00:00:00 Category shifts during the pandemic and introduction of guests.
00:01:14 Explanation of Estée Lauder’s Center of Excellence and its role in strategic transformations.
00:02:12 Differences between supply chain resilience and firefighting and their implications.
00:05:01 The concept of “poly crisis” and the importance of building resilience for long-term planning.
00:07:11 Discussion on various classes of shocks and their impact on supply chains.
00:09:54 Discussing the complexity of supply chain management and vendor incentives.
00:10:56 Phase transition in supply chains and the impact of post-pandemic disruptions.
00:12:33 Importance of supply chain resilience and adjusting for market shifts.
00:14:07 Using digital technologies to align demand and supply during disruptions.
00:16:36 Automation in decision-making for better efficiency and addressing blind spots.
00:19:47 Discussing process automation and its role in reducing mundane tasks.
00:20:24 Importance of strategic automation and scenario planning in anticipating future disruptions.
00:21:28 Gauging effectiveness of resilience initiatives and the difficulty in measuring preparedness.
00:24:01 Vendor perspective on selling resilience solutions and the temptation to overpromise.
00:27:00 Importance of humility and skepticism when assessing the potential of technology in achieving supply chain resilience.
00:29:54 Discussion on the need for intentional preparedness and stress testing in global supply chains.
00:30:41 How to mimic systemic events like COVID to test supply chain resilience.
00:32:33 Importance of long-term financial perspective and intangible aspects in supply chain resilience.
00:35:02 Accuracy in high-dimensional forecasts and simulation tools for supply chain resilience.
00:37:01 Importance of a 3-year time horizon for supply chain resilience and market growth.
00:39:27 Discussing the challenges in forecasting and planning for the future.
00:40:37 Lengthy software upgrades and their impact on supply chain management.
00:41:49 The role of human intervention and automation in supply chain resilience.
00:43:10 Cultivating a culture of resilience through employee selection and training.
00:45:01 Complexities of projecting supply chain states and managing disruptions.
00:49:53 Making strategic investments in technology for resilience.
00:50:07 Closing remarks and appreciation for the guests.
Conor Doherty interviewed Joannes Vermorel, founder of Lokad, and Jay Koganti, VP of Supply Chain at Estée Lauder’s Centre of Excellence, about supply chain resilience. They discussed anticipating disruptions, strategic planning, and investing in technology, data analytics, and employee training. Vermorel emphasized stress testing and accurate forecasting in simulations, while Koganti highlighted long-term thinking and training in resilience. Both agreed on the need for a cultural shift towards a mindset of action and resilience, using technology like automation and simulations to prepare for potential scenarios.
During the interview, Conor Doherty, the host, discusses supply chain resilience and firefighting with Joannes Vermorel, founder of Lokad, and Jay Koganti, Vice President of Supply Chain at Estée Lauder’s Centre of Excellence. Koganti explains that he has been with the company for 17 years and has been passionate about supply chain digital transformation for the past two decades.
The Centre of Excellence was formed about eight years ago to sustain improvements from a large enterprise transformation on an ongoing basis. The center focuses on strategic transformations for the company, including digital transformation, new planning systems, and new distribution centers.
Koganti distinguishes between resilience and firefighting, explaining that firefighting is heroic and crisis-driven but is exhaustive and unsustainable. Resilience, on the other hand, involves responding to a shock or crisis and recovering quickly. It is more structural, systematic, and sustainable. Vermorel adds that while having employees who are willing to save the day is a positive sign, many emergencies could have been anticipated and avoided.
The pandemic led to significant category shifts, with people staying at home and using more skincare products instead of makeup. This situation highlights the need to focus on long-term strategies and anticipate potential disruptions several years in advance. Both Koganti and Vermorel emphasize the importance of strategic thinking and resilience in supply chain management.
The focus of supply chain resilience should be on preparing for the shocks that have the potential to disrupt the entire supply chain, rather than just addressing the shocks as they happen. This requires strategic thinking and planning, anticipating potential risks and developing contingency plans. In order to build a resilient supply chain, companies need to have a long-term vision and invest in new technologies, data analytics, and training of employees.
Jay Koganti, as the Vice President of Supply Chain at Estée Lauder’s Center of Excellence, emphasized the importance of digital transformation in building a resilient supply chain. By adopting new technologies and data analytics, companies can have better visibility and control over their supply chain, and respond quickly to disruptions.
Overall, building a resilient supply chain requires a proactive approach, rather than just reacting to the shocks as they happen. It requires investing in new technologies, data analytics, and employee training, as well as having a long-term vision and strategic planning. By doing so, companies can not only survive but thrive in an increasingly complex and unpredictable business environment.
