Scary E-Commerce Math

By Kevin Gue | 09/15/2017 | 1:53 PM

November and December in Louisville look like the corn harvest in Iowa—all hands on deck moving packages and loading trucks. My non-scientific survey of local distribution center managers suggests that volumes increase between 2–5 times their normal levels during the peak. That got me to wondering, just how bad (good) could this get? In 10 years, how many packages will be moving, and more importantly, how will those packages get moved?

To begin, let's consider only retail sales because they account for the majority of e-commerce shipments. Let's also assume dollars as a surrogate for number of shipments. In other words, a 10 percent increase in e-commerce retail sales means a 10 percent increase in shipments that need to be handled.


Data source: U.S. Census Bureau


The figure nearby shows that e-commerce has grown as a percentage of retail sales from less than 1 percent in 2000 to more than 8 percent in 2016, an 8-fold increase in percentage terms. But percentage growth doesn't tell the whole story, because retail sales are also growing every year along with the overall economy. Accounting for retail sales growth, the second figure shows that overall e-commerce sales has been growing in excess of 15% CAGR (combined annual growth rate) during the past seven years.

Data source: U.S. Census Bureau

Fifteen percent growth per year can't go on forever, of course, so we need another way to think about the growth of e-commerce during the next 10 years. Very likely we are in a standard S-shape technology adoption curve, in which the adoption rate is low at first, then peaks as the technology gets hot, then slows again as adoption reaches steady state. No one knows what percentage will constitute “steady state” for e-commerce in retail, but even conservative estimates are pretty scary: if retail grows at 2.5% per year and e-commerce rises to just 20% of retail sales in the next 10 years (which I believe is conservative), the total e-commerce sales volume, and by proxy the total number of shipments, will rise by 3.2 times. You read that right: the number of packages handled will be approximately 3.2 times current levels. If e-commerce rises to 30 percent of total retail, e-commerce shipment volume will increase by a factor of 4.8. At 50 percent, the factor is 8. And we haven't even thought about peak volumes, which are 2–5 times steady state or more.

How does this get done?

Data source: Bureau of Labor Statistics

The traditional answer, at least in the U.S., has always been “more labor,” but the labor market is already tight and shows no signs of improving. Average U.S. unemployment is currently just above 4 percent (see figure), and the Bureau of Labor Statistics estimates that for demographic and other reasons, the labor force will grow only about 0.5% between 2014–2024.

Allow me to summarize: the labor pool projected to grow by 0.5% between 2014–2024 is supposed to support an e-commerce distribution industry projected to grow by 300%–500% in the next ten years? Really?

Something must change.

I believe that "something" is widespread adoption of material handling automation. Fortunately, recent advances in autonomous vehicles, robotic order pickers, and decentralized control algorithms make it easy to envision a future in which a very few people will work alongside a very many machines to make all this distribution possible.

What will be really exciting to watch, is how fast we can produce the machines necessary to enable this explosion in automation. If next week a company began selling a truly autonomous order picking robot for $20,000—a robot that traveled at least as fast as a human and picked individual items faster and with greater accuracy than a human—could they build them fast enough?

This day is coming.

[This post also appears on my personal website as "Doing the Math on E-Commerce Growth."]

Logistics Automation and Us

By Kevin Gue | 01/05/2015 | 9:05 AM

Several years ago I began telling my children that I thought the greatest challenge for their generation was sensibly integrating technology into everyday life. It seemed at the time (and I continue to believe) that the blind adoption of rapidly advancing technology would have unknown and possibly deleterious effects on the human condition. Sounds like the subject for a nice book, eh? Alas, Nicholas Carr has beaten me to it with The Glass Cage: Automation and Us.  51lESeR-C+L

While recognizing the industrial and social benefits of automation, Carr points out that our inventions no longer help us accomplish work, but rather do the work for us. If the work is mindless or backbreaking, so much the better, but technology now threatens to rob us of many of the experiences that make us human. As I have written elsewhere, this point really resonates with me. I am reminded of a rhetorical question posed by a friend several years ago: "Why do I have to remember anything, when I can just look it up?" That is a serious question. The answer, of course, is that remembering—knowing—is a critical part of what makes us human. Machines look things up; humans know.

