00:00:08 Introduction to the quantitative-qualitative paradox in supply chains.
00:00:33 Challenges in measuring savings with a quantitative approach.
00:02:00 Paradox between highly quantitative methodology and the need for qualitative understanding.
00:03:10 The role of subjective judgment in evaluating numbers and measurements.
00:06:00 The importance of qualitative understanding in using data to improve businesses.
00:08:01 The challenge of comparing pre- and post-covid supply chain situations.
00:10:22 Emergency measures taken during the crisis and evaluating supply chain performance.
00:12:57 The complexity of supply chain problems and the frustration of not having a single number to represent the solution.
00:14:10 The danger of overemphasizing technical solutions and the importance of interpretation in supply chain management.
00:15:53 How often supply chain practitioners should revisit decisions and the iterative process of qualitative thinking.
00:17:38 Adopting the “Day One” mindset from Amazon’s philosophy to tackle problems with a fresh perspective.
00:19:12 Utilizing qualitative approaches to improve problem-solving in supply chain optimization.
00:21:36 How number analysis can help make better choices for a company’s battles.
00:22:31 The importance of understanding the interplay between qualitative and quantitative methods in supply chain management.
Summary
The interview between Kieran Chandler and Joannes Vermorel, founder of Lokad, delves into the complexities of supply chain optimization and the need for a balance between quantitative and qualitative approaches. Vermorel stresses the importance of understanding and making sense of the numbers through qualitative judgments, as supply chains involve complex systems with numerous potential points of failure. He suggests that businesses should engage in data analysis to build an intellectual model of what optimality means for their specific supply chain, leading to judgment calls that are not purely quantitative. The discussion also highlights the challenges of simplifying complex problems and the importance of serving customer demand better.
Extended Summary
In this interview, Kieran Chandler, the host, speaks with Joannes Vermorel, the founder of Lokad, a software company specializing in supply chain optimization. They discuss the mix of quantitative and qualitative approaches in supply chain management and the challenges in balancing the two.
Vermorel explains that Lokad’s quantitative supply chain approach is often met with skepticism and questions about specific dollar savings. While the company’s methodology is highly quantitative and relies on advanced statistical tools and a programming language dedicated to predictive optimization, Vermorel admits that it is difficult to provide simple, precise measurements of savings. He highlights that it is often misleading to promise a guaranteed dollar amount of savings, as the complexities of supply chains require a more nuanced understanding.
Although Lokad’s approach is rooted in quantitative methodology, Vermorel emphasizes the importance of qualitative judgment in making sense of the numbers. He explains that every measurement, even in a highly quantitative setting, is subjective to some extent, as supply chains are complex systems involving people, machines, processes, and software. Consequently, eliminating personal judgments entirely is nearly impossible.
Vermorel illustrates the subjectivity of measurements by discussing the turnover metric. He points out that even though turnover is a quantitative figure, it is influenced by various factors, such as promotions, discounts, and stockouts, which are subject to qualitative judgments. Similarly, when looking at stock levels, there are multiple factors to consider, including lead times, reliability of suppliers, and the risk of obsolescence. All of these factors require qualitative assessments.
Chandler and Vermorel also discuss the difficulty of predicting the future and the importance of qualitative judgments when dealing with uncertainties. Vermorel notes that even with the most advanced predictive tools, there will always be unknowns that require human intuition and experience. This is especially true for supply chains, where complex networks and interdependencies create numerous potential points of failure. While quantitative methods can help identify risks, qualitative assessments help prioritize and address them effectively.
Finally, Vermorel highlights that supply chain optimization is not simply about reducing costs but also about achieving strategic goals. He stresses that companies should consider qualitative aspects such as brand image, customer satisfaction, and long-term growth potential when making supply chain decisions. In this context, it is crucial to balance quantitative and qualitative assessments to ensure optimal outcomes.
The discussion between Kieran Chandler and Joannes Vermorel revolves around the complexities of supply chain optimization and the need to balance quantitative and qualitative approaches. While Lokad’s methodology is heavily quantitative, Vermorel emphasizes the importance of qualitative judgments in understanding and making sense of the numbers. Supply chain management requires a mix of both approaches to address the inherent uncertainties and complexities of the field and achieve strategic goals.
Vermorel explains that Lokad started with the classic forecasting approach, producing a single number for demand on a daily, weekly, or monthly basis. However, they soon realized that this approach lacked consideration for uncertainty. As they delved deeper into the numbers, they found that there were many nuances and subtleties at play. This led to the development of more advanced numerical analysis tools and the need to make judgment calls on which paths to take in optimizing supply chains.
The host brings up a quote by William Edward Deming, which says that without data, you’re just someone else with an opinion. Vermorel agrees but adds that being reliant solely on data can be naive. He argues that the idea of engineers simply crunching data to come up with an optimal solution is absurd because the concept of “optimal” is highly subjective. Instead, Vermorel suggests that businesses need to engage in data crunching to build an intellectual model of what optimality means for their specific supply chain. Paradoxically, this often leads to judgment calls that are not purely quantitative.