During the interview, the host, Conor Doherty, asked Joannes Vermorel, the founder of Lokad, and Jay Koganti, Vice President of Supply Chain at Estée Lauder’s Centre of Excellence, about the challenges of supply chain optimization. Vermorel identified two broad categories of supply chain disruptions, namely, internal problems and external shocks. External shocks can include problems with suppliers, transport, or government intervention, whereas internal problems can include defects within an organization that become magnified and affect the entire system. Vermorel also noted that supply chains are complex systems and that even minor dysfunctions can lead to major problems due to the tight coupling of different areas. He added that vendors may invent problems in convincing ways, creating even more complexity in the landscape. Koganti agreed with Vermorel and added that companies need to prepare for disruptions by creating buffers and investing capital in resilient approaches. He gave an example of a drug maker that had buffers in place and was able to meet the demand for a particular medicine during a pandemic. Koganti also discussed how digital technologies could help companies realign demand to supply during category shifts. Vermorel noted that predictive optimization technology faces challenges in dealing with disruptions that do not exist in historical data. Finally, Vermorel cautioned against overestimating the capabilities of modern statistics in dealing with deviations from historical data.
In a recent interview, Conor Doherty spoke with Joannes Vermorel, the founder of Lokad, a software company that specializes in supply chain optimization, and Jay Koganti, Vice President of Supply Chain at Estée Lauder’s Centre of Excellence. The discussion centered around building resilience in supply chains. Vermorel explained that statistical approaches are widely used to make decisions in supply chain management, especially when operating at scale with hundreds of sites and products. He also pointed out that statistical technology cannot anticipate or mitigate crises, but automation can free up teams to focus on thinking about the blind spots of automation. Koganti agreed, adding that automating mundane activities frees up intellectual capacity and that strategic automation can be used for scenario planning to prepare for potential disruptions. When asked how to gauge the effectiveness of resilience initiatives, Koganti explained that it is difficult to measure, but time to respond and time to recover are important metrics to consider. The discussion ended with the acknowledgment that assessing muscle resilience is challenging but important.
difficult to simulate a systemic event like COVID-19 in a vacuum to measure how resilient a supply chain is. However, companies can adopt stress testing methodologies similar to software vendors to prepare for peak disruptions. It is important to be intentional about building resilience into the cultural DNA of a company and not just focus on efficiency. Being resilient means being able to face detrimental situations, survive and recover quickly. As a technology vendor, it is important to be realistic about what technology can and cannot do to build resilience in the supply chain. Humility is key in approaching solutions, as the stakes are high and involve the survival of companies. The fragility of supply chains can be increased significantly through technology alone, and while it can help with resilience to some extent, the stakes are broader than just the supply chain. Overall, the aim should be to reduce the potential vectors of weakness or weakness vectors in the supply chain, not necessarily to make it stronger.
During the interview, the host, Conor Doherty, asked questions to Joannes Vermorel, founder of Lokad, and Jay Koganti, Vice President of Supply Chain at Estée Lauder’s Centre of Excellence, about supply chain resilience. Joannes emphasized the importance of stress testing and simulation to prepare for unexpected disruptions in supply chains, such as market shutdowns due to COVID-19. He also mentioned the challenge of adopting a financial perspective that considers long-term investments and the value of intangible assets, such as goodwill.
Jay highlighted the significance of a three-year time horizon in assessing the resilience of supply chains, as most actions, such as setting up new factories and DCs, can be achieved in that timeframe. Joannes pointed out that thinking decades ahead is necessary to prepare for major crises that have a probability of occurring around 4 percent. He emphasized the need for accurate forecasting in simulations, as an illusion of resilience could lead to arbitrary inaccuracies. The discussion centered around the importance of anticipating and preparing for unexpected supply chain disruptions through stress testing and simulation, as well as the challenges of adopting a long-term financial perspective and accurately forecasting future events.
been happening in the real world. During the interview, Conor Doherty, the host, talked to Joannes Vermorel, founder of Lokad, and Jay Koganti, VP of Supply Chain at Estée Lauder’s Centre of Excellence, about supply chain optimization and resilience. They discussed the challenges of projecting the future state of supply chains, expressing intent, and the need for automation and process automation. Vermorel talked about the difficulty of projecting the future state of a supply chain from a high-dimensional perspective. He also discussed the need to express an intent and the lack of mathematical tools to do so. Koganti emphasized the importance of intentional long-term thinking, strategic thinking, and capital allocation. He also talked about the need for people to be trained and upskilled in resilience, through simulations and experimentation. They both agreed on the need for a cultural shift towards a mindset of action and resilience. Koganti also discussed the possibility of shorter lifecycles with incremental improvements, rather than decade-long projects.
