View the December 19 seminar recording
Jennifer Symoun
Good afternoon or good morning to those of you to the West. Welcome to the Talking Freight Seminar Series. My name is Jennifer Symoun and I will moderate today's seminar. Today's topic is International Urban Freight (I-NUF) Conference Presentations, Part 2 – Assessing the Impacts of e-Commerce
Before I go any further, I do want to remind you to call into the teleconference for the best audio quality. If you are listening to the audio over the computer and experience any issues, I am unable to fix them as audio quality will vary based on your network connection, computer, speakers, and other factors. Please also keep in mind if you are calling into the teleconference for the audio, you will need to mute your computer speakers or else you will be hearing your audio over the computer as well.
Today's seminar will last 90 minutes, with 60 minutes allocated for the speakers, and the final 30 minutes for audience Question and Answer. If during the presentations you think of a question, you can type it into the chat area. Please make sure you send your question to "Everyone" and indicate which presenter your question is for. Presenters will be unable to answer your questions during their presentations, but I will start off the question and answer session with the questions typed into the chat box. We will also take questions over the phone if time allows and I will provide instructions on how to do so once we get to that point.
The PowerPoint presentations used during the seminar are available for download from the file download box in the lower right corner of your screen. The presentations will also be available online within the next few weeks, along with a recording and a transcript. I will send a link to the recording in the next day or so and will also notify all attendees once all materials are posted online.
Talking Freight seminars are eligible for 1.5 certification maintenance credits for AICP members. In order to obtain credit for today's seminar, you must have logged in with your first and last name or if you are attending with a group of people you must type your first and last name into the chat box.
PDH certificates are also available for Talking Freight seminars. To receive 1.5 PDH credits, you will need to fill out a form. Please see the link in the chat box. Certificates will be emailed one week after the seminar. A seminar agenda has been included in the file download box for those who need to submit an agenda to their licensing agency.
Finally, I encourage everyone to please also download the evaluation form from the file share box and submit this form to me after you have filled it out.
Today we'll have three presentations, given by:
Our first presentation will be given by Jean-Paul Rodrigue of Hofstra University. Dr. Rodrigue's research interests mainly cover the fields of transportation and economics as they relate to logistics and global freight distribution. He has authored 6 books, 32 book chapters, more than 50 peer reviewed papers, numerous reports, and delivered more than 150 conferences and seminar presentations. Dr. Rodrigue developed a widely used online reference source and textbook about transportation, the Geography of Transport Systems, now in its fifth edition. In 2019, he was the recipient of the Edward Ullman Award for outstanding contribution to the field of transport geography by the Association of American Geographers.
I am not sure if you might be on mute, we are not hearing you.
Are you there? Are any of the presenters hearing anything?
Presenters
I am not hearing anything.
Jennifer Symoun
We may have lost your line we are not hearing you Jean-Paul.
Jean-Paul we will give you another second or so or will switch over to another presenter. You might want to try hanging up and dialing back in.
We will move on to the next presentation and come back.
Monique, I'm going to introduce you and start with your presentation first. Monique Stinson is a Computational Transportation Scientist at Argonne National Laboratory and a PhD candidate at the University of Illinois at Chicago. As part of the Vehicle and Mobility Simulations group within Argonne's Energy Systems Division, she conducts fundamental research on the relationship between energy use and transportation in the U.S. Prior to working at Argonne, Monique worked at Cambridge Systematics, RS&H, and the Chicago Area Transportation Study. She is a member of the Transportation Research Board's Freight Planning and Logistics Committee and previously served on the Transportation Research Board's Bicycle Transportation Committee.
Monique Stinson
Thank you, Jennifer. Thank you, for having me, I'm excited to be here. Today I'll be presenting on the results of some e-commerce analysis we did looking at the impacts of e-commerce on regional travel and energy use, examining both impacts of household shopping as well as parcel delivery trips. My colleagues on this work were Annesha Enam, Amy Moore, and Joshua Auld.
Just to set the broad stage for the research agenda that we have been pursuing for the last few years, obviously there is a lot of disruption happening in the transportation system right now. There is electrification of vehicles, there is charging infrastructure questions, there is a lot of changes in goods movement, shared mobility, automation, and so on. This has a lot of impact on human behavior as well as on transportation and energy use. Today we're going to focus specifically on the impacts of e-commerce as a major disruption and the implications that it has had for regional travel and energy consumption.
We are all familiar with the traditional paradigm of people going out to the store to buy goods and coming back home or going back to work. We are all familiar with the paradigm we have seen arise in the past two decades with the rise of e-commerce, where now goods are being transported, not necessarily from the store to home by a person, but from a fulfillment center or depot to homes or businesses by a medium-duty truck.
So, the research question we have is that as these traditional physical shopping trips are being replaced by virtual shopping events, the e-commerce generated trip, what is the net effect on regional Vehicle-Miles Traveled (VMT) and Fuel Use or Total Energy Consumption? It's a very important question and has a lot of profound impact on regions, but to date we don't think anyone's really looked at this in comprehensive detail. So, our scope is focusing on the last-mile of delivery and we are looking specifically at Chicago Metropolitan Region where the area that we include covers 16 counties; a lot of urbanized as well as rural and suburban areas, so it kind of covers the gamut of urban geographies. The approach we use is modeling these shopping trips as well as delivery trips using Activity-Based Travel Demand modeling in conjunction with Dynamic Traffic Assignment. We have a tool that Argonne has been developing for about a decade now called Polaris. It's integrated activity-based with dynamic traffic assignments, and it's very detailed, so it allows us to explore the trade-off people make in their travel decisions. Decisions like where to work, what kind of cars to own, how often to use transit or nonmotorized modes, how to schedule trips, and so on. All of that information is sort of combined with this highly detailed Dynamic Traffic Assignment where congestion is done in a way where both passenger vehicles and trucks are modeled together, so we get a comprehensive evaluation of their impacts, their interactions with each other, and their impacts on VMT as well as energy use.
