Escolar Documentos
Profissional Documentos
Cultura Documentos
Automation in logistics:
Big opportunity, bigger
uncertainty
As e-commerce volumes soar, many logistics and parcel companies
hope that automation is the answer. But as this second article in our
series on disruption explains, things are not so simple.
April 2019
The history of logistics is also a history of Three cheers for automation
automation, from the steam engine to the forklift At first blush, more automation seems like the
to today’s robotic pickers and packers. So today’s answer to three problems facing contract
fevered interest in new machinery, after a lull logistics companies.
of several years, has plenty of precedent. Many
trends are thrusting automation toward the top Start with a shortage of workers. It’s no secret that,
of the logistics CEO’s agenda, not least these at least in the United States, labor markets have
three: a growing shortage of labor, an explosion tightened. Unemployment rates are at a 50-year
in demand from online retailers, and some low, and wages are increasing. Some of the largest
intriguing technical advances. Put it all together, e-commerce facilities currently require 2,000 to
and McKinsey Global Institute estimates that the 3,000 full-time equivalents, an order of magnitude
transportation-and-warehousing industry has more than traditional distribution centers employ,
the third-highest automation potential of any and need to add even more workers during the
sector.1 Contract logistics and parcel companies holiday peak season, when labor is most scarce.
(which, for sake of convenience, we will call simply
“logistics companies”) particularly stand to benefit. While many of the jobs that might be automated
(Automation is also on the table at other transport are currently difficult to fill, that’s not to say that
companies, such as trucking companies and port automation will have no effect on the workforce: it
operators. See sidebar “Automating freight flows: will, and companies must reckon with the significant
Changes for every sector”.) costs to their employees and communities. In 2017,
the US Bureau of Labor Statistics estimated that
Yet for all the excitement, most logistics companies nearly four million Americans work in warehouses
have not yet taken the plunge. For every force as supervisors, material handlers, or packers. That’s
pushing companies to automate, countervailing almost 3 percent of the total labor force; collectively,
factors suggest they should go slowly. We see five they earn more than $100 billion in annual wages.
reasons companies are hesitating: the unusual com- Automation won’t make all these workers redundant,
petitive dynamics of e-commerce, a lack of clarity of course, and many can be reassigned to new jobs
about which technologies will triumph, problems that involve collaborating with and maintaining the
obtaining the new gizmos, uncertainties arising from new machines. But if even a portion of these jobs
shippers’ new omnichannel-distribution schemes, are lost, it will still represent significant upheaval.
and an asymmetry between the length of contracts
with shippers and the much-longer lifetimes of E-commerce, the second trend, is remaking the
automation equipment and distribution centers. entire logistics industry. The inexorable rise of online
sales is well documented. In the United States, for
This is the second in a series of five articles on example, growth has averaged 15 percent annually
disruption in transport and logistics. In the first, we over the past decade, and the range of goods
examined the implications of autonomous trucks. has expanded dramatically. That’s been good for
Automation is no less potent a force. In this article, logistics companies. We estimate that, out of every
we will review the reasons automation is coming to $100 in e-commerce sales, these companies (or
the fore, examine the five factors that are hindering e-tailers’ in-house logistics units) are collecting
investment, and lay out strategies that can position $12 to $20, a massive increase from the $3 to $5
contract logistics companies to prepare for an spent on logistics in a typical brick-and-mortar-
uncertain future. retail operation. (It’s important to note that, in our
1
Michael Chui, James Manyika, and Mehdi Miremadi, “Where machines could replace humans—and where they can’t (yet),” McKinsey Quarterly,
July 2016, McKinsey.com.