The challenge is to build technology that relieves man of the burden of work without robbing him of the satisfaction of work. Here Carr gives a huge shout out to the human factors research community, which knows a great deal about the interaction of humans and their machines, but which Silicon Valley has little interest in accommodating if that means limiting what can be done (and how much money can be made). Utopia, we are told, is life without work, instructing BeerBot to fetch us a cold one while we lie on the couch watching Netflix. Come to think of it, why can't BeerBot just anticipate that I need a cold one!

Baxter the Robot
Baxter the robot. Source: BBC News, which included the funny caption “Baxter can work happily alongside human co-workers.” We ask, can humans work happily alongside Baxter?

Back in the industrial world, automation has changed laborers and craftsmen into button pushers and caretakers, making sure nothing goes wrong. But as Carr points out, automation is not just a threat to the blue collar workforce. White collar jobs that involve design, analysis and decision making are very much in the cross hairs. If corporations are willing to replace workers with robots, why would they hesitate to replace a multitude of managers with Mr. Algorithm? They won't.

The effect of technology on employment is neither the point of Carr's book nor the point of this post. I am more interested in the effect of logistics technology on us. Is our industry developing machines and devices that improve the human condition, or are we developing machines and devices that solely improve ROI? Are these objectives mutually exclusive? Must they be?

What is really interesting to me is the prospect that, in an environment of scarce labor resources, companies that develop and implement human-centric work environments with "human optimized automation"—perhaps at a higher cost—might have the last laugh. What exactly "human optimized automation" looks like is still an open question, but I am sure it doesn't look like a row of buttons and toggle switches. Here's hoping that Carr's book gets a wide reading in our industry, and that suppliers and end-users alike find a way to develop automation that serves rather than dehumanizes us.

[This post also appears on my personal website.]

Just how bad is the labor problem?

By Kevin Gue | 09/16/2014 | 1:21 PM

Recently I visited the distribution center of a local 3PL that supports e-commerce and other distribution operations. At some point in the visit I asked my host if he was having problems with labor availability. "Oh, yea." He went on to tell me that his firm was considering locating new clients into DCs away from Louisville for exactly this reason. Let me say that again: his firm was moving business away from Louisville, Kentucky (home of the UPS Worldport and about 100 miles from the population weighted center of the continental U.S.) specifically because they were having trouble finding workers. OK, so I'm partial to Louisville, but can you imagine moving e-commerce business away from Louisville when you already have a DC there? That's how bad the labor problem is.

Help Wanted

Driving away from the DC, it occurred to me that labor availability is an obvious constraint on supply chain network design problems, but I wondered if there had been any work in this area. It is tempting to think that labor availability is "just a capacity constraint," until one realizes that competitors are also trying to scoop up that capacity. Mix in the uncertainty of labor availability data in the first place, and you have a very hard problem.

The labor shortage in distribution earned an entire chapter in last year's U.S. Material Handling and Logistics Roadmap. Participants in the workshops identified two classes of labor that were particularly troublesome: skilled workers, such as those who could program PLCs and do installations of automation, and unskilled workers who had to have a solid work ethic and a willingness to learn. One participant told a prototypical story, "If I have two open positions, I'll receive a hundred résumés. I discard ninety immediately, due to no high school degree, failed drug tests, felonies, and so on. Of the remaining ten, I'll hire seven. Five won't show up for work for all five days of the first week—they didn't realize the job would be that demanding. That leaves me with two." No one in the room thought he was exaggerating, and no one was laughing.

We are past the time of complaining about this problem. 

What has your WMS vendor done for you lately?

By Kevin Gue | 08/29/2014 | 9:16 AM

One of the challenges identified at last summer's workshops for the Material Handling and Logistics Roadmap was the use of optimization tools to improve warehouse operations. It was widely believed that the state of the art far exceeds the state of practice. That is, we know how to optimize certain aspects of warehouse operations, and yet very few firms are making use of the algorithms and tools that make it happen. How come?