A lot of development in Polaris has been focused on the passenger side. For the last year or year and a half we have implemented a freight model in Polaris. I won't go into the details of the other features of this model, but it does include commodity flow information as well as regional trip information. And then it has a parcel delivery component and that's what I will spend the next few slides talking about, so you get a sense of the approach.
So, to do this exercise we use this model along with several scenarios we have come up with as part of the Systems and Modeling for Accelerated Research and Transportation, or SMART mobility program, that is run by the Department of Energy Vehicle Technology. Now specifically I'll talk about the methodology for assessing e-commerce impact.
Step one, we generate household delivery demand within the Polaris model. We basically develop a behavioral model that uses data from the Whole Traveler survey that was collected by our partners at the other national labs. We developed this model so then as we apply it in Polaris, we generate e-commerce demand for each household in the region. Step two is then coming up with parcel delivery supply to serve all of this demand. To do this, the demand points are basically set into the supply component and all of the demand is carved up into these small areas for delivery tours to serve. Also, as part of that, the traveling salesman problem algorithm is used to determine an efficient stop sequence for each set of tours.
Step three is routing these delivery trucks within Polaris. And so, this shows an example of a single tour that actually the model for this is based on data from the Columbus Ohio area where about 120 stops per tour are made, so there has been more recent data on this. It looks like some of the other parcel carriers now like Amazon is ramping up and shooting for 200 or 200+. This kind of gives you a sense of the efficiency of the delivery trips that come out of this modeling process. Then the fourth step is, as these tours are routed in Polaris with the Dynamic Traffic Assignment, we are then able to compute the vehicle-miles traveled, and also those results are integrated with SV Trip and Autonomy, some of our other in-house tools, in order to estimate the energy use associated with all these trips. I wanted to briefly mention, the Household E-commerce Demand Model is a behavioral model, so it's sensitive to household characterizes such as income level, number of children, number of vehicles. It's pretty consistent with what we observe of people that are generating e-commerce demand. They do tend to be higher income households with more children, reflecting busier parents, especially dual-worker households. Households more vehicles, they can stop at the store on the way home from work or resident households within a walkable area have less e-commerce demand. Polaris runs in 6 second time increments, so we want to take advantage of that level of temporal resolution, so we use a Temporal Disaggregation Algorithm and assume that parcel truck deliveries are uniformly distributed between 9:00 AM and 6:00 p.m. And then other trucks' trips are distributed throughout the day in accordance with some other data.
That is basically an overview of the approach. I want to spend a little bit of time discussing the assumptions then that go into the model scenarios, where we are actually running e-commerce scenarios and looking at what-if scenarios, at what would happen. We have three time frames: the baseline, short-term, and long-term, and you can think of it as roughly 2020, 2025, 2040. For e-commerce delivery rate, we did runs with the baseline rate for all of the time frames. And then we also looked at what if that rate increases. So, the baseline rate is one delivery per week per household, and this is consistent with NHTS and also the Whole Traveler survey. For the short term we looked at what would happen if that tripled; and in the long-term, what if it were five times as great. Another important aspect of this scenario run was the Vehicle & Powertrain Technologies. We have three levels here: one is the baseline technologies, so today's technologies; one step up from that is the BAU, or the business as usual, and that's a slight increase in electrification compared to the baseline; then the Vehicle Targets Office, the targets, is yet even higher in terms of electrification. Finally, the last column shows other important assumptions, especially long-term we look at two different scenarios. We know TNC use is increasing; that's the transportation network companies like Uber and Lyft and others. We also think there may be some private AVs (autonomous vehicles) on the horizon. We don't really know, so we wanted to test both. In one version of the long-term future, we look at a high TNC use with low private autonomous vehicles, and that's privately owned by households. And then we look at a variation where instead of that we look at low TNC use coupled with high private av use. So, that's important to keep in mind.
Now I will discuss some of the findings from this analysis. Just to show you the context, I want to make sure it's clear what travel markets we are focusing on. We are really focusing on, in the rest of the analysis I'm going to show, the medium-duty parcel delivery truck trips. And those we will use the abbreviation MDT. Secondly, the passenger shopping light duty vehicle trips. This pie chart shows the baseline vehicle-miles traveled for all travel segments. You can see the parcel MDT is actually only 400,000 miles of 300 million miles, so it's a very small amount in the base here. The passenger shopping on the other hand is maybe 30 or 40 times that amount, it's quite substantial. Passenger shopping is roughly 6 or 7% of total VMT. It's quite a lot in the metropolitan region. First, we will see what happens if moving into the future – short-term, long-term – if the e-commerce rate stays at the current day level, which is one delivery per household per week, and what we see is very slight increase maybe 10 or 20% is commensurate with population growth anticipated for the region. On the left it shows the VMT estimate and on the right, it shows the energy use estimate. Also, please note that the blue in all of these bar charts is a medium duty truck delivery estimate; you can see here it very small in the base here. Then the orange is the light duty vehicle shopping estimate.
Now, keeping that same delivery rate one per week per household, we want to see what is the impact of changing vehicle technology. So, the left actually in the kind of faded color is the energy use that just from the previous slide the slight increase due to population growth. And then on the right, it's the same amount of e-commerce, but just with these improved vehicle technologies. We have the Business As Usual and VTO shown for short-term and long-term. So, comparing to the base year, the vehicle technologies alone would generate anywhere from 10 to 40% energy savings. It's pretty significant and something to keep in mind with the rest of the results I show you.