Automation will affect the supply chain automation-ready capabilities rather than line a longer-term prospect, rail operators
far beyond the walls of the warehouse and simply automating old processes. And they and governments are investing in technol-
sorting center; it will change the way goods can apply better project discipline to en- ogies that lay the foundation. Positive train
flow across all modes (exhibit). In the first sure that automation investments account control (PTC) is a long-desired step toward
article in this series, we addressed the for all attributes of port operations. an automated future: its data links allow
impact of autonomous trucking, a critical for real-time automated control of sets of
automation technology, on roads, rails, Of the remaining transport modes, au- trains. Several European and US railroads
and ports. And our colleagues recently tomation in ocean and air freight is quite have PTC schemes in the works, and a few
produced a detailed look at other forms possible but will probably not move the have fully implemented them.
of port automation. They find that while productivity needle much. In rail, automa-
ports are accelerating their adoption of tion will likely begin in terminals, which offer Over time, railroads will continue to search
automation, they are not yet recouping controlled environments and repeatable for opportunities to automate the main line,
their costs. Moreover, while operating
2019 processes. Intermodal terminals will but some limits will persist for the foresee-
expenses are falling as expected (by
Automation in logistics: Big opportunity, 15 to likely seebigger
increased use of autonomous
uncertainty able future. For example, trains traveling
35 percent),
Exhibit throughput
sidebar of is falling as well (by hostlers to move containers to and from
sodebar heavily populated routes or hauling haz-
7 to 15 percent). Port operators can take trains. Autonomous cranes are also likely to ardous materials will likely continue to need
several steps to get the most out of auto- emerge in the near term. While the physics human oversight.
mation. Among other moves, they can build of trains makes automation on the main
Exhibit
2
Rick Braddock, “To compete with Amazon, big-name consumer brands have to become more like it,” Harvard Business Review,
June 14, 2018, hbr.org.
Exhibit 1
Multishuttle system Analytics tools Optical recognition Conveyor connection Management system
Typically used with an Algorithms that help operators Sensors that scan items (often A connection between 2 Analytic and digital systems that
automated storage and retrieval analyze performance, identify on 6 axes) to apply sortation disparate conveyor systems integrate analytics, performance
system (AS/RS) that moves trends, and make predictions and other logistics. Examples that often uses decision logic reporting, and forecasting tools,
goods (mostly on pallets) in 3 that inform operating decisions, include a conveyor’s diverts, to affect the flow of items. allowing managers to easily
dimensions to store and retrieve often using machine learning to laser-guided vehicles, and Typically, connections integrate control a full system such as a
items without human improve over time. camera-based movement different systems of flow, for warehouse.
intervention. of drones. example push and pull flows.
Smart storage 3-D printing Swarm AGV1 robots Smart glasses Picking robot
Storage solutions that use Also called additive Autonomous guided vehicles Glasses that augment and assist Systems with robotic arms that
advanced analytics and digital manufacturing, this process that operate freely or on digital reality of wearers—for example, mimic human picking motion.
tools to place and retrieve items creates parts by adding layers tracks to bring items (often from by displaying directions to Picking robots can be fixed (with
in the most efficient way, of a material (metal or plastic, a storage rack) to a picking storage locations for picking— goods brought to them) or mobile
adjusting storage media based typically) to create a desired station based on instructions reducing inefficiencies of (traveling to storage to pick
on the product, picking, and shape. from the order-flow software. searching. items).
order characteristics.
1
Autonomous guided vehicle.
Source: McKinsey analysis
day as well as instant delivery, are a potent step in processed 812 million orders, eight times more than
that direction, and logistics companies will have to on a typical day. If logistics companies are to fulfill
carefully monitor the pace of change. customer expectations during peaks, they will have
significant spare capacity for three-quarters of the
A particular challenge of serving e-commerce year. And if they do not build sufficient capacity for
companies is that demand is very spiky, easily peaks, e-commerce giants have further incentive to
doubling around Christmas or Singles’ Day. On build their own capabilities, as Amazon did after the
Singles’ Day 2017, Cainiao, Alibaba’s logistics arm, 2013 Christmas season.
Warehouse automation technologies can shuttles moving in three dimensions on rails to determine the appropriate action. Auton-
be broadly categorized into devices that as- attached to the structure. omous palletizers use robotic arms to build
sist the movement of goods and those that pallets from individual units and cases, of-
improve their handling. In the first group, New handling devices automate the pick- ten using advanced analytics to determine
we’ve already seen automated guided ing, sorting, and palletizing of goods. Pick- the optimal placement for each box.