Let's take order picker routing as an example. In many manual order picking environments, the WMS batches orders into a pick list for the worker. Deciding which orders should go into the batch is actually a difficult problem, but let's skip that for now. Once the batch is determined, the lines must be sequenced such that the picker travels as little as possible. That is, we wish to minimize the length of the picking tour, thereby maximizing the worker's productivity.

[Source: www.directindustry.com]

Sequencing pick locations to minimize the length of a tour is a form of the famous "traveling salesman problem" (TSP), which asks in what sequence a salesman should visit a collection of cities in order to minimize his travel time. The TSP has received more attention in the operations research community than perhaps any other problem, and we now have very efficient methods for solving very large instances. For order picker routing, that means optimal methods exist to solve almost any problem. And by "optimal" I don't mean "very good" or "excellent." I mean we can generate picking tours that are provably as short as possible, and we can do it very quickly.

Which brings me back to the WMS. How many WMS vendors actually optimize order picker routes? I don't know the answer to that question, but an anonymous vendor in a land far, far away once confessed with chagrin that they used the most naïve algorithm imaginable. When asked why, the engineer told me, "Customers haven't requested optimization." There you have it—a perfectly rational explanation! What this vendor was telling me was essentially, "We build into our product what our customers want, not what we think they might want."

And now the $64,000 question for WMS customers: Have you asked for optimal routing methods? Do you know they exist? Have you asked for optimal slotting methods, which also exist? Have you asked for details on the order batching algorithms under the hood?

If you are a WMS vendor that uses optimal routing algorithms for order picking, by all means write to let me know. I'll be glad to update this post. If you are a WMS user wishing your WMS was as good as it could be...

You do not have, because you do not ask.


Why Uber and friends should keep you up at night

By Kevin Gue | 06/10/2014 | 8:16 AM

The drumbeat of a new "sharing economy" continues with excitement I haven't seen since 1999. For those with day jobs who seem unable to keep up with this stuff, the sharing economy (or crowdsourced economy) is based on the insight that much of today's commercial activity—especially in logistics and transportation—essentially duplicates functions that ordinary people do in their ordinary lives. Example: everytime I drive 1.5 hours to the Atlanta airport with three empty seats in my car, dozens of people are paying $40 for a shuttle to exactly the same place at about the same time. Why? Because until now there has been no coordinating mechanism to avoid the duplication.

Recent news that Uber, a crowdsourced transportation service with big ambitions, was valued at $18 billion should make executives in the logsitics industry quiver more than just a little. What happens to the logistics industry when millions of automobiles on our nation's highways become miniature freight transportation providers? "Anyone driving from Atlanta to Louisville this morning? Got room for 4 medium sized cartons?" Can you say, "Nationwide Transportation Company with Near-Zero Capital Investment?" I thought you could.

We're all truckers in the sharing economy [image: npr.org]

Again, the only obstacle to this business model has been coordination, but Uber and other companies are now exploiting the mobile computing revolution to provide both the necessary coordination and a low-cost means of paying all these new drivers. Whether or not a sufficient number of folks sign on to make this a threat to the current people transportation industry is an open question, and there is another leap to make it work for time-definite freight delivery. Nevertheless, I believe new crowdsourcing models will play a significant role in both industries in the near future.

How about this scenario: For my flight tomorrow, TripIt knows I'll be leaving my house around noon. CrowdLog (made up name!), which has a contract with TripIt, sends a notification during the night asking if I can stop by a local auto parts store to make a delivery to a customer in LaGrange (on the way to ATL). I tap "Agree" and pick up the part on my way out of town and deliver it to a Cracker Barrel at Exit 18, from which another "crowdsource participant" picks it up and makes the delivery right to the door of Mr. Intended Customer. Presto—no distribution center, no conveyors, no forklifts, and no 18-wheelers!

That gets me excited enough to lose sleep tonight. How about you?