Now we are going to be looking at the impacts of increasing e-commerce based on that tripling and quintupling I mentioned earlier. Basically, that VMT grows by 300 to 500% for the parcel MDT, which is what you would expect given the increase in e-commerce. What we find is that actually, instead of a total increase in looking at all retail VMT, we don't see an increase; there is actually a decrease compared to the baseline. This decrease is anywhere from 32% up to 50% for the future years. For the long-term, there is some difference and that's due mainly to whether the high private AV is sort of the paradigm for the future or whether high TNC ends up being predominant. That's kind of driven by an assumption that it's widely discussed in our circles. So, there is a big question; if people are using AV's probably the value of travel time is going to go down because they are not taking on the burden of driving. In general, there is a belief that trips will be a bit longer, and so that's really what we see reflected in here.
So, this is the VMT. There is a slight difference looking with and without the vehicle technology improvements. So, 34 to 56% versus 32 to 50%. The ballpark is 1/3 to about a half savings for the future. Likewise, energy use is also significantly reduced as the levels of e-commerce increases. Looking at the comparison to the base here, we see roughly 40 to 50% savings for the short-term if e-commerce triples, and up to say 72% if it grows by five times for the future years. Again, here there are some notable differences due to the assumptions around private AV versus use of high TNC. So, energy is where we really need to account for vehicle technology improvements. Because this 52-72% savings here over the baseline, a lot of that is due to vehicle technology improvements and the rest of that is due to e-commerce. So, after we account for the vehicle technology, there is still a huge savings on energy side.
Finally, I wanted to mention the results from a corner case we did where we studied the complete replacement of household shopping trips with e-commerce deliveries. And this is from an earlier version of the model, so some of the base numbers are a bit different, but the impacts are the same in terms of directionality and magnitude. The maximum VMT savings in this case is 80% and the maximum energy savings is about 50 to 60%. So, again, the e-commerce delivery system does by all accounts seem to be more efficient than the household shopping paradigm. It also highlights that there is some room to improve truck efficiency, so that remains to be seen; whether electrification will be affordable, what the range of batteries will be and so on.
Why does this happen? Fundamentally, it's because all the shopping trips that are being replaced are on average between 7 and 8 miles, according to the 2017 National Household Travel Survey. If you think about it, one additional stop on a delivery tour is just a small amount in comparison to that 7-8 miles roughly, a half a mile for each additional delivery trip. So, that's really driving these results.
Although e-commerce currently is generating a lot more delivery truck traffic and if e-commerce continues to increase, we will still see a lot more delivery traffic. But overall, given that these delivery trucks are replacing these inefficient shopping trips, there is an overall net reduction in VMT and energy use. And it's important to keep in mind that the results differ somewhat; whether private AV becomes the paradigm or TNC, so on, as well as the market penetration of electrification technologies. So, if you want more information, this is a lot of detail; we do have a paper published and some conference proceedings from earlier this year, as well as a couple of capstone reports that DOE is working on and they're in progress, but my understanding is that they will be released in the next few months. And an acknowledgment to our sponsor, the DOE (Department of Energy). That's all, thank you.
Jennifer Symoun
Thank you, Monique. I know we have a number of questions in there. Please continue typing them and we will get to them at the very end. We are now going to go back to Jean-Paul Rodrigue of Hofstra University. Hopefully the phone issue is all clear now. You can go ahead.
Jean-Paul Rodrigue
Good afternoon, everybody, from very chilly New York this morning. I am going to talk today about an overview of an Emerging Geography of Freight Distribution, which is impacting retail and I'm going to look more specifically at Amazon in terms of the location distribution centers, as well as emerging specialization of the logistical function that Amazon is doing, which is very impressive. To move on with this, from my perspective I put the impact of e-commerce on freight distribution into 4 categories. I'm going to cover 3 in this presentation. The first category is the change in distribution pattern; the growth of home delivery and the change in last-mile logistics, which was very well covered by Monique. Another consequence which is much less discussed, is essentially if you want to define e-commerce. in my perspective, it's a switch of the inventory from the retail store to the warehouse. That's essentially what it is; it's a change in the real estate footprint of all of this. Overall when e-commerce would say it penetrates the market you see radical changes in the footprint of different facilities and it's very clear that retail across the United States is suffering, not because the sales are going down, but because the footprint is shifting from the stores to distribution centers. The third aspect, which is very interesting and it's going to be big focus today, is the emergence of all these new logistical facilities. It's amazing the names we come up with and their function. These facilities did not exist 5 or 10 years ago, and I'm going to give you an overview of what is emerging these days. Last, but not least is the process of vertical integration, mostly I would say spearheaded by Amazon, which is taking control of the distribution process, taking control of physical access and of course logistical assets regarding distribution. From what I understand, Amazon is now claiming that half of all of its parcel deliveries are now internal to Amazon. It remains to be seen if that it's entirely true or not, but that's, in a sense, the story.
The next slide is simply to show you the change in retail logistics, focusing on the phase of digitalization, which is the emergence of new facilities. The best way to talk about that is to tell you what I have been doing, which is in a sense taking the data provided by a private consulting firm which has been collecting this information, and I simply took the addresses and geocoded the addresses and put a location virtually on the ceiling every distribution facility owned by, let's say Amazon in the US, and cross checking the data in terms of the square footage, of function of the facility. And overall this point it's changing very rapidly where we have around 441 facilities in the United States, which is quite impressive. And I was able to identify 6 specific functions. Some of these locations are also co-located, which is kind of interesting by itself.
