My to be Thesis – An Economic and Transportation analysis of last-mile delivery

At the half way stage of my PhD, I present to you, my to be Thesis. If you have read my previous blogs, then you can very well guess what this is going to be about. Yes, lets talk about
E-commerce!

In the winters of 2015, breathing the smog filled air of Delhi, I thought to myself – what is the role of vehicular traffic on the deteriorating air quality and how can we foster sustainable urban mobility? An year later, having done some foundational courses on transportation, specifically public transportation, I had even more questions to ask. And in a quest to answer these questions, I set myself on a grueling, hopefully rewarding path of PhD (or a self imposed निर्वासन/exile as I would like to say). In the years that followed, my focus shifted from individual and public mobility to a more a challenging aspect of urban transportation – freight mobility.

In the past few years, transportation systems have radically evolved – re-defining individual mobility as well as commodity flow. On one hand, the advent of shared mobility services such as Uber and Lyft has provided on-the-fly mobility options for commute. On the other hand, the rise of e-commerce has dramatically transformed the shopping experience and has redefined the way carriers distribute goods, affecting delivery lead times and service reliability. The common denominator in both being the Internet. In one of our recently published work, Jaller and Pahwa (2020), we found only 4% of all daily shopping activities to be internet based. This presents a huge scope for the e-retailers to capture larger market shares and improve their profits and revenues. And they do so by making lucrative offers to its consumers, offering free shipping, free returns, same-day, 1-hr/2-hr expedited (rush) deliveries and more. This has made last-mile ever more demanding, both in terms of economic as well as environmental sustainability.

In general, it is argued that since delivery trucks consolidate demand and optimize delivery routes, e-commerce can result in substantial reduction of negative externalities (emissions and congestion) from shopping related travel. This has been documented in various research work. In fact, in Jaller and Pahwa (2020), we found that e-commerce can reduce Vehicle Miles Traveled (VMT), often used a proxy for congestion, by 80% if e-commerce becomes a dominant channel for shopping. However, any benefits accrued by shopping online are wiped out by rush deliveries as it compels the e-retailer to ship packages at lower consolidation levels, make more delivery tours, thus increasing distances traveled and emissions. Free return is yet another way in which e-retailers pursue larger market share. The tendency to return a product is not founded in the functional error of the product but can be ascribed to lack of information about the fit and suitability of a product. The apparel segment especially observes high rate of return, which can go as high as 45% (Cullinane et al., 2017), i.e. almost every other product is returned. And although reverse logistics incur additional operational cost (also emissions) to the e-retailer, not providing free return puts the retailer at the risk of losing the customer. This compels the e-retailer to provide free return, face the additional operational cost, produce more emissions just to keep the consumer happy. Hence, in general, one can conclude that the environmental efficiency of deliveries reduces significantly as companies make such lucrative offers to its consumers. The above results and discussion therefore highlights the importance of stakeholders, in particular, for consumers in consolidating their demand, for e-retailers in consolidating the deliveries, and for planners and regulators in managing the urban freight system to foster sustainability.

To keep pace with the growing needs of e-commerce, last-mile operators have developed, tested and implemented various last-mile strategies around the globe. This includes use of micro-hubs, which are consolidation facilities often used in conjunction with alternate fuel delivery vehicles, such as electric trucks and cargo bikes, thus limiting the use of diesel trucks within the city limits. On the other hand, the e-retailer can set up collection points, wherein the customers can collect packages from. While this reduces the operational cost for e-retailers since the onus of making last-mile is transferred to the customer, this strategy results in additional external cost from individual customers traveling to the nearest collection point. More recent strategies include use of a mobile micro-hub in conjunction with drones and robots. That on the right is a Kiwi delivery robot, already in service at the UC Berkeley campus, and under testing at UC Davis.