vehicles (AGVs) that move cases and pallets. ing systems typically include a robotic arm
New twists are the equipment and software with sensors that can determine the shape Beyond the machines that mimic human
needed to retrofit standard forklifts and and structure of an object, then grasp it. hands and arms, other innovations will
make them autonomous. The new gear can Some devices remain fixed and have goods improve the productivity of people in ware-
be switched on whenever needed—peak brought to them (often by AGVs). Others houses. Drones are already in use in the
seasonal shifts, say—and the forklift can re- travel to the goods and retrieve and move warehouse for inventory management and
main manual when demand is slower. Other them at once. Magazino’s new TORU cube outside the four walls for yard management.
recent technologies include swarm robots is an example of the latter. We expect to see much greater adoption
(most famously, Amazon’s Kiva robots) of drones for these uses. Exoskeletons
that move shelves with goods to picking With the e-commerce boom, efficient augment human motion with mechanical
stations and advanced conveyors that can sorting has become increasingly important, power through gloves or additional support
move goods in any direction. Advanced particularly in parcel operations. Advanced for legs. The systems feature electric
automated storage/retrieval systems (AS/ conveyor systems use scanners that can motors that augment the person’s own
RSs) store goods in large racks, with robotic pick up bar codes on any side of a package strength to allow them to move more goods
(for example, heavier items) or move goods
more easily and safely.
Technology racing ahead investment at all. The cost of removing and replacing
We combed the industry and found more than equipment, much of it not fully depreciated, would
50 technologies that could further automate put unlucky investors in a deep hole.
some part of the supply chain, including many
in logistics (Exhibit 2). All are much more than a Purchasing woes
twinkle in some technologist’s eye, but none are Even if a logistics company makes a great choice
yet in widespread use. The question that confronts about the automation equipment to buy, it can run
logistics companies (and warehouse companies) is into another problem. The leading warehouse-
simple enough: Which ones will take off to yield the automation manufacturers have enjoyed strong
greatest return on investment? revenue growth of 15 to 20 percent annually since
2014. At many, order books are now full. In 2017, the
Finding answers is much more difficult, of course order book at Vanderlande Industries reached an
(see sidebar “Automation technologies to watch” for all-time high. Our conversations with many would-
our thoughts on the first few horses out of the gate). be buyers, especially at parcel companies, suggest
No one wants to buy technology that becomes that manufacturers operating at full capacity cannot
obsolete shortly after acquisition. Not only would even provide them with quotes.
that leave a company less efficient than competitors
that made better choices, it would also leave it Part of the problem is that the manufacturers are
worse off than those competitors that made no not yet at scale. Many companies, including the
Exhibit 2
Maturity, in stages
1
Speed of innovation adoption based on maturity.
Source: McKinsey Supply Chain 4.0 Innovation Survey
market leaders, are focused on a narrow range of order is “overspec’d,” or more expensive than it
technologies and solutions. That may change: the might have been. We have seen purchase prices for
industry is in turmoil, with significant M&A activity the same equipment vary by as much as 50 percent.
underway. Notably, large technology conglomerates
are investing in automation start-ups. For example, Rapid change in shippers’ distribution networks
in 2015, Siemens took a 50 percent stake in Brick-and-mortar retailers are reacting to the
Magazino, a start-up that builds automated picking e-commerce onslaught in part by evolving their
robots. Once the dust has settled, some larger distribution networks into omnichannel systems in
companies that are better able to meet demand may which consumers can purchase and receive items
emerge. Then again, such companies will also have through any channel. They might purchase online
stronger pricing power. and take deliveries at home, the classic e-commerce
model. Increasingly, they can order online and pick
A related issue is some confusion at logistics up in stores. Or they might purchase in-store and
companies about which advanced equipment they receive shipments at home, an option that menswear
truly need. Often the equipment on the purchase company Bonobos and other companies offer. And
Shippers—the manufacturers and part because they cannot find logistics Shippers should expect their partners to
retailers that hire logistics providers to companies that will invest enough in auto- seek contracts in line with the life cycle
move their goods—will also grapple with mation to meet their needs. of automation investments. Put another
automation in coming years. As new way, logistics companies will seek to share
technologies come online and omnichannel Beyond the level of investment, shippers some of the technology upside—and some
delivery becomes more common, most will and their logistics partners must also con- of the risk—with customers.
need to revisit their long-standing in-house tend with the complexity of omnichannel.