So, moving forward, this represents my take in a sense of the facilities in the distribution channel that Amazon is basically creating. In the past, not so long ago, it was put forward to two channels, with the classic Postal Service and the UPS and FedEx, which were the classic 3PL in parcel deliveries. Now, what seems to be happening is Amazon is taking control of the procurement/fulfillment process. But also now is creating sortation centers, delivery stations, and now for the last year and a half or two, the new air hub facilities. This is ongoing and is very impressive as well. Plus, of course, the fast delivery hub which I will call Amazon Prime. All of this is essentially internalizing the parcel delivery process. I will give you an overview of all these facilities in the United States. And at this point a map of all the facilities, and this is something like 169 million square feet which is about 6 square miles, just to give you an overview of the footprint. So, Amazon is right now 6 square miles of warehousing space in United States. This is color coded and I will give you an overview of each and every one of these in detail in the next few minutes, so let's move on.
In terms of the growth of this footprint, which at this point is exponential, I would say very impressive/ And I was able to suggest that there are 4 stages in this emergence. The first step is, of course, the early years were Amazon was doing its niche testing what was the potential of e-commerce. It was pretty slow with 3 to 5 distribution centers covering the whole United States. Then, as the e-commerce became more accepted, the footprint of Amazon started to fully emerge and to become a full-fledged e-commerce retail platform. Which leads to the third step where Amazon gets used massively but starts to feel problems in terms of the ability of its distribution and starts to open new facilities which are sortation centers and eventually delivery stations, as well as inbound cross dock facilities. And now we are in the fourth step, which is again very quick, which is a full process of vertical integration were Amazon is taking control of the whole supply chain, from the procurement all the way down to the door. Of course, by doing that, Amazon is having a substantial amount of footprint, which essentially this is the distribution of this footprint in terms of the average facility size. There's a remarkably constant size in the facility that was very impressive in terms of the footprint they take. It means that Amazon gets a lot of fulfillment centers, and when it needs to expand it does not increase the fulfillment center, it replicates it, it creates another one approximately the same size. It seems to be quite interesting in terms of what it's doing.
So, let's start off with the first facility at the beginning of the procurement process, which is called an inbound cross dock facility. This is essentially a warehouse that receives containers, or FTL/LTL loads on one side. And the purpose of this facility is to create deliveries for the e-fulfillment facilities. This is essentially where all the material inputs of Amazon go in. As far as I can tell there are no deliveries taking place to the e-fulfillment center from third party suppliers. I could be corrected, but most of the inbound logistics goes to the cross docking facilities. The next map shows the location of those 10 facilities in the United States. The pattern is relatively simple. It's two inbound maritime gateways: Los Angeles-Long Beach, New York-New Jersey. The other facilities are located nearby major facilities in the United States. The purpose is basically to supply the whole network of Amazon according to these facilities.
The next facility, which is the one we know the best, is the e-fulfillment center. Average size something like 800,000 square feet. It is a cross dock facility on a single side. Most of the orders are processed here. What is very impressive is there is a specialization in terms of what is being shipped. And there are 3 major categories: small sortable, large sortable, and large non-sortable, which is rated to the size of the parcel itself. And everything which is small is basically what fits in a small box. Again, the large is the same thing, which fits with a box which can be put to a parcel carrier. And what is non-sortable is when the shipment is the box itself, such as the televisions, the printers, and all the other large orders. When you look at the location of the facility it is a market coverage strategy, it's simple. If you live in Wyoming, North Dakota, you might be a little bit sad as there is no Amazon distribution e-commerce facility. It's a territorial coverage strategy with about 180 facilities approximately. This is essentially what their purpose is in terms of the number of facilities, the function. It was impressive that the largest number of facilities are what we call the non-sortable because they are bulky, which means that they are more expensive to ship, they're more difficult to handle, therefore Amazon has created a large number of these facilities in a wide area of market to minimize the distribution cost. You will notice the last facility is called returns. There are only two facilities in the United States to process returns, which is a pretty smart strategy. By this I mean every return is run through a cost-benefit analysis; if it costs too much to return, most of the time Amazon says you don't need to return the product back. If it's brought back, it's basically a slow route and takes a long time to process. They don't care too much about this, so they only use two facilities in the United States to take care of that.
The next facility is a sortation center. Again, a smaller cross dock facility mostly located in the periphery of a metropolitan area. Basically, that is Amazon taking control of the post office, or what UPS and FedEx are doing, that is to sort parcel by ZIP Code. So, all the traffic shipped to metropolitan area is likely to enter this sortation facility and the parcel loads are going to be broken down into ZIP Code and sent to either a third party delivery company or to a delivery station. If you look at the distribution of these facilities in the US, 46, and again all the major metropolitan areas and the periphery are serviced by these centers.
The next which is pretty much the Amazon Prime strategy, which is an independent facility, Fast Delivery Hub. They're quite small, I was surprised, I was expecting them to be bigger. But they are designed to observe Amazon customer behavior and is capable of knowing, in advance, the amount of what people order on average per day within a specific area within a ZIP Code and is able to pre-position orders which are in high demand in these facilities. Therefore, from these facilities, Amazon is able to respond within 48 to 24 hours on the demand by being within the central areas. It's an interesting facility, which is the intersection of the fulfillment center, but also a facility that is designed to cope with what we call last-mile logistics. Which brings us to their geography; again, pretty much in all major market areas. I found that 4 of them are co-located with fulfillment centers, but they are separate. The other fulfillment centers, if there is a fast delivery hub, they are a separate part of the warehouse.
Moving on with something that is brand-new and very difficult at this point because it's emerging as we speak, that now Amazon has entered the air parcel delivery service directly competing with FedEx, UPS, and DHL. In some ways, most of these facilities are relatively small at this point, co-located next to a major airport. This map of them. You have 5 facilities at this point, but they are coming fast. Over the next few months there will be a new one coming up, within the next couple after that yet 5 or 10 more will come online, but all the squares are the air hub facilities served by Amazon within the United States to third-parties. But overall Amazon is moving in within the air system quite rapidly. All of this to move high value goods in the fastest way possible.
