This site depends on JavaScript to run. Please enable it or upgrade to a modern browser that supports it.

ASCM Insights

Episode 55: Yossi Sheffi and the Magic Conveyor Belt


Bob Trebilcock: Welcome to The Rebound, where we'll explore the issues facing supply chain managers as our industry gets back up and running in a post-COVID world. This podcast is hosted by Abe Eshkenazi, CEO of the Association for Supply Chain Management, and Bob Trebilcock, Editorial Director of Supply Chain Management Review. Remember that Abe and Bob welcome your comments. Now to today's episode.

Bob: Hello and welcome to today's episode of The Rebound. Yossi Sheffi and The Magic Conveyor Belt. I'm Bob Trebilcock.

Abe Eshkenazi: I'm Abe Eshkenazi.

Bob: Joining us today is Yossi Sheffi. I don't think Yossi needs an introduction to this audience, but just in case, he's a professor at MIT, where he's also the Director of the Center for Transportation and Logistics and the Director and Founder of the Master of Engineering and Logistics Program. If those aren't enough, Yossi is the author of at least nine books, including his latest The Magic Conveyor Belt, Supply Chains, AI, and The Future of Work. Yossi, welcome.

Yossi Sheffi: Thanks for having me.

Bob: Well, we're glad to have you. The last time we had you as a guest, I think we've had you after every recent book has come out. The last time was following the publication of A Shot in the Arm that was your book on the development of the COVID vaccines. Now, that was a time of incredible disruption. As you look at things today as we've come out of COVID, how do you assess the state of supply chain today? What's working? Where are we struggling? Then what responses to the pandemic do you think are permanent?

Yossi: First of all, I think that-- Let me get rid of one misconception, the chain broke down. In fact, my point is that supply, it was their finest hour, so to speak, because the pandemic was not created by supply chain managers. It was created by outside forces and logistics manager, supply chain manager, procurement, distribution, what have you, reacted valiantly. I mean, think about just the food, from one day to the next, restaurants were closed, universities were closed, industrial parks were closed, yet the food kept coming.

Only people in the business understand that the machinery that was used to send stuff to, say restaurants, is not the same machinery that is used to pack stuff for supermarket with all the ingredients on the package. The machinery was not there. People changed their eating habits. They had less fresh food, a lot more comfort food, bread, pasta, cans, yet nobody went hungry. By and large, it worked extremely well.

It drives me nuts when people are saying that the supply chain was broken. No, supply chain acted under incredibly difficult condition and worked well. This, of course, created several things. First of all, it elevated the role of supply chain in cooperation, in the media and with the government, but it also gave, I saw in few organizations, a newfound confidence, "Hey, we went through this, we can go through anything."

Also, a lot of collaborations that were generated during the pandemic, when everybody was facing the same danger, held on. Relationships even between competitors are better than they were before. Now, in terms of the other part of the question, I should say, what we are still struggling-- Of course, we are still struggling with transportation, procurement, but mostly now a lot of inventory in the system, which is a result of so-called bullwhip effect.

We could have predicted it. In fact, I predicted it over a year ago, I wrote a blog saying there will be too much inventory in the system because everybody was over-ordering. They thought that the demand during the pandemic will continue forever. Well, nothing goes forever. A lot of retailers, a lot of manufacturers have a tremendous amount of inventory which impact the system. They order less, so suppliers having to lay people off.

We see a lot of people are laying people off and we are slowly marching into a recession. I still think we'll have a recession, not sure how deep it's going to be. There's still some struggling there. Of course, because of this, the rates of transportation, logistics, other services are low and people are struggling. Anyway, the last part of the question had to do with the responses to the pandemic. That, what I think, is permanent.

I think a lot less than what people think. I think that at the end, the main pressure on corporations is to cut cost, increase service, increase revenue. There's a lot of talk about reshoring and keeping more inventory and getting out of China. On the margin, yes, we see some companies are starting to move out of China. In many cases, it is only the last stage of production, the assembly.