In this interview, Conor Doherty, the host, interviews Joannes Vermorel, the founder of Lokad, and Jay Koganti, Vice President of Supply Chain at Estée Lauder’s Centre of Excellence, about supply chain optimization and resilience. Vermorel emphasizes the importance of correctly modeling complex supply chain scenarios and answering basic management questions in order to avoid programming errors that could lead to incorrect conclusions. He also notes that there is a lack of mathematical instruments in the literature that can provide the desired properties for the supply chain of tomorrow. Koganti emphasizes the importance of being intentional about building resilience into the supply chain and using technology, such as automation and simulations, to create real proxies for potential scenarios. He encourages investing time and energy into making resilience a backbone of cooperation.
Conor Doherty:Welcome back to LokadTV. I’m your host Conor, and as always, I’m joined by Lokad founder Joannes Vermorel. Today, we’re going to talk about supply chain resilience while firefighting, and to help us with that, we’ve invited Jay Koganti to join us. He’s the Vice President of Supply Chain at Estée Lauder’s Center of Excellence. Jay, welcome to Lokad.
Jay Koganti: Thank you.
Conor Doherty: Um, so, Jay, just introducing myself, you know, being with the company for close to 17 years, I do a lot of strategic transformations with the company, and supply chain digital transformation is my passion and my work for the past two decades. Very excited to be here. Thank you. And, just to get things started, I’m just curious in terms of the Center of Excellence, what exactly is that? What do you guys do there? What are your duties?
Jay Koganti: Yeah, I think the Center of Excellence, we formed almost, uh, now 8 years back. We came out of a large enterprise transformation, then. I think our initial intent is how do you really sustain the improvements on an ongoing basis? It could happen in a large corporation, you do one big transformation, then you’re back to normal. But we want to make sure there’s continuity and continually innovate. So, as part of that, we formed the Center of Excellence. We do strategic transformations for the company, whether it is a digital transformation or whether it’s new planning systems or new distribution centers. You know, a lot of those things come from the Center of Excellence.
Conor Dohert: Yeah, so it sounds like you’re the perfect person to ask this question to then. In terms of resilience and firefighting, when you were explaining that to other people as part of these transformations, what exactly do you tell them in simple terms?
Jay Koganti: Yeah, no, I think, intuitively, a lot of people will think, you know, firefighting is resilience. I mean, the way, clearly, the distinction is, even in America, we call it whether it’s firefighting or diving catches. It’s like saving the day, right? There’s a crisis that comes in, someone just heroically jumps in and saves the day. But we know, it’s exhaustive, it’s not sustainable. It’s a one-time deal, whereas resilience comes from how do you really respond to a shock or a crisis and then recover fast to the original phase. It’s more structural, more systematic. That’s a fundamental difference because you kind of firefight every day, but resilience is something you could operate on an everyday basis.
Joannes: My thoughts? Yes, I mean, the interesting thing when I think about resilience is that among our clients and companies that operate in the supply chain in general, there are tons of very, very accidental emergencies you see. So, yes, it’s perfectly fine to have in the company people that are willing to go to great lengths to save the day. It’s a very positive thing, it means that people care. But the where I would say it may reflect a reality that is not so good for the company is that what I see is
Joannes Vermorel: I’m sorry, let me start over. There are things that could have been avoided in the first place, and should not have required any kind of heroic gesture to be fixed. The thing could have been solved, and especially, I mean, if you look at it, we are a software company. So, we are interacting first and foremost with the IT plumbing of the companies. IT-wise, I see systems that face one crisis per day, sometimes several, and for reasons that should have been avoided in the first place. Both compatibility glitches in the transition process from one version to another and due to the fact that those supply chains are very complex. So, you have tons of people, obviously, but you also have tons of software, tons of layers. You end up having, at least on the software side of the picture, tons of very man-made problems. Problems that do not reflect any kind of grand challenge. It’s not a natural disaster; it’s a man-made disaster, and it’s very accidental. The thing that is quite interesting is that I see that for let’s say, in terms of a disaster, like a warehouse taking fire, people go to great lengths to avoid those sorts of accidents, but there are other areas where incidents keep happening, and it seems that companies learn very slowly from their mistakes. They do learn, the situation is getting better, but it looks like it’s a fairly slow process. Also, I think you could say in the past two, three years, there are a lot of major disruptions we all know, COVID, post-COVID recovery rebound. There are also a lot of small disruptions. It’s a poly crisis, as someone coined it. There are always multiple crises emerging, amplifying. They feed on each other. There’s a recurring theme for most of the supply chains in the past year or so. That’s where resilience is becoming a big topic. How do you really build it? And as you pointed out, I don’t think this is about accidental incidents and responding to it. It’s more structural to me. You have to almost think in these things. You need to plan and anticipate and prepare in a strategic horizon. You need to look at a two to three-year horizon. How do you prepare yourself for this kind of poly crisis situation?