This brings us to the delivery station, which is close to the last step. These are cross docking facilities mostly coming from sortation centers. And at this point what happens is that the cargo is brought into individual delivery vans for a longer reach, so that facility is essentially the last-mile average size, somewhat metropolitan area, and at this point it's where the roots are either to Amazon vans or to third-party players. I forgot to mention earlier that on from the sortation center, Amazon might decide as well to ship the parcel to the post office instead of to the delivery station. Which means often when a customer receives a parcel delivered by USPS it could be the last-mile. That is, the whole process could have been internalized through Amazon facilities, but for the last-mile through the sortation center, the parcel was delivered to a local post office for the last-mile delivery. A lot of those facilities are relatively small size; most average 75-80 thousand square foot type of facilities.
And last, the local freight station, the locker, which Nico will discuss a little more in detail. Just to mention that Amazon has built a lot of those to offer a customer opportunities to pick up parcel. I did not do an inventory of those; there are too many and it's difficult. I see 3,000, 4,000, 5,000 of those locker facilities within the United States and that's the very last mile in city logistics.
What I could say based upon these observations is that Amazon is creating a very clear emerging geography which specializes in facilities. Each of them for filling some kind of function; there is a functional specialization taking place in terms of what Amazon is designing, and all of these facilities did not exist just 5-10 years ago. Amazon in e-commerce is forcing an entirely new form of logistics requiring very specialized facilities to fulfill. Overall there is three layers for logistics: sortation center, delivery station, local freight station; that is the first observation. The other observation is there is a horizontal and vertical integration process taking place. That is, Amazon is spreading horizontally by replicating facilities to more efficiently cover markets, because it wants to minimize delivery time and there are many reasons behind minimizing delivery times. Of course, it's customer satisfaction, but also it wants to replicate the consumer impulse buying process. That is, a consumer buys something and it wants to ship as fast as possible to avoid buyer's remorse. That is also something interesting that needs to be researched, that's a driver behind this. As I was mentioning, there's also a vertical integration process taking place, where the whole element of the supply chain is now being internalized through different sorting facilities all the way down to the delivery trucks. In a sample I'm doing in New York, now Amazon accounts for 25 to 30% of all parcel deliveries are now taking place through Amazon trucks. That's pretty much about it. If you want more information you can have a look at the Geography of Transportation Systems, or the online book we are currently developing as part of a research consortium I am part of, which is called Global City Logistics; we're going to unlock a full online textbook about urban freight distribution and e-commerce is going to be a very big part of this material. Thank you very much.
Jennifer Symoun
Thank you, Jean-Paul. I encourage everyone to continue typing questions. We are going to move on to our final presentation, which will be given by Nico Boyd, a Los Angeles-based transportation planner at Fehr & Peers. He is particularly interested in understanding how to maximize the efficiency of freight movement in urban contexts while balancing the needs of all transportation system users. Prior to joining Fehr & Peers, Nico completed master's degrees in City and Regional Planning and Civil Engineering at the Georgia Institute of Technology.
Nico Boyd
Thank you. Thank you to Jean-Paul and Monique for your presentations. My name is Nico Boyd I am a transportation planner at Fehr and Peers. We are a full service transportation planning and engineering consulting firm with offices around the country, but as was mentioned I am based in Los Angeles. The title of my presentation today is "How Do Smart Growth Cities Take on Rocket Increase of E-commerce? Assessing Effective Strategies for California cities." I will start with some background on the origins of this presentation, I'll get into some of the challenges of urban growth, I will talk about e-commerce and retail trends we are seeing, and then I'll get into smart growth: what it is and why it is pertinent to this presentation. I will talk about California Sustainable Growth Initiatives, specifically state-level policies that have an impact on smart growth and e-commerce. I will talk about challenges for urban goods movement and highlight best practice case studies we can learn from here in California. And then I will end with future considerations.
So, a little bit of background. Fehr and Peers has been helping Caltrans update the 2019 California Freight Mobility Plan. As part of the update we were tasked with preparing a white paper assessing the interconnections between smart growth and urban goods movement, specifically with the goal of looking for ways to align the objectives of each. Really it was our research into that relationship that spawned this presentation.
According to the World Bank, 55% of the global population lives in urban areas today. And it's estimated that by 2050, 68% of the global population is going to be urban. We are seeing more commercial activity than ever before and in the U.S. our cities have historically been designed around the car, incentivizing driving over other modes of transportation and driving greenhouse gas emissions and increasing the amount of impervious surface in cities, which has a direct impact on water quality, and consuming large quantities of open space. So, as cities continue to grow, if we are not careful to moderate increases in vehicle travel, we are going to continue to see the same trends play out into the future.
Amidst the challenges of the urban growth that I just described, we are also seeing explosive growth in e-commerce. Younger generations are increasingly doing more and more of their shopping online for a wide variety of goods and services. In 2018, we saw a 16% growth in e-commerce. And in the graph here on the right, you can see how quickly e-commerce has grown as a percentage of total retail, even overtaking total retail growth as recently as 2016. As part of our update to the Freight Mobility Plan, we administered a targeted limited public outreach survey online via Facebook and in person via a mobile app. And we found that 60% of our respondents receive 1 to 2 online deliveries a month, 35% receive 3 to 5, and 3% receive more than 5 online deliveries every month. At the same time that e-commerce is exploding, we are also seeing large decreases in traditional retail. In 2018 we saw an 87% decrease in new retail spaces opening in the 54 largest U.S. markets. The long-term implications of that for cities are not totally certain yet, but we do know that there are going to be impacts on things like sales tax revenue, land use, and the provision of infrastructure as cities and industries start navigating how to repurpose retail space for new uses. In California, we are seeing new fulfillment and distribution centers emerging outside of major urban centers for online retailers like Amazon, and also big-box retailers like Target and Walmart. It's very clear these companies are doubling down on the growth of e-commerce with major investment in infrastructure to support their distribution chains. Zeroing in on Amazon in particular, Amazon has established 9 new hubs that cluster around major urban centers in California to capitalize specifically on the proximity to airports and seaports. The table on the right highlights projected growth in total cargo tonnage at California's major airports and as you can see LAX is projected to grow from almost 1.7 million tons in 2011 to over 3 million tons in 2040. This is really sizable growth and it definitely underscores the need for proactive planning to accommodate the increase we are going to see in goods moving activity.
