in the procurement event. Each shipper, what lanes did and did not materialize? Why? How did we make decisions about how we defined our lanes? Things like that. And then, use that very strategically about what lanes should we actually include in the procurement event this year, and what can we carve out, and maybe get rid of some of the excess that we didn't need? And maybe even more importantly, as I said, the size, the cost, the time of the procurement event is driven by the number of lanes procured. But by freeing up a lot of that time from the lanes that are not important, we can actually be focusing our attention on procuring the important lanes, the higher value, higher volume lanes, and maybe make sure we have better agreements with our carriers. We've spent more time on those important lanes, rather than lanes that are very unlikely to materialize. David Correll: So, keeping with the theme, and maybe stretching it a little far, but thinking about classic monsters and movie monsters, in some ways, all monsters are just exaggerations of the things that we as humans fear in ourselves. So, werewolves are exaggerations of our basic human animal nature. Vampires can be seen as send-ups of our tendencies for vanity and avarice and cheating death. Mummies, you can think about pointing out our fear of our own mortality. Which is a long way to build up to asking you, do you think there's an undesirable human trait that ghost lanes, as a monster, exaggerate or point out to? Is there a dark side of humanity that the ghost lanes you discovered point out? Angela Acocella: I knew there would be a philosophical piece of this conversation, knowing you. No, I think that's a really nice kind of question to take a step back, and understanding, what is the human behavior that we're seeing in this behavior? And I think it comes from our preference for known costs now over unknown costs later on. So, we're downplaying the significance of guaranteed, upfront risk or cost. In this case, it's, as I'm saying, the administrative costs of running the procurement process. And then we're overemphasizing the unknown or the perceived or the potential cost later on, in this particular setting of searching for carriers and/or relying on the spot market. So, it's this perception of, "I prefer the devil I know versus the devil I don't know," right? I know I'm going to have these administrative costs now. I'm willing to incur those now, and that way, later on, I don't have to deal with it. That's one. I think the second one is also that we place value on flexibility in the face of uncertainty. So, this is a way to hedge our bets. We want to make sure that we have as many options open later on. Again, just in case. And we actually have seen this established, actually really well established, in management research. There's a really interesting controlled experiment where we actually saw that, I didn't do this, other very smart researchers did this, but they found that we're actually willing to knowingly overinvest upfront in order to keep the door open for flexible options later on. So, we're seeing that these behaviors that we're seeing from shippers is kind of a natural human behavior. We're willing to incur known upfront costs, overemphasize that known versus the unknown later on. And in this case, in the specific context of transportation services, it's those administrative costs, sifting through those bids, loading the carriers, the prices, all of that, into our transportation management systems, for every single potential lane, even if it may not materialize, just to avoid that unknown cost and that volatility potential later on. What hasn't been studied, and I think that's what was really interesting here is we place value on flexibility, but does it actually pay off, right? Are there costs to that flexibility? Are there costs to that tendency to rely on the known knowns versus the known unknowns, so to speak? And I think in our research, we got to explore that, no, there are costs, and I don't think it is worth it to have this coverage approach, to stick to the knowns that we know. David Correll: Gosh, that's really interesting. So, if I take that argument and sort of stretch it out over the horror movie argument, the ghost lane can be seen as a manifestation of shippers' own aversion to uncertainty. Angela Acocella: Yes, yes. David Correll: I have it too. I can't blame them. Angela Acocella: I know, we all do. David Correll: I understand. Gosh, thank you so much for all the great work that you've done contributing to the FreightLab during your time here, and where you're going to next, and thank you for spending the time with us today to share your most recent research with the broader community. Angela Acocella: It's always a pleasure to be here, and it's great to be back for just a little bit. Announcer: All right, everyone, thank you for listening. I hope you enjoyed this edition of MIT Supply Chain Frontiers. My name is Arthur Graw, Communications Officer for the center. I invite you to visit us anytime ctlmit.edu, or search for MIT Supply Chain Frontiers on your favorite listening platform. Special thanks to Paweł Feszczuk for the music, Something Scary Is Behind Us, on freemusic.org. Until next time, everyone.