People are doing the last stage in, say, Vietnam, so they can put a label of made in Vietnam rather than in China, but a lot of the deep supply chain, many, many tiers, are still in China and companies invested decades and billions of dollars in building it in China. It's not easy to get out of China. We see some regionalization in order to create more resilience. Some companies are moving out of China, moving towards the Americas or the extended EU, but it is not yet a tsunami.

We see some of it. I'm not sure how much of it will stay. The last point is that people are talking about the end of just-in-time. The media is talking about it doesn't understand Just-in-Time. They think that just-in-time is a method to reduce cost so we'll have less inventory. They don't understand just-in-time is a method to create high quality. The original Toyota manufacturing system was there in order to increase quality. As a result of increased quality, you have less rework, less recalls, and of course, in general, lower cost.

That's what JIT and the Toyota manufacturing system in general are about and this is not going away. Anyway, I gave you a long answer to a short question.

Abe: Yossi, let's jump into your new book. What was the catalyst for writing it? Why was this the right time to tell that story?

Yossi: The right time to tell the story was that people were bugging my wife. They said, "We hear so much about the supply chain." I'm not talking about people in the profession. I'm talking about friends, neighbors. "Your husband is in supply chain. What is this? What is this thing that people are talking about?" I mean, for us in the profession, it's amazing that people have no clue but most people have no clue.

Rather than meet every one of my wife's friends and neighbors one-on-one I decided to write the book. The beginning of the book is explaining what supply chains are, but more than that, trying to explain the complexity and trying to get people to a mindset that they should not be upset when some item is not on the supermarket shelf or the Amazon warehouse. In fact, they should be amazed and delighted when it's there.

Once they understand the complex processes that are involved in getting it there, that are involved in extracting material from the earth, going through a series of tiers of manufacturing of parts and bigger parts and sub-assembly and getting into assembly, moving it all over the world, going through dozens of legal and custom regimes and just getting it to the supermarket shelf. Once they understand it, they get a new appreciation for what it takes to do it.

I'm trying to get people to do it, to understand it. Then as I was writing it, I started to look at technology in the supply chain because some of the people who I talked to still thought that logistic and supply chain management is about people driving trucks and moving boxes. They don't understand the use of technology in the profession, become a very sophisticated use of the latest method because the operation, because the networks are so complex and so wide-ranging.

I started talking about the tech involved and I also ended up-- This was just luck, or I don't know, prescient when I started talking about AI even before November last year when ChatGPT came out and everybody started talking about AI. I even included something about ChatGPT at the last minute and inserted it into the book. The book came out in February this year so basically, I had it done by December.

I included all this in order to explain to people the use of technology. Once I started talking about the use of technology and I started to explain more about the role of AI in society in general, because this is infrastructure technology that has a lot of implications as we see today in the media.

Abe: Yossi, first now I know how to get to you, and that's to go through your wife. That's thank you for that tidbit.

Secondly, have we created this problem with the expectations of just-in-time that prior to the pandemic, there was very little hesitancy on individuals to order and receive things same day? Have we conditioned them to expect more than the supply chain could deliver?

Yossi: Clearly, the profession was a victim of its own success. The fact that with delivering day in and day out, anything that people wanted, unbelievably fast, people thought that that's how God created the earth. That's normal. Nobody realized what it takes to do something like this until there was a disruption. Then we say, "Oh my God, this is what it takes." I tried to highlight some of this in my book.

Bob: Yossi. I want to pick a little bit at some of the things you and Abe just talked about both complexity and meeting consumer demand. For years, when I was the editor of Supply Chain Management Review, I was trying to get somebody to write something about complexity and how it was impacting supply chain management. It really fascinated me, and I know it's part of what you deal with.

One of the things I wondered, I don't know if this is anecdotal or if it's a trend, but I've had conversations recently with Proctor and Gamble with one of the major candy companies called Ferrara, one of the things they do is Candy Corn, and I'll give you the example in a moment. Then with a leading retailer. The Ferrera, the Candy Corn episode was that during the pandemic, because they had limited ingredients, they had to really rethink their SKU portfolio about what they were going to make and what they weren't.