So, what is smart growth? Smart growth is essentially an urban planning philosophy whose goal is to reduce automobile dependency and improve the livability of cities. The focus is really on increasing urban density and providing a mix of different land uses within close proximity to one another so that people can realistically make the choice to get around using public transportation, they can get around by walking, biking, but are not necessarily being forced into a car for all of their trips. In California, as I mentioned before, we have a number of statewide policy initiatives that have an impact on smart growth. The first on the left is Assembly Bill 32 which kind of sets the stage for the other two pieces of legislation that I'm calling out here. AB 32 essentially required the state to reduce greenhouse gas emissions to 1990 levels by 2020, which was actually achieved in 2018. But AB 32 gave the California Air Resources Board the authority to regulate sources of greenhouse gas emissions, including those that come from the transportation sector. Moving on to Senate Bill 375, this one empowered the California Air Resources Board to set regional targets for greenhouse gas emission reduction, and it further empowered cities and counties to play a role in achieving those regionally set targets. So, as part of the regional transportation planning process that regions are already undergoing periodically every 5 to 10 years, regions are now required to develop what's called a Sustainable Communities Plan, which is essentially detailing the strategies that they are going to use to achieve their greenhouse gas reduction targets. Then lastly, Senate Bill 743 has a direct impact on how we manage congestion in California and transportation impacts. The previous metric for transportation impact analysis was level of service, which as many of you know intersections are essentially given a level of service letter grade indicating the level of congestion and delay, and mitigations then focus on reducing congestion which in some cases involves adding lanes or other things that ultimately have the act of incentivizing more driving. But the new metric we are switching to is Vehicle-Miles Traveled and our mitigations are now going to focus on reducing VMT instead of facilitating its growth.
So, the policy landscape I just described creates challenges for the urban good movement and the growth of e-commerce. These policies are largely directed at reducing automobile trips, but the explosive growth of e-commerce is by some estimates increasing the number or at the very least the frequency of deliveries, especially considering retailers like Amazon are able to provide next day or even same-day shipping and delivery. And then, at the same time that delivery activity is increasing, we are also seeing increases in personal transportation through transportation network companies like Uber and Lyft. The combination of these and other factors as well is leading to increased competition for curb space, which is something that's happening nationwide not just in California. But one thing we are noticing that is something we want to start paying attention to is that the increase in delivery activity does pose potential safety challenges for active transportation users. Delivery trucks often have to, for example, double park in a bike lane to make their deliveries, occasionally they have to block sidewalks, so this is another challenging issue we need to pay attention to.
Next, I am going to highlight some best practices from around the world that we identified, specifically strategies that we felt could be successfully implemented in California. And I will call out as I go along some of these strategies are perhaps more suitable for residential delivery, some are most suitable for commercial, and then some kind of straddle the fence, they could be suitable for both.
Our first strategy is Urban Consolidation Centers. These are essentially large warehouses on the outskirts of the city that are paired with smaller decentralized warehouses within a city. Heavy commercial vehicles representing potentially multiple carriers will bring in goods to the large warehouses on the outskirts, and then from there, medium and light commercial vehicles will consolidate the cargo and transfer goods into the city either to decentralized warehouses or in some cases directly to consumers. This helps to reduce the congestion that can be caused by heavy vehicles entering the city to make deliveries. Our example here comes from Padova, Italy. In 2004, they developed their Cityporto logistics scheme, and essentially, it's an urban consolidation center located about two miles outside of the city center. It has close access to major highways and they also have their own fleet of environmentally friendly delivery vehicles. So, they are able to realize some emissions reductions by using fewer, smaller vehicles and that couples with the fact they are environmentally friendly. A couple things to note, participation by carriers in this program is totally voluntary. But for carriers that do choose to participate, there are some benefits. For example, those who use the Cityporto vehicles are allowed 24 hour entry into the city, whereas carriers that are independent might not have that 24 hour access. Cityporto vehicles have permission to use the bus lanes to make deliveries, which is a timesaver. And then they also get dedicated loading bays at the final delivery destination.
Moving on to Dynamic Routing and Intelligent Transportation Systems. These are all about providing real-time information to travelers to allow them to make informed choices about how they are going to reach their destination. We know that drivers making local deliveries typically have different information needs than local thru traffic. For example, delivery drivers benefit from knowing if there are weight or height restrictions along their routes, as well as things like loading zone availability that might impact their delivery timing. By some estimates, nonrecurring causes of congestion, things like roadwork, traffic incidents, and poor weather, can cause as much is 60% of congestion. But using real-time traffic data and information on the road network and land-use in the area, delivery vehicles can be routed in real-time to avoid congested routes or routes that are otherwise unsuitable for freight movement. Our two examples here, the first comes from Detroit where dynamically routing trucks has yielded as much as a 45% reduction in drive time compared to historical congestion data. And then in Vienna, their ILOS project has resulted in a 60% reduction in travel time, 15% reduction in average distance traveled, 20% reduction in fuel use, and 30% reduction in cost as a result of switching to dynamic routing. And I'll note that this strategy could work equally well for commercial or residential delivery, as could urban consolidation centers.