Conor Doherty: That actually leads perfectly to something I did want to talk about, and Joannes, I’ll come to you first about this because when we talk about resilience and firefighting, it seems like that’s the response. So, resilience is in response to something. Firefighting is certainly in response to an event, and those events are, I believe, shocks. And there are multiple classes of shocks, and not all of them are existential threats to one supply chain. So, as Joannes first, and I will come to you, Jay, could you expand a bit on the classes of shocks, and which ones are the ones that we should really focus on from a supply chain perspective?
Jay Koganti: The categorization of the shocks is very difficult because there are so many of them. Broadly speaking, we could say there are demand shocks where your customer, for one reason or another, disappears, either because they cease to be interested in your products or because they can’t afford to or because they go to a competitor. That would be on the demand side. On the supply side, there are a bunch of different kinds of shocks, like natural disasters, labor strikes, geopolitical unrest, port closures, pandemics like COVID-19. There are also technological shocks like cyber-attacks or failures of critical IT systems. Finally, there are financial shocks like currency fluctuations or credit market
Conor Doherty: So, can you talk a little bit about the different types of shocks that can occur in a supply chain?
Joannes Vermorel: Supply side, there are shocks that prevent you from serving your customers. That can be because your factory doesn’t operate properly, your supplier doesn’t operate properly, or you can’t transport the goods. So, there are two broad classes of shocks. But you can have tons of other shocks. For example, it can be a problem with litigation, or it can be a new regulation that just prevents you from doing things the way you were doing it previously. Even during the pandemic, the problem was the lockdowns, and in this case, you have a heavy-handed government intervention that just decides to shut down your business. Maybe it’s in the general interest, but the result is that your business is shut down from your supply chain perspective, or one of your critical suppliers, and that creates a problem for you.
Joannes Vermorel: So, just a categorization, but as we have seen, for example, with Southwest Airlines, sometimes it’s just a software problem that puts your company to a halt. Sometimes it’s something that is just purely internal, which reflects your own organization or some defects of your organization that become magnified to the point where the company as a whole grinds to a halt. Just because, again, supply chains are systems. When something starts to function, it tends to ramify and propagate into other areas. When you have a company that is heavily optimized, things tend to be very tightly coupled, and so the points to pay for efficiency is that when something starts to dysfunction on one side of your system, that might ramify and impact plenty of other systems. You could easily decouple everything, but you would lose economies of scale and plenty of other things.
Joannes Vermorel: So, the bottom line is that you’re faced with plenty of things that can potentially go wrong, and even vendors like myself, Lokad, a software vendor, are prone to invent problems because that’s even more things that you can sell. So, you have these problems where it’s not just difficult in the sense that there is so much diversity, it’s that there are plenty of people in the space who have a direct incentive in making things up because if there is a new problem, then it means that you can potentially solve a solution. I don’t dismiss the fact that there might be problems. I’m just saying that it makes the landscape even more complex when you have people that can make up problems in very convincing ways.
Jay Koganti: Do I think those are all spot on? My view is there’s a lot within the supply chain. I believe there’s a big phase transition. So, in terms of post-pandemic, we could see a lot of structural disruptions continue to happen. And I think that’s going to be at least in my view, it’s going to last for the next couple of years. Whether it’s inflation-related macroeconomic challenges, a place going out of business, the energy crisis we are seeing, and all of those things, having an amplifying effect. But also, I think a lot of companies are not prepared for it. Whether it is a major disruption, I’ll give you an example. Even recently, in the U.S, there’s a trifecta of the viruses going on.
Conor Doherty: So Joannes, can you tell us about your experience with supply chain optimization during the pandemic?
Joannes Vermorel: Sure, Conor. During the pandemic, we saw a resurgence of viruses, and people were running out of basic medicines like cough syrup. So, I thought, why not prepare for such situations? I talked to a friend who said that only one drug maker, Claritin, was prepared because they had buffer capacity to support the manufacturing. This made me realize that we need to think about how to put buffers and invest our capital in the right place to create resilience in the supply chain. This is not the strongest point historically for supply chains, as we are all driven by efficiency.
Jay Koganti: Yes, Joannes is right. We need to prepare and anticipate by creating buffers, so we can approach things in a more resilient way. During the pandemic, we saw major category shifts, with people using more skincare products while staying at home, and more makeup when going out. This presented a challenge of how to run down the manufacturing assets for one category and ramp up production in an agile way for another.
Conor Doherty: And how did you deal with this challenge?
Jay Koganti: Initially, it was a bit of a surprise, but we quickly realized that we needed to do it on an everyday basis manually. For a large company with 300-400 assets globally, we had to leverage digital technologies to understand how the demand was shifting and realign it with the supply. This required a systemic approach that we have built and are now ready to use for any future category shifts or rebounds.
Conor Doherty: And Joannes, can you tell us about the challenges faced by companies like Lokad in predicting and optimizing supply chain during such events?