The CEO of the company spent four hours at Walmart one day watching what people bought in terms of candy corn. He came back and said, "We've put all this effort into 27 flavors of candy corn," and they really only want one or two flavors. They did a SKU rationalization, and P&G said that they did something similar during COVID, but they've continued post (pandemic).

Then the last little anecdote was a retailer I spoke with recently said, there's just less emphasis for them on trying to get same day and next day delivery. Because what they found is, yes, some customers want it, but most customers don't care. As long as they get it within a reasonable period, long setup for this. A, how is complexity impacting supply chain? B, are you seeing companies rethinking? Do we need all those SKUs? Do we really need to do next-day delivery?

Yossi: For good or for worse, I've been around this for some time now. Every recession, we call it the accordion phenomenon. Every recession, companies lower, the number of SKUs get less, and it does several things. First of all, it makes sure that these best-sellers are on the shelves all the time. Second, it reduces cost because you have less changeovers in manufacturing. You see, during the pandemic, I think I gave an example in my book, what is it?

The people who make a Progresso Soup got down from 90 varieties to 30 or 40 varieties. Interesting, you mentioned Procter and Gamble. We interviewed the CEO of Procter Gamble at MIT. He said something, this goes to the age of how tough it was and how heroic the effort was during the pandemic. Procter and Gamble makes Tide, which is the number one detergent. They did not have all the ingredients every month. They had different ingredients. They were reformulating Tide continuously based on the ingredients that they have. They had to do the quality and make sure that people would not notice what's going on, that the quality was still there. They were doing it. The fact that during the pandemic people, first of all, go back to your point, reduce the number of SKUs.

Every time there's recession, there's pressure. Then marketing takes over and it starts to market the, "Oh, maybe we'll do a red one," or maybe we'll do one that goes backward" or whatever other variety is there. It goes back. That's just a historic view. This is a general trend. It happens all the time. Not necessarily the pandemic, the pandemic was a big shock, of course, and it led to changes in demand and all this. Okay, my prediction is that it goes back the way it was because marketing will take over and start coming up with new varieties and new ways in order to get market share, in order to increase revenue on the margin.

It's interesting you talk about complexity, because in very few companies when you increase the variety, increase the number of SKUs, you increase complexity and you actually increase the cost. Very few companies can withstand what I call the marketing pressure, can do a real analysis. Okay, we have 30 varieties. What will be the cost for all of them? Not, or for the 31st one. Okay, we'll add the 31st one. The usual analysis is marginal analysis or incremental analysis. What will be the cost of adding a new one? Very few times people look at what will be the added cost on the end, on all the 30 other varieties that now we have to make 31.

That's harder to measure, and very few people do. The complexity increases for the entire portfolio, not only for the one, not only the added SKU. Complexity is hard to measure, actually. There's no specific measure of complexity. It pops up in so many places. As I say, coming back to the point where there's pressure, cost pressure in particular, people reduce complexity, but then it creeps back up.

Abe: Speaking of that, Yossi, AI and other technologies specifically robotics, they seem to be on the top of mind. In almost every conversation that we're having with supply chain leaders, AI is in the title of your book. From your perspective, how is AI and automation transforming supply chains today?

Yossi: Yes, today, to some extent, a lot more in the longer term, let's say 5 to 10, 15 years, right now, we see robotics, of course, every warehouse operator is putting new robots in the warehouse. That's a trend that started just before the pandemic and during the pandemic going through the roof. Warehouses are putting robots at the huge clip. In terms of AI in general, we forget that AI is already here, in the sense that we use it, we don't even realize we're using it.

When we talk to any customer service function of a company, it used to be “press one if you want this, press two if you want this.” Now, you just speak this AI on the other side that picks up what you're saying and tries to find a solution. We are not even excited about it. It's no big deal because we're used to it. There's AI in a lot of the things you do. I don't know if you guys use a program called Grammarly, which fixes your writing, the AI that fixes your writing. It's not quite ChatGPT, but it fixes existing writing and it also gives you comments as you write, like a souped-up Microsoft Word.