From a regulatory aspect, efforts to mitigate the negative externalities from freight include road tolls, congestion pricing, geo-fencing, load factor controls etc. And while such measures can potentially bring down externalities, it is equally important to note that freight is fundamental to economy. Any regulatory restriction can affect the economic sustainability of last-mile distribution (Tamagawa et al., 2010) and in turn the local economy. This gives more impetus to develop last-mile structure that can sustain operations even under strict regulations, thereby bringing down externalities at the minimum cost to the e-retailer.


Hence the objectives of my thesis are to
1. Identify the diverse conditions in which e-retailer operates
In an economic sense, the last-mile distribution environment comprises of three key agents – the consumer, the e-retailer itself (this includes all the firms responsible for the last-mile), and the regulatory body. Since the consumer dictates the demand it is imperative to realize how this demand manifests. In particular, it is necessary to understand the online shopping behavior, which is not just limited to the purchase behavior- studied in Jaller and Pahwa (2020), but also encompasses lead-time choice (no rush delivery, same-day/2-hr/1-hr rush delivery), shipping choice (delivery to home, delivery to a collection point) and product return behavior. In addition, the consumer, and in general the larger community, pressurizes the e-retailer (through the regulatory body) to develop environmentally cleaner means of distribution. The e-retailer on the other hand develops optimal (mostly least cost) strategies to carry out last-mile distribution- to a certain extent in response to the demand, to some extent in response to the regulatory framework, but also in pursuit of higher market shares and higher revenue as discussed above. This in turn brings economic prosperity for the community. These interactions are hence fundamental in understanding efficacy and sustainability of last-mile strategies. The figure below depicts the above discussed interactions between the different agents in the last-mile distribution environment. Hence, the first objective of this work is to identify and develop, through the combination and interaction of the three key agents, real-life environments under which e-retailer operates and serves the market.


2. Model last-mile delivery operations
When it comes to developing tools to assess last-mile operations, the main objective is to model the process and the sequence in which the customers need to be visited. This, referred to as the Vehicle Routing Problem (VRP), originally addressed 60 years ago by Dantzig and Ramser (1959) as “The Truck Dispatching Problem”. Since then many variants of the problem have been developed and addressed. In the context of last-mile service, it is common to observe – uncertain demand, uncertain travel times (stochasticity); evolution of market with new customer demands, customer unavailability (dynamism), and time-of-day dependent operations and traffic conditions (time-dependence). Hence, in light of such complex operation environments (which will be precisely identified as a part of this work), this work will build a state-of-the-art heuristic for Time-Dependent Dynamic-Stochastic Capacitated Vehicle Routing Problem with Time-Windows (TD-DS-CVRPTW) – more on this in another blog.

3. Assess the sustainability of different last-mile distribution structures and regulatory policies
Sustainability is founded on three fundamental pillars – Economic Viability, Environmental Protection and Social Equity. In our recently presented work, Pahwa and Jaller (2020), we developed a cost-based sustainability assessment model to evaluate the economic and environmental aspects of sustainability for point-to-point (door-to-door) last mile delivery.  The cost model evaluates Economic Viability as economic costs which includes the Fixed and Operational cost, while Environmental Protection is evaluated in terms of the social cost of emissions, modeled as External (emission) cost. Modeling Social Equity for last-mile delivery operations is however beyond the scope of this cost model. Nevertheless, certain regulations such as geo-fencing disadvantaged communities and other local/regional policies that restrict truck movement to reduce exposure to emissions in such disadvantaged communities will be tested in this work.

With this I hope to develop a holistic understanding of what strategy suits best under different operating conditions, strategies that can result in reduced externalities at minimal cost to the e-retailer.

Notes for nerds
Evaluating the environmental impacts of online shopping: A behavioral and transportation approach – Jaller and Pahwa (2020)
Evaluating costs and distribution structure in last-mile deliveries under short time-windows – Pahwa and Jaller (2020)
Improving sustainability through digitalization in reverse logistics – Cullinane et al. (2017)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s