and outsource decisions. Shippers inter- Take one example: to operate efficiently, an Shippers cannot outsource completely
ested in automation must first determine omnichannel retailer must either open the the intricacies of automation and the best
whether they have the capital and know- full inventory system to the logistics com- practices of automated warehouses. To
how to invest effectively in automation or pany so that it can route orders between be a smart customer requires enough
whether it is more economical and easier to stores and fulfillment centers or add steps knowledge of automation to evaluate bids
outsource increasingly complex warehouse to the order-routing process to determine intelligently. Contract logistics companies
operations to a logistics company. The whether the order remains in-house or is we speak with often see automation listed
same uncertainties about omnichannel that sent to the logistics company. prominently, yet typically with sparse
hold back logistics companies’ investments detail, in requests for proposals. Shippers
in automation can also constrain shippers. Supply-chain managers should also expect frequently know they want automation but
However, our analysis indicates that ship- changes in their negotiations with logistics don’t know what kind they need. Getting
pers are investing more in automation than partners. As contract logistics players add a fair shake from logistics companies will
logistics companies are (see section “Five more fixed costs in the form of automa- require shippers to stay aware of technol-
reasons for hesitation” in article), in large tion, their strategic flexibility will decrease. ogy trends and understand well how these
might meet their needs.
of course, they can still go to the store and walk out in the past. Shippers have tried to cut costs by
with their purchases. On top of that, consumers more frequent tendering and have sought greater
demand ever faster delivery, which requires more flexibility to respond to rapid changes in customer
local storage capacity, further driving complexity. demand. The trend has exerted significant pressure
Building a supply chain to support an omnichannel on logistics companies. Because they typically
system is highly complex (Exhibit 3). develop sites with a particular customer in mind,
they need to calculate carefully the investment
With all this complexity comes a lot of uncertainty: required to add a new customer. With a significant
Where should new fulfillment centers be built? What initial investment required, logistics contracts are
share of B2C orders should they accommodate? often not profitable for two years. That leaves only
And perhaps the biggest question: How much and a year or so of profit before renegotiations begin.
what kind of automation are ideal? Shippers are Big investments in automation would push the
asking the same sorts of questions (see sidebar break-even point back further, leaving logistics
“The shipper’s perspective”). companies at even greater risk that a customer
would change providers, which would leave the
Too-short contracts facility empty and automation equipment unutilized
Most logistics contracts run for about three years, while the third-party-logistics company searches
sometimes longer. That’s much shorter than for a new customer.
Exhibit 3
Vendor
E-commerce Omnichannel
fulfillment center fulfillment center Hub and Logistics Retail Omnichannel
spoke clusters infrastructure infrastructure
Network strategy Network strategy Network strategy Network strategy
— Use segmented — Locate near hubs — Rely on own — Build smaller
approach to place and competitor footprint of brick- e-commerce
inventory in larger facilities to benefit and-mortar stores fulfillment centers
(regional) and from availability close to customer close to customer
Consolidation in Dark Store with cross- smaller (local, of labor and — Train store
logistics cluster store trained workforce metro) sites collaborative associates to serve Applicability
transportation both e-commerce — For geographies
Applicability and store visits with high population
— National/regional Applicability density
— For low volumes Applicability (metropolitan areas)
— Reduced capital- — For areas around
expenditure retail footprint
investment — Reduced capital-
Customer locker expenditure
or drop box investment
In the future, contract planning might get even more Contract logistics
difficult. E-commerce requires dense networks, The big changes we’ve discussed—the simultaneous
especially in urban areas. But no single customer rise of e-commerce, omnichannel supply chains,
has the scale to support a full-scale network. and new automation technologies—present
Logistics companies must therefore build fulfillment contract logistics with a great opportunity to
centers and purchase automation technology sharpen its value proposition, which has historically
before demand is known, let alone contracted. relied on one of two factors:
Ashutosh Dekhne is a partner in McKinsey’s Dallas office, Greg Hastings is an associate partner in the Charlotte office, John
Murnane is a partner in the Atlanta office, and Florian Neuhaus is a partner in the Boston office.
The authors wish to thank Knut Alicke, Tom Bartman, Alan Davies, Mark Staples, Adrian Viellechner, and Markus Weidmann for
their contributions to this article.