There's a lot of these things that are already happening. Some things that are interesting maybe and more in my wheelhouse. I wrote a lot about resilience and risk management. I know some companies, few, I must admit at this point, they try to develop risk management strategies. Part of this is finding out what is the financial health of critical suppliers. The traditional way to find out was looking at Dun & Bradstreet and other financial statements. These are backward-looking. They're at least one and usually two-quarters behind. It's a back-looking view. It doesn't help you to find out what's going on right now, and then you can do it.

People are looking at this. You can do only so many and companies may have tens of thousands of tier-one suppliers, let alone deep-tier suppliers, companies are using large language models in order to look at older suppliers. For each supplier, look at the social media, regular media, try to find out if they are executives leaving, if there are lengthening terms of payment to their suppliers, if they have some failed project and if their deliveries are not 100%. This gives an indication that right now something is going awry in this supplier. The response will be to send a person there to find out what's going on, to actually inspect it. This is something that could not be done without a large language model that can look at tens of thousands of suppliers and look at a lot of information on each one of the suppliers and of course, huge computing power to do all this. These are examples, there are many others, of course, in manufacturing in many other areas.

Bob: Yossi, when we talk about technology, I think to some degree automation and robotics, but certainly the discussion around AI, we've had congressional hearings on the threat of AI. From your perspective, do we need to fear AI or automation in general in the supply chain?

Yossi: The discussion on this is wide-ranging, and when you say in general, there are people on extremes, on both sides. There are people like Bill Gates, Yuval Harari, who wrote in The Economist a very interesting article thinking that ChatGPT and similar programs can be the end of democracy, the end of the human race. On the other hand, there are people just as credentialed who think that this AI will usher in an age of plenty and will have new medical solutions, new ways of growing plants, new ways of fighting climate change, everything will be great. As always, I think the truth is of course in the middle, but it's hard to find out what's right.

A lot of the angst, by the way, is about job losses. I must say that of course, nobody knows the future. When nobody knows what's going on, when we do forecasting, every type of forecasting, we look in the past. A large part of my book, or I don't know, part of my book is looking at the various industrial revolutions and documenting the fact that people always thought, this time it's different. Well, it was not. Every time, few jobs were lost, but many, many more jobs were created.

You can deal with all the angst about job losses. The US is at, what is it, 3.5, 3.6% unemployment. Historically low unemployment and lots of positions open despite all the headlines about companies shedding off some people. There's still very low unemployment in the United States. It's hard to justify the high level of anxiety. In some sense, it is understandable because we know the people who might lose their job. We look at the supermarket when you have the automated teller machines basically, customers are checking out themselves and say, "Okay, some of these people may lose their jobs." We don't know yet the new industry and the new jobs that will be created.

One example, with technology, it usually increases jobs in the same industry when implemented technology, but it also created new industries. An example is Ford, in the beginning of the 20th century, put in the assembly lines. Before that, there were about several thousand people working for Ford, while building one car at a time. Teams of artisans, basically building one car at a time. They put the production line, the assembly line, and the number of employees of Ford went from a few thousand to 150,000 in the heydays of the model team.

This was not the main impact of the assembly line. The assembly line created cars that were a lot more affordable. Because cars were more affordable, people build highways, hotels, motels, restaurant. A whole industry with millions of jobs developed. The whole hospitality industry, developed. This was not Henry Ford's intent when he started putting in the assembly line. He changed the jobs of the people who worked for Ford. The number of jobs increases, but then there were adjacent industries that were created almost as a side effect. This is what is so hard about predicting what will be the jobs in the future. Clearly, we know there'll be more jobs in technology, but there'll be a lot of other jobs, very hard to predict.

Abe: Yossi, among the topics that we talked to supply chain leaders is about their digital transformation which seems to have taken on an extraordinarily, high focus for a lot of organizations. Yet what we're hearing is that the transformations really aren't delivering the results that they anticipated. From your perspective, is automation or technology enough to take our supply chains to the next level of performance?

Yossi: You have to look at the past. It helps a lot. What is it? Robert, who was the Nobel laureate economist from MIT, who wrote during the third industrial revolution when computers were coming in, software. He wrote, "You can see computers everywhere, but not in the productivity numbers." Then it took time and we started seeing productivity. I think it's the same thing.