Thoughts on Augmented Reality in the Warehouse

By Kevin Gue | 05/09/2014 | 8:06 AM

I was recently sent a video of Google Glass being used in an order fulfillment center. The prospect of using Glass (or similar devices) in warehousing was mentioned by a participant at one of the Material Handling & Logistics Roadmap workshops this summer, so the idea wasn't new to me. But seeing the video for myself got me to thinkin'.

In a word, the video is impressive. Whether the implementation is real or constructed for video I don't know, but it's easy to imagine that such functionality will soon be with us, if it isn't already. Very much like voice picking technology, Google Glass directs the worker to proceed to location such and such and "pick three" items, after the location is illuminated by a green rectangle cast before the worker's eyes. No mistake: that's the location.

While driving his lift, the worker sees a green arrow indicating which way to turn to get to the next location, presumably following the shortest route. We have therefore solved the "how to direct the worker's route" problem.

After illustrating a couple of picks and put-aways, the forklift supposedly suffers a malfunction, which is detected by Glass, and the operator is directed to proceed to a maintenance area. There, a maintenance technician appears in the Glass to help the operator effect a simple repair. Off goes the worker for more Glass-directed picking. What could be better?


If you know me, you're probably expecting a contrarian view. I'll get to that in a minute. But first, it turns out that research on applications of "augmented reality" in warehousing have been underway for a decade. The first papers I could find on the subject appeared around 2005. Color me embarrassed! The authors are from Germany, where so many of the latest developments in logistics technology are happening.

There is much to like about these developments. As we wrote in the U.S. Roadmap for Material Handling and Logistics, wearable computing offers the industry a significant opportunity to improve operational control and reduce costs. As the positions and activities of workers become more and more transparent to "the system," human error and inefficient behavior (e.g., picker routing) will become increasingly rare. All to the good.

But one part of the video gives me pause. When the forklift operator slides back the battery cover to investigate the source of a malfunction, a technician appears instantly in the Glass to provide expert advice on the repair. "What is the voltage? Just plug it in and you should be fine." I couldn't help but think, "He's been robbed! Let the man solve his own problem!" In our drive to make all things as fast and easy as possible, we've robbed the operator of the joy of problem solving—diagnosis, critical thinking, problem solving—and more importantly, the sense of accomplishment from having repaired that which was broken. Call me the Industrial Romantic, but this just makes me sad!

I can hear the laughter through my monitor. "Hey Kev, how about you plug into the real world of ROI and quarterly earnings reports like the rest of us!" Having spent most of my academic life thinking about ways to reduce logistics costs, I am not unsympathetic to this objection. But I'm also a worker myself, and much of the satisfaction I derive from work comes from solving problems and accomplishing tasks I feel are important. I couldn't help wondering what value workers provide in a Glass-directed life—beyond 10 very capable fingers to pick, put, and push.

When I shared all this with my 17 year old son, he said "Dad, did it ever occur to you that not everyone enjoys problem solving as much as you?" Well there you go! This is what makes this subject so interesting to me: what seems to one person "death by robotic instruction," is to another a faster way to get stuff done (think voice picking). All hail, the diverse workforce. Aren't we humans wonderful?

The opinions expressed herein are those solely of the participants, and do not necessarily represent the views of Agile Business Media, LLC., its properties or its employees.

About Kevin Gue

Keving Gue

Kevin Gue is a professor of industrial engineering at the University of Louisville, where he holds the Duthie Chair of Engineering Logistics and serves as Director of the Logistics and Distribution Institute (LoDI). His research addresses the design and control of logistics systems, with a focus on distribution, warehousing, and material handling. He is co-inventor of the warehouse aisle designs known as the Flying-V, Fishbone, and Chevron, work for which he received multiple best paper awards and was awarded the Technical Innovation in Industrial Engineering Award from IIE in 2009. Kevin is a former president of the College-Industry Council on Material Handling Education and is Editor-in-Chief of the recently published U.S. Roadmap on Material Handling and Logistics.


Popular Tags

Subscribe to DC Velocity

Subscribe to DC Velocity Start your FREE subscription to DC Velocity!

Subscribe to DC Velocity
Go digital