Our next strategy: Neighborhood Package Pickup Points and Automated Parcel Systems. These exist in a variety of forms in different cities and contexts; I'm sure many of you have seen them. The concept is essentially the same regardless of geography. Carriers will make deliveries to a centralized hub where customers can then come retrieve their packages. This allows carriers to make one delivery trip instead of several, which has a congestion and emission-reducing effect. In the U.S., I'm sure many of us have seen Amazon's lockers. A lot of them are based neighborhood-wise inside of Whole Foods stores and they've also started partnering with apartment management companies to provide apartment-based lockers. These lockers can accept deliveries from multiple carriers, not just Amazon. Delivery drivers spend less time parked outside or idling while deliveries are made door-to-door, because they no longer have to go door to door, they can make all apartment building deliveries to a centralized hub within the building where tenants can retrieve them with typically a personalized access code that gets sent to their smart phones. And depending on the size of packages, this can also work for package return as well. This strategy is clearly mostly suitable for residential types of deliveries.
Our last strategy is Off-Peak or Overnight Deliveries, and this one really functions as a complement to other strategies I've discussed, because by incentivizing deliveries for large retailers and grocers during the off-peak hours, congestion can be reduced during the peak hours when residential or other types of deliveries are taking place. Our first example comes from New York City, specifically Manhattan. Freight deliveries in Manhattan exceed 100,000 a day. There is significant competition between trucks and other road users for a very finite amount of road space. From 2009 to 2010, the city decided to try to do something about that and they launched a pilot program where they basically had 25 receivers and eight carriers switch to distribution and receiving activities during off-peak hours, which they define as 7:00 p.m. to 6:00 a.m., for at least one month. As a result of the pilot, trucks spent an average of 30 minutes at the curbside making deliveries compared to 100 minutes before the pilot. This pilot program was so successful, they have since expanded it to include more than 400 shippers and receivers and is no longer a pilot program it's a totally bona fide city program. One thing I should note that's important is that many of the businesses that participated in the pilot program had to pay staff to be present at night to receive deliveries, and that was subsidized during the pilot program. But once that subsidy ended, some of the businesses decided it just wasn't cost-effective to continue paying staff to be available overnight to receive deliveries, so in those cases they reverted to daytime deliveries.
Our last case study comes from Barcelona where the city partnered with two supermarket operators to switch to a system for quiet nighttime delivery. This was done using vehicles that are typically quieter than vehicles that might be making deliveries during the day. Through this pilot program, the city was able to realize greater delivery efficiency at these grocery stores. They were actually able to replace 7 daytime deliveries with 2 deliveries by larger, quieter vehicles outside of peak hours. Delivery times were reduced, and lower transportation operating costs were achieved. This was accompanied by a reduction in congestion and greenhouse gas emissions. Following the initial pilot program success, one of the two supermarket operators was so pleased with how it went that they actually rolled the program out to all their stores nationwide where local regulations permitted that to happen.
Lastly, future considerations. One of the big unknowns in the realm of transportation right now is how autonomous vehicles are going to transform the transportation landscape. And freight is certainly going to be affected by this. By some estimates AVs are going to reduce congestion and improve travel time, which would be extremely beneficial for goods movement. However, in the near term while AV's are being deployed, there is going to be a mixed fleet of AVs and human-operated vehicles for some time, and the coexistence of these 2 types of vehicles might actually increase congestion and slow down goods delivery in the short term. So, that's something we want to be cognizant of. Another factor for consideration is the role of urban planning in urban goods movement. Urban planners have historically not taken the needs of goods movers into account in their planning processes in this country, and that's very much to our detriment, because goods movement is fundamentally our economic backbone. So, moving forward, urban planners would do well to take the needs of urban goods movers into account. Lastly, future proofing of our infrastructure. We need to make sure as we attempt to solve the problems of today that we are being mindful of our future needs as well. Problems that exist today might not necessarily be problems in the future, so we need to take steps to ensure that our proposed infrastructure is still relevant down the road so that we are not wasting investments we are making today. That's it for my presentation. I think at this point all of us are accepting questions.
Jennifer Symoun
Yes. Thank you, Nico. I will start by going through the questions that have been typed into the chat box. I encourage everybody to continue typing in questions and we will get to them in order they came in. I'm going to go back to the top and start off with some questions for Monique.
A person was interested in the specific energy intensity of any typical parcel delivery, over the last-mile, in BTU/ton-mile.
Monique Stinson
As far as BTU, there is a way to convert from kilowatt hours to BTU. It can definitely be done; divide the total energy by the tons and the miles. We haven't tabulated it that way, but I think the results would be interesting to look at and I think it would be consistent with what we've found so far.
Jennifer Symoun
They are also interested in specific cost of last-mile delivery in $/ton-mile.
Monique Stinson
We do have some data on the types of goods people have purchased through e-commerce. In theory we could come up with something. So, I guess that's one direction we could take that cost; the value of the goods that are being moved. Another direction we could take it is the cost of transportation. Just like the previous question, it could be handled by relating the energy cost of the trip to the cost of fueling plus the cost of electrification from any plug-in activity that's going on. It's doable. I can't say we've looked at it, but that could be interesting to look at in the future.
Jennifer Symoun
Another question for you. Did you compare your explicit truck routing to the results you would get from using Daganzo's method for estimating the delivery route distances?
Monique Stinson
No, we have actual data from one of the major parcel carriers that by some estimates carries maybe between 30 and 50%. It's a wide estimate, but they carry a significant portion of e-commerce deliveries. That part of the approach is completely data driven. I am quite familiar with that method. I do like it. It's something that could be used in the model moving forward, but we didn't do any comparison of that. That's more sketch planning. That's kind of what do you do if you'd don't have a method already or you don't have the data? Daganzo's method is a great fill-in. Since we have really good data, we just use that.
Jennifer Symoun
Another question for you. Is this assuming e-commerce is replacing personal shopping trips? I have heard that personal shopping trips haven't so far been reduced by e-commerce deliveries, as people still shop for groceries, recreation, and to try things on.