There's a lot of new technology and new automation right now. We did not find the secret sauces of how to make people work better with computers, how to make the combination work better. We still have a workforce that is not used to working with high levels of automation, and it takes time to train a whole workforce as jobs are changing and as even management is not sure how the jobs are changing and when it is best to integrate people with automation. We see a lot of trial and error.

I visited the plant of Mercedes in Germany. The plant has people, in the plant, people holding iPad-like devices and running robots. Then, to my surprise, talk to them, they reduce the number of robots and put more people in. "What's going on here? Everybody's digitizing and you go the other way." Say, "Yes, the digital is too rigid. We now build on the assembly line, EVs, and hybrid and internal combustion engines with all. Germans are ordering cars the way they like it with all the options. It is too volatile. We need people who deal better with the volatility than any algorithm that we can develop."

It takes a lot of trial and error and see-- There are various ways of integrating people into automation. One way is that people are in the loop. For example, the workers in Amazon warehouses, they are sitting in one place. The bins are coming to them. They do some operation pick up and the bins go away and a new bin is coming. They're in the flow. Then there's the other model when people are monitoring and are not in the flow.

For example, when you talk to your customer service organization and there's talking words and somebody else and the AI basically answers. The AI can deal with simple problems at one point, either you get frustrated, or the AI realizes that it doesn't know the solution and it kicks it up to a human. Here's a combination when the AI works on the simpler problem and there's the human coming in to work on the more complex problem. When you have to understand how simple the problem can be.

I remember years ago being in a Dell call center and realizing that almost half the calls where my computer doesn't work, and then the customer service representative will say, "Well you have to plug it into the electrical outlet and then you have to turn it on." This is half the calls. You don't need people to man such a customer service center. This can be done by computers, by AI. That's what's going on now. There are several models of AI working with humans. Some of it we all understand. Some of it is still trial and error so I agree. For the most part, we're still trying to figure it out. In general, it'll take a lot longer to see the productivity numbers out of digitization.

Bob: Yossi, you just really tackled what I was going to ask you next, which was about the future of work, because supply chain at the end of the day is still a lot about getting something from point A to point B. There's still a lot of mechanical, where manual processes. Just taking what you just talked about, if you were summing up, how do you see the future of work in the supply chain evolving?

Yossi: Talk about supply chain management. There is work in the warehouse and driving the trucks, and there's work in planning, managing, monitoring supply chains. In fact, one of the biggest angst of people is that the new AI may start displacing managerial jobs, white-collar jobs. Not only the robots may be replacing blue-collar people. Not sure. First of all, when activity becomes less expensive, for example, because it requires less people, there's more of it. Example, in the early 70s, ATMs were getting to be widely used. There were about 300,000 tellers in the United States and everybody thought that there'd be no more tellers in 10 years. Turns out today, there are over 600,000 tellers.

Why? Because opening a bank branch became so much less expensive when a lot of the work is done through the ATMs or online that it became very inexpensive to open a branch. I don't know about any other places, but when you walk to the street of Newburg Street in Boston or any other downtown street every third storefront is a bank. There's now a lot more bank branches. There are a lot more people working in this. We talked before about for example, in adjacent areas when so something developed with one does not expect and of course, there's a whole issue of technology.

Before, Google searches we never had the people optimizing Google, Google ads for example. There's a lot of news, new professions in tech. When we talk about supply chains specifically, we see there will be a lot more automation on the floor so to speak. It's not going to happen quickly. It'll take much longer than people realize. Example from other fields: In 1892, AT&T developed the automatic telephone exchange. Until then, there were mostly women sitting there and plugs into hall connecting people. By 1950, there were still 350,000 telephone exchange operators. Only by 1980, the job basically disappeared, or it means it took nine decades of innovation to go through the through the economy. Things are a lot slower than people think and the evolution is a lot slower for a variety of reasons.