Joannes Vermorel: Yes, Conor. One of the key challenges is that the things that call for resilience capacity do not exist in the data. This is especially true for Black Swan events or events that deviate from the norm. Modern statistics are not able to extrapolate very far, and we need to be careful not to overestimate what they can do. So, the challenge for companies like mine, which primarily leverage historical data, is to build predictive optimization technologies that can anticipate such events and redirect resources accordingly.
Conor Doherty: So, Joannes, can you talk a little bit about supply chain optimization and the role of data in that process?
Joannes Vermorel: Sure, Conor. So, in supply chain optimization, data plays a crucial role. However, the data involved is very mechanical, and there is no intelligence in it. Even if something seems obvious to humans looking at the process, it may not be apparent from the data. The data only shows cyclicities and patterns that are more of the same. This is why statistical approaches are still widely used, especially when operating at scale, with hundreds of sites, products, and thousands of SKUs.
Jay Koganti: I agree with Joannes. When operating at such a large scale, we need mechanization to help us make decisions, because it’s not practical to employ thousands of clerks to manage inventory levels manually. However, this approach has its limitations, because the data may not reflect what is happening in real-time, and it may take weeks for the data to catch up with the actual situation.
Joannes Vermorel: Exactly, Jay. At Lokad, we address this problem by automating mundane decisions, such as those related to manufacturing, and other areas of the supply chain. This frees up time for people to chase what our numerical recipe is not going to capture. People need to have the bandwidth to think about the blind spots of our automation, and what is not even captured by our numerical recipes.
Conor Doherty: So, you’re saying that automation can actually help teams to spend more time thinking about things that are not captured by the data?
Joannes Vermorel: That’s right, Conor. Automation can actually free up time for people to think about the blind spots of our numerical recipes. This is crucial, because no statistical technology or data-driven technology can anticipate or mitigate all shocks. However, by liberating teams, we can give them the luxury of being able to think long and hard about those crises that are happening, and apply corrections based on their human insights on top of the system.
Jay Koganti: I completely agree. Automation is not a substitute for having smart and dedicated people who can spend time thinking about those crises happening and what it would mean to do something that is in the interest of the company over the long term. Acting in the interest of the company long term is a difficult problem, and there are many paths and potential decisions that may be unclear.
Conor Doherty: Jay is that aligned with how you would approach building resilience, a mixture of automation but with management overseeing it?
Jay Koganti: Yeah, I totally agree on the process automation part. Any mundane activity, automating gives a lot of bandwidth. I think that’s definitely a big part of the firefighting conversation. If you’re doing everyday firefighting, your intellectual capacity is exhausted, you’re not focusing on the big things. So I think definitely, that’s how I see a lot of digital transformation really focus on. But I will add one more dimension also. There is a big opportunity to strategic automation around scenario planning, for example. It’s like a forest and trees, right? So you could do a lot of trees every day based on small, mundane activities. Also, how can you anticipate two or three years ahead in terms of what could be a peak disruption looks like? What could be major demand shifts that could happen, right? What could be scenarios that really prepare you as a company, whether it’s new distribution centers or new factories? Our demographic shifts and all those things also need a large degree of automation. You can’t really run hundreds of Excel sheets and try to simulate those things. It’s very hard to do it. I think one thing we all need to invest, and we have done a lot of investment in that area, is definitely the scenario planning capabilities. So really, you could prepare and build a structural resiliency as part of your organization.
Conor Doherty: Well, then that leads to my next question, which is essentially how do you gauge the effectiveness of one’s resilience initiatives? So, for that question, presume that a shock has not occurred. There’s been a sustained period of normal activity, but you want to know, well, am I spending my money wisely? Have I actually erected a sustain a resilient business model?
Joannes Vermorel: Yeah, I think that’s a very hard question, right? I mean, a lot of clients, you don’t know what’s going to come and hit you, right, and how you’re going to respond, right? It’s very hard to anticipate that. Also, just one other anecdotal thing, you know, a lot of companies struggle with this because we don’t write capital allocation for this, right? So you rather want to spend money on a market growth, you know, web promotions. You don’t want to put an extra buffer capacity because it’s, you know, you don’t know, is that going to be really useful or not, right? So I think the point about how do you measure is one of the measures we believe is very important is time to respond, right? When the crisis comes in, whether it’s a smaller or a bigger one, how fast you’re responding, right, whether it’s a people technology, all of that. The second is very important is time to recover or rebound. It’s not only responding to it, how you’re going to bring it back to the original state or original phase. I think these are the two important metrics. Again, it’s very hard to do for every small crisis, but for bigger crises, definitely, we look at it, you know, “Hey, this crisis came in, you know, when did we respond, right? When did we get into the steady state phase?” You know, that’s a good measure of your systems and people, how prepared you are. If that makes sense?