Monique Stinson
That's a great question. This aspect as well is completely data driven. This has been another big research question. Are people still going out shopping? Because we know that kind of activity is still going on. People go out shopping with family and they kind of want to walk around and it's fun. A lot of people do that and try things on. Then also, the grocery industry is trying desperately to break into e-commerce, but that's pretty much the lowest retail segment. That being said, that kind of stuff is still going on and grocery is still very low, but for these other aspects it's not a complete elimination of physical shopping trips. It is a replacement of some shopping trips. That is what we've found and the data we got from the Whole Traveler survey also backed that up. There was a question asking people how many deliveries they are receiving and whether or not those deliveries are basically replacing a trip. And there is approximately a 1 to 1 replacement, so that's what we used.
Jennifer Symoun
Is this covering only parcel deliveries, and can you define parcel delivery?
Monique Stinson
This is only covering parcel deliveries. I go by the definition that Commodity Flow Survey and Freight Analysis Framework use, which is 150 pounds or less. In terms of how we use that in the analysis, we do assume that major shopping, such as for appliances or cars, we assume that's still done in person. Certainly that's also increasing in the last few years, but a lot of that is still being done in person.
Jennifer Symoun
We'll now move onto some questions for Jean-Paul Can comparisons be made without determination of specific cost and energy intensities for each different mode of last mile distribution?
Jean-Paul Rodrigue
I am a little bit confused by this question. I have not looked into this. I have looked at the location of the facility and the commercial strategies behind this. I have not plotted the delivery routes and the average distance involved and the average number of parcels per delivery, so I do not have this information. But on the other side, what I know based upon the samples I have made, that e-commerce deliveries are growing at about the rate of 15% per year. We are currently in the worst possible world in terms of you have more truck delivery and we are in a phase of transition, so people have not completely adapted shopping behavior. There is a lot of shopping at the same time. I don't know how far this transition is going to go. But, again, I don't have specifics about these. Again, I don't have specifics about drones as well, it's not something I have looked into; I don't have that data at this point in time, but it's sure going to be some kind of niche. To what extent e-commerce is more sustainable than the standard retail pattern. Based upon the data that Monique has presented, it looks like it quite a lot of gains to be made and when we're going to get a level of transition between retail based shopping and home deliveries. This remains to be monitored and fully measured.
Jennifer Symoun
How can drones be more efficient than ground transportation?
Jean-Paul Rodrigue
I have no idea. I've not covered this topic, so I cannot tell you. I can only speculate there is a niche. What I've seen is that companies are experimenting with this technology. They are going to run the numbers and it will be for a very short, high-intensity deliveries. We will find out. Is it going to be good for food deliveries? It can be an interesting niche as well, but I cannot answer that question with any form of authority.
Jennifer Symoun
Is the MWPVL facility data you used available, or is it proprietary?
Jean-Paul Rodrigue
It's proprietary, but they make it available online to display. So, you can have a look at it. But they're the ones who compiled. I only presented the data here in aggregate fashion. I have to be a bit careful about this. But you can have access by talking with them if you want to purchase it, it's not a big deal. But I have only been using it at the aggregate level.
Jennifer Symoun
Nico, some questions for you. considering that the 900k elevators in the US are "self-driving vehicles", wouldn't any sort of 'grade-separated' freight or parcel autonomous system offer the same advantages that elevators enjoy?
Nico Boyd
Yeah, absolutely. Grade separation is definitely a goal that we strive for, but implementing it unfortunately can be very costly on transportation networks. So, where the funding is available to develop grade separation, I think it's very successful and that's definitely something to strive for, but it's just not always doable in terms of cost in all contexts and circumstances.
Jennifer Symoun
Is Detroit's dynamic routing a city-created platform, or leveraging a private/existing platform?
Nico Boyd
So, I want to say that the impetus to create the program came from the city. The actual implementation of it is likely leveraging a public-private partnership. I'm not totally clear on the details about that. I can look into that further and send those to anybody who's interested in learning more.
Jennifer Symoun
We have a question I will put this out to any of the presenters. Has anyone studied specific energy intensity of tap water vs. bottled water?
[Silence]
Most likely not. I think we have gotten through all of the questions typed in. We do have some time left, so I will give people the opportunity to ask questions over the phone if they would like to or you can continue to type in questions. If you want to ask a question over the phone press *5 and I'll open up your phone line.
We do have one question.
Caller Question
Nico, this is Robert DeDomenico with Cargo Fish. My question is that when you answered in the affirmative that grade separation does tend to be expensive, did you consider grade separation might also apply to capillary scale of system, such as each municipal utility really is water, gas, sewer, electric, and information? Very capillary scale and their expense is not on the same magnitude. But they are grade separated. So, did you consider that in your answer? Or is that perhaps outside of what you considered in the scope before you answered?
Nico Boyd
Yeah, that was really outside of our scope, considering the context of my presentation was really focused more on delivery of commercial goods. That wasn't something we considered, but is definitely something we could look into for future projects and expansion of the research we've already done. Thank you for that point.
Jennifer Symoun
It does not look like we have any other questions over the phone, nor do I see any more questions typed in. We will go ahead and end for the day. If you think of any other questions continue typing them and I will go back to them. I want to thank all three of our presenters for presenting at today's seminar. I want to thank everybody who was listening in from the audience. I will send out a link of the recording of today's webinar in the next day or so. The January Talking Freight seminar is not yet available for registration but once it is I will send notice through the Freight Planning LISTSERV. Please note that the January seminar will be held a week later than usual due to the TRB Annual Meeting. The Freight Planning LISTSERV is the primary means of sharing information about upcoming seminars. I also encourage you to join the LISTSERV if you have not already done so.