For example, unions are in many calls in the way. Regulations are in many cases in the way. For example, Italy disallowed the use of GBT. We know, the US, EU, other countries are talking about regulating modern AI. Then there's the issue of acceptance. Today's airliner, whether it's a 787 or A350, can basically go gate to gate with no pilots. Very few of your audience will go on a metal tube that flies at 35,000 feet across the Atlantic when there's nobody in the cockpit, even when it's monitored from the control room down on earth, right? It's, you need somebody in charge the same way. We'll see what the impact or the acceptance of people going on the highway with autonomous trucks with no drivers running behind them at I don't know, 80 miles an hour.

That can be disconcerting. We'll see if people will accept this or there'll be a similar fight to what the railroad we're mounting against long combination vehicles with all the scary heads of mothers seeing trucks coming onto them. So there's a whole issue of social acceptance of a lot of this innovation. As I said, some of it, already here. We use Google Maps. It has AI built into it. Looking at the-- when it gives you the, what will be the congestion close to your destination, half an hour from now it actually uses some forecasting and try to put some forecasting together using basically machine learning to try to get you a good time and try to choose the right route. It happens in the background, we don't even think about it.

How do I see the future of work in the supply chain evolving? I would say the number one thing is slowly. I don't see too many jobs disappearing. I see the nature of jobs changing, just like the nature of jobs in Amazon Warehouse is now different in the sense that the items are coming to the picker rather than picker go to the item. Okay, it's changing but still, there are 1.2 million warehouse workers for Amazon. Jobs may be changing. There'll be more working with technology, but I don't see big changes happening fast. Over time, yes, we'll see some changes.

Abe: Yossi, last question. Obviously, you talked to a lot of supply chain professionals and leaders out there. What are the two or three things that you're telling them to pay attention to? We've gone through a whole host of challenges and opportunities. How do they prepare? What are you focusing on? What should they be focusing on?

Yossi: It's really, talking about supply chain leaders more business leaders in general. One of the new challenges that I see that has nothing to do with technology is the polarization of the workforce and the community. Political polarization. How should companies respond? We saw all the cases, whether it's Target or Disney or the beer, Bud Light, how should the company respond when in the community they're in, or when there are pressures from certain groups when you know that whatever they do, they'll alienate the other half of the population?

We are bifurcating society. How to deal with it, that to me is one challenge. Companies used to think that they better avoid any political stand. They may not have the pressure to take a political stand and with the downside that comes with it of alienating the other side, whatever stand they take. I am not sure how this is going to serve. Companies will have to find ways of dealing with it.

In terms of focusing attention on what comes next, specifically in the supply chain, because of the pandemic and the Russian action in Ukraine, people are focusing on resilience. It is in fact, for many companies taking the attention away from sustainability and fighting global warming. Because we always say that the one thing, the one constant is change. Today, it's more so than ever. When companies are even putting in software, they're putting in software that can help them reformulate and redesign and replan the supply chain much faster than before.

The ability to put in parts or materials from new suppliers and make sure that they can be qualified at a much faster rate. Building in these capabilities is what many companies are focused on, which is the result of not only the pandemic, it's the pandemic, it’s the political situation, it's geopolitical situations, the fear about China. All of this is creating unease, uncertainty and the need to prepare companies are actually, many of them are engaging to justice.

Abe: Yossi, really tremendous information. There's always a lot more to unpack and follow up. Unfortunately, that is all the time that we have but special thanks to our guest Yossi Sheffi. Take a look at his new book, The Magic Conveyor Belt. I'm also pleased that is going to be joining us and presenting at our annual event, ASCM in Louisville again on September 11, 13th. You can find out more about that at Finally, a special thanks to you for joining us on this episode of The Rebound. We hope you’ll be back for the next episode. For the Rebound, I'm Abe Eshkenazi.

Bob: I'm Bob Trebilcock.

Abe: All the best everybody. Thanks.

Yossi Sheffi will deliver a keynote, The 5th Industrial Revolution: Why Human and AI Collaboration is the Future, at ASCM CONNECT 2023: North America in Louisville, Kentucky on September 12, 2023. 

Use of Cookies

We use cookies to personalize our website’s content and ads, to provide social media features and to analyze our traffic. We also share information about your use of our site with our social media, advertising and analytics partners who may combine it with other information that you’ve provided to them or that they’ve collected from your use of their services. You consent to our cookies if you continue to use our website.