Conor Doherty: Absolutely, absolutely. And by the way, when you say it’s difficult because it’s to assess this sort of muscle resilience and why I found it…
Conor Doherty: It’s very intriguing and quite funny that many of my peers, I’m speaking about enterprise software vendors, largely evade the question. From a vendor perspective, it is such a good thing. You’re selling something, and if it doesn’t work, the company goes bust and disappears. So you end up with the witness, and if it works, then you can claim that it was thanks to you. So you see, as a vendor, there’s this sort of asymmetry where if you fail, nobody will notice your insulation. And yeah, you’re completely insulated. But if you succeed, then you can actually take the claim that it was thanks to you, at least in part, which makes, you know, at least for me, I’m trying to really figure out whether these sorts of things work or not. But indeed, it is very, very difficult. And for a vendor, it is very tempting to just pretend that you have something that works because due to this asymmetry, even if it’s not the case, you are largely insulated from the fact that whatever you’re proposing is just not working.
Joannes Vermorel: But you see, the bottom line is that the way I approach that is a bit like, you see, there are some broad topics that are just incredibly difficult and evasive to approach, like happiness at work. You know, those sorts of things are incredibly hard to just go straight on. So the way I approach that, I’m thinking of what is it that I might be doing that might be making the situation worse, and let’s try to avoid that. So it’s a much more modest ambition, you know? It’s not about making the clients more resilient. It’s starting by, what can I stop doing or stop selling that would just stop making the client company even more fragile than it is? Indeed, you were mentioning the dimensioning of the buffers. If, as an inventory optimization service, if you go for the leanest of the leanest, with a super short-sighted perspective, you can actually give the impression to your client that you’ve said an enormous amount of working capitals. And if you look at, for example, software companies, they all claim that they have reduced inventory by 30 or even 50. But if twice a decade, due to that, your entire supply chain blows up, was it wise? Yes, on the short term, you were saving a lot of working capital. But if you twice a decade, you blow up your entire car now of Goodwill with all your clients just because you have, like, a major short Edge, it might not be such a wise thing to do. But as for the vendor, you know, the technique is “Take the Money and Run.”
Jay Koganti: So, if I can just summarize that before I come back to you, Conor, because I want to make sure I understood that, is it your position that you want to start resilience is essentially reducing the potential vectors of weakness or weakness vectors, excuse me, not necessarily the idea of making you stronger? It’s just removing the areas of weakness in the supply chain?
Joannes Vermorel: No, I believe that the metrics given by our guests were actually quite spot on. It’s your capacity to face something that is very detrimental, survive the detriment, and then go back to whatever was the initial situation. So I think that’s pretty much resilience, you know? Don’t die and then recover, and ideally, do it fast. But the challenge is that literally when there is a shark, it’s not just the supply chain. You know, finances all over the place.
Conor Doherty: Marketing is always a place for sales, etc. So it’s obviously a company-wide problem. And the way I see that is that, at least from our position of being a vendor, there is some humility to be had. Because if you promise a cure and you don’t deliver, it’s very bad. You know, it’s the Hippocratic Oath: first, do no harm. And so, as a technology vendor, you have to be realistic in what your technology can do. I’m very skeptical that through technology alone, you can make a supply chain resilient. I would say not even close by a long shot. But what I can see is that through technology alone, you can make the supply chain pretty fragile. That, I’m very convinced about. The other way around, fragility, you can go really, really far with technology. Resilience, to some extent, technology can help. But I believe that to be very resilient, the stakes are much higher and much broader than just supply chain. That would be my take. And that’s where I think we have to approach this, at least from a solution standpoint, with humility because the stakes are super high. Because, at the end, we are talking about the survival of companies. So, as far as those entities are concerned, stakes can’t be possibly higher.
Jay Koganti: Yeah, no, I think those are all good points. To me, it’s spot on. But also, I think it needs to be intentional. It needs to be part of a cultural DNA. It won’t automatically happen. At least I think you need to be intentional in terms of how you put the right capital investment because a lot of times, supply chains are all about efficiency play – how well you run your working capital, your service, and all those things. A lot of times, these are secondary thoughts and about the buffers and all that. It’s wasted capital. That’s how people think about it. But given the global shift, whether it’s re-globalization, reconfiguration of the supply chain, resource constraints, or climate issues, if you add up all those things, I think for the next decade or so, we need to be very intentional about it. Preparedness also comes from stress testing. One of the things I personally believe in is how you stress test your system for a big disruption.
Conor Doherty: That’s what I was going to ask you about.
Jay Koganti: Yeah, so I think that’s a big part of this as well.
Conor Doherty: Perfect segue, actually. How exactly, if I can just follow up on that with Jay, how exactly does one mimic or create, in a vacuum, a systemic event like COVID, for example? How does one create that in a vacuum, subject our supply chain to that, and then measure how resilient we are? I mean, I can’t conceive of that, so I’m just curious. How would that work?
Jay Koganti: Yeah, I think the best way, I mean, I don’t think we do very well in global supply chains. That’s part of the issue, right? Or, you know, pandemic and post-pandemic. I think probably the best companies that really do that are software vendors. So they have this traditional stress testing methodology that prepares for peak disruption. At least, I think, parts of the supply chain are realizing and trying to build capabilities and adopt these methodologies and techniques, for example.
Conor Doherty: It’s very hard to mimic a major disruption like COVID, right? You can’t bring your business to a standstill and do the testing. But what you could do is anticipate, right? If a major DC shutdown occurs, your supply lines are cut off. What are your alternative routes where the inventory is flowing from? If the demand has a major shift, if the market is completely shut down, how do you respond to that? Those things you don’t need to necessarily just be reactionary. You could actually anticipate, right? And you need to simulate, you know, simulate and see whether you’re ready or not. That kind of stress testing, I think we all should invest, definitely.
Jay Koganti: Well, that actually builds on a point that you said before. Yeah, and here I would have, I mean, bouncing on what you just said, two things come to my mind. First on the financial take, and then on the simulation for the stress testing purposes. On the final stake, the interesting thing and the challenge that I face, you know, look at, as a general rule, I strongly advocate for a financial perspective. But historically, those sort of approaches had, I would say, a bad press because, especially in the 80s and 90s, there were even plenty of Hollywood movies where you would see that the villain has some kind of incredibly short-sighted financial approach where the villain of the story does terrible things with an incredibly short-sighted perspective. You know, splitting companies in tiny bits just to make a little bit with a very, very short horizon. So I believe that, indeed, nevertheless, if we want to be really efficient, we have to count those dollars, but it means to count the dollars with a long-term perspective, which is a very difficult exercise because suddenly you can’t just trust your usual, what I call, the first-order financial instruments, you know, what appears in your books. If you’re Estée Lauder, for example, you have a globally known brand. It took decades to build that. The value is to a large extent, both completely intangible, but also very real. And so, if we want to count those dollars, it means that there are plenty of things that will never appear in the books, but they are still very real and they need to be accounted for. So, that’s, I would say, that my answer is that on the financial part, the main challenge that I have is to bring my prospects and clients to adopt this financial perspective, but that includes tons of things that are both very rational but very made up. That’s also, that’s a paradox, you know, because you’re thinking literally decades ahead. So, there are things that are made up in the sense that if we’re looking at the Goodwill, you know, there is no scientific measurement of the Good Will of clients, especially if you think two decades ahead. Nevertheless, it is very important, so it needs to be done.
Joannes Vermorel: And then back on the sort of simulation stress test here, um, I see that from a vendor perspective, one thing that interests me is that, for example, when you do a simulation, you’re doing a forecast, you know, of a kind. And the interesting thing is that when you’re doing this High dimensional forecast, what is your accuracy? Look at I have some very, I would say, we have developed some, um, some interesting techniques about how what does that mean that your simulation is even accurate?
Conor Doherty: So, we have a lot to unpack here. Let’s start with Joannes. Joannes, you mentioned earlier about the accuracy of supply chain simulations. Can you explain why this is a non-trivial problem?
Joannes Vermorel: Yes, Conor. The accuracy of supply chain simulations is a very difficult problem because it’s not a one-dimensional thing. It’s not just about having the accuracy of one element. You have to project the future state of the supply chain, which has tons of interdependent factors. So, it’s a very tricky question to answer.
Conor Doherty: I see. And you also mentioned that most people who promote these technologies don’t question the underlying accuracy of those simulations. Can you elaborate on that?
Joannes Vermorel: Yes, Conor. If you don’t have dedicated mathematical instruments to evaluate the accuracy of these simulations, you don’t even know if they’re accurate or not. So, you could be arbitrarily inaccurate and have an illusion of resilience, which is even worse. You may have an illusion of resilience based on tools that have made up forecasts with no assessment of accuracy whatsoever.
Conor Doherty: That’s a good point. Jay, earlier you mentioned a time horizon of two to three years when talking about resilience. Can you explain why that’s significant and how you measure resilience in that time frame?
Jay Koganti: Yes, Conor. When we look at market growth, sales opportunities, demographic shifts, and other indicators, we need to have a much longer horizon. But when it comes to supply chain, we can pretty much do everything within a three to four-year horizon. We can set up a new factory, a new DC, new lanes, etc. Also, it’s very hard to predict longer horizons because the possibilities are too many. So, a three-year horizon is a good enough indicator for us to think about capital allocation and assets.
Conor Doherty: I see. Joannes, you mentioned thinking decades ahead. Can you explain why that’s important and how it’s related to probabilistic forecasting?
Joannes Vermorel: Yes, Conor. The paradox is that if you think about major crises happening, let’s say four times a century, and we take the 20th century, which had two World Wars plus other events, we have to approach this by thinking about probabilistic forecasting. It just…
Conor Doherty: You know, the resilience comes from that action mindset, orientation, how do you respond to it in the most agile way? That is also a cultural thing, right? You have to really create that cultural DNA, and it takes a long time.
Jay Koganti: Well, there are two points there. If I can just jump in, I will ask you Joannes as well, but when you talk about inculcating or cultivating that sort of DNA, as you call it, how do you go about that? Because again, it is very much a mind and a cultural shift.
Jay Koganti: I think one, you have to be choiceful about the people you bring in, right? That’s one thing. The other thing is how do you upskill them? A lot of the preparedness comes from actually simulation. You ask the question, “Hey, your Distribution Center is going to shut down for the next week, how would you go about it? Tell me.” So, that’s the kind of things you’re really training and upskilling people to think about. These kinds of things could happen, right? So you’re really building resilience as part of a curriculum for the teams. That’s what I meant by, you have to create that very intentionally, whether it’s training, simulation, or even thought-related experimentation. You have to do all of them very intentionally.
Joannes Vermorel: And the way I can bounce back on how we can approach resilience, again, from my technologist perspective, which is a very narrow viewpoint, is that there are some really super tough challenges. Literally, the literature of the sciences is very much lacking. That’s what drives me and what we’re researching. The first thing is, literally, if you want to project the future state of a supply chain from a high-dimensional perspective, it is an incredibly difficult problem to even express what accuracy means.
Joannes Vermorel: So, you have this very basic problem, which is, if I want to project one product, that’s a time series perspective, I know, and the literature is full of metrics. There are plenty of instruments to tell you whether your projection is accurate or not. If you start playing this game with literally the state of a complex system, this is very, very difficult. And here, there is literally no instruments that I’m aware of that let you do that in a clean manner. So, there is literally a fundamental problem in the sense that we are even lacking the very mathematical tools to do that. So that’s the first class of problems.
Joannes Vermorel: The second thing is, if we want to introduce a shock, that means that we want to be able to express an intent and some sort of disruption that will arrive in the system. So, we have first, we want to project the system into the future, but on top of that, we want to be able to express an intent. And here comes the question of, can we do that expressly? Syntax, by the way, when I say “express this intent,” will come through some sort of programming tools. You will program your behaviors, specify the disruption, and the question will be, does it give you any kind of correctness by design to properly capture your intent?
Joannes Vermorel: I know it’s a very abstract question, but just take a look at the problem through the lenses of the vendor. I’m supposed to build a technology that can project the state of a system, that’s already very difficult. And then I’m saying that if you do some sort of “what if” on top of that
Conor Doherty: …that our “what if” simulation was reflecting truthfully what would have happened for something that never came to pass, and thus that it captured correctly the intent of the person that was expressing this scenario. It is a very difficult problem, but a very interesting way, and there are tools that can make those exercises more correct by design.
Joannes Vermorel: What’s really interesting to me is that supply chain professionals have to deal with a lot of problems, so they have to be able to potentially find answers to a lot of concerns that could be raised by management. The way I approach this is that, yes, we want to have a multi-dimensional perspective where you can project the future state of the supply chain, but you also want to be able to answer tons of frequently basic questions: more facilities, less facilities, other locations, cannibalization, disruption by a competitor. It’s not actually fantastically complicated questions, but the way I approach that is, I want to be able to craft tools that give some degree of correctness by design, in the sense that you can avoid some very basic programming mistakes. Because, again, one of the dangers that I see is to gain false confidence. You do something, you say, “Oh, that’s okay, we did the ‘what if,’ everything is fine,” but actually there was a bug in the simulation and everything is not fine, and you just concluded the opposite of what you should have concluded. So, that’s the sort of thing that I would say keeps me up at night, thinking how can we have those tools? And again, these are very open questions in terms of mathematical instruments. There is not much known in the literature that gives the sort of properties that we would seek for the supply chain of tomorrow.
Conor Doherty: Well, gentlemen, thank you. I’m mindful of our guests’ time, so I think, as his customer, we’ll give the last word to Jay. Is there anything you’d like to share with the audience, any resilience call to action that you’re dying to get off your chest?
Jay Koganti: Yeah, resiliency is already a big boardroom conversation right across the supply chain. Naturally, the pandemic really amplified this topic. I think just be intentional about it. There’s a big opportunity on the technology side. I definitely see, even as we talked about the mindset, the capital investment, but also there’s a big play for the technology. Technology could really help in terms of whether it’s automation or even the simulations, making it a real proxy for what could happen that could even give a mental resilience. So, I think just be choiceful about what technologies you choose and really spend a deep investment of your time and energy to make resilience the backbone of the corporation. That’s all I could say.
Conor Doherty: Well, on that note, gentlemen, I think I’ll draw things to a close. Joannes, thank you very much for your time. Jay, thank you very much for yours. And thank you for watching. We’ll see you next time.