When asked for an example, Vermorel discusses the challenges that clients have faced during the economic downturn caused by the COVID-19 pandemic. Businesses experienced less sales, excess inventory, and significant shifts in their sales channels. In this context, clients have asked Lokad to prove that their supply chain management has improved compared to the previous year. Vermorel points out the difficulties in making such comparisons, as the pre- and post-pandemic situations are significantly different.
He explains that, in order to compare performance, businesses need to make high-level judgment calls that cannot be based solely on numerical data. For example, they need to consider factors such as store closures and the shift from in-store to online sales. Additionally, supply chains had to be shut down and restarted in response to the crisis, further complicating the comparison.
The interview highlights the importance of both quantitative and qualitative analysis in supply chain optimization. While data is essential for understanding the intricacies of supply chain management, judgment calls and a more holistic approach are necessary for making effective decisions. This is particularly relevant in times of crisis, when businesses must adapt to unprecedented situations and make comparisons to previous performance with care.
They discussed the importance of a quantitative approach in supply chain optimization, while acknowledging the necessity of qualitative judgment. Supply chains should be uneventful, with no remarkable events or bottlenecks, and the right balance between conflicting economic forces. Vermorel emphasizes the need for an iterative, qualitative review of the system rather than a constant quantitative review, suggesting that companies should focus on fixing the forecasting engine instead of manually reviewing forecasts. The discussion also highlights the challenges of simplifying complex problems and the importance of understanding and serving customer demand better.
Vermorel discusses the importance of balancing qualitative and quantitative approaches in supply chain optimization. He highlights the concept of “day one” thinking, which emphasizes a fresh perspective on problems. Vermorel suggests that while quantitative methods can provide valuable insights, it is crucial to focus on the problem definition qualitatively. He emphasizes that numbers can help identify the right battles for a company, ultimately leading to better results. The back-and-forth between quantitative methods and qualitative thinking can expand the horizons of supply chain directors, helping them approach their supply chain in innovative ways.
Full Transcript
Kieran Chandler: Hey, at Lokad, the mix between a quantitative approach that builds upon data and a qualitative one that focuses on business decisions has often been seen as somewhat disruptive. As such, today on Lokad TV, we’re going to look at this balance and understand which is more important in somewhat of a chicken and egg scenario. So, Joannes, our topic today is all about that quantitative-qualitative paradox in supply chains. It’s a bit of a mouthful, so what’s the idea behind this?
Joannes Vermorel: The idea is that, at Lokad, we are pioneering something referred to as the quantitative supply chain approach. There are many people who challenge us and say, “You have a quantitative approach, so how much is my company going to save in dollars one year from now if we do that?” The answer is actually very difficult because you’d say, “Well, you have an approach where you measure everything in dollars and optimize that, so why can’t you have a very simple measurement of the savings in dollars?” The answer frequently turns out to be a lot more complicated than that, and it requires quite a lot of qualitative understanding to make any sense of any quantitative figures that we can give. Usually, the naked figures, if I say, “Oh, your company is going to save $10 million, guaranteed,” are completely misleading.
So, the paradox is that although we have a highly quantitative methodology with tons of highly numerical tooling and technologies that are literally just crunching numbers at scale, doing advanced statistics, it’s highly quantitative. Lokad is a programmatic platform, so people are doing programming, which is again crunching numbers with a fairly advanced tool, a programming language dedicated to the predictive optimization of supply chains. It’s deeply quantitative in terms of day-to-day operations, tooling, and tech. And yet, what I’m saying is that usually, the only way to make sense of that is to make qualitative judgments, which is a bit of a paradox for a methodology that prides itself as being highly quantitative.
Kieran Chandler: People watching this might find that a bit surprising. I mean, you wrote an entire book on the quantitative supply chain. So, what is it about a qualitative judgment? What are the sort of things that we should be looking at there?
Joannes Vermorel: One of the most important things is to understand that every number, every measurement that you can make, is highly subjective. You would think that you can have something completely rational, objective, and independent, where you have eliminated all sorts of personal judgments, but the reality is that it’s nearly impossible to do that. Supply chains are very complex and include plenty of people, machines, processes, and software. They live inside an ecosystem that you barely comprehend, so at best, you can make judgment calls even when you want to assess things like turnover. There is a great saying in…
Kieran Chandler: Can you tell us more about the business philosophy behind your company, Lokad?
Joannes Vermorel: The philosophy is reflected in the business saying that “turnover is vanity, profit is opinion, cash is king.” It means that depending on what you look at, you’re either looking at something that is very real or something more fuzzy, to be interpreted in different ways. That’s the crux of our products.
Kieran Chandler: Can you elaborate on that?
Joannes Vermorel: As we were developing tools for more quantitative analysis and data crunching, we progressively realized how much depth there was, for example, in having better and better forecasts. When Lokad started, we had classic forecasts that gave us one number for the demand on a daily, weekly, or monthly basis. But we realized that we were missing something; there was something entirely absent from our methodology, and that was uncertainty. Our classic perspective was incorrect because we were dismissing uncertainty altogether. So, as you go deeper into the numbers, you see that there are tons of nuances and subtleties at play. This requires judgment calls, which are highly arbitrary.
Kieran Chandler: So, what you’re saying is that without having a bit of a qualitative understanding, there’s a limit to what you can do with data?
Joannes Vermorel: No, not really. You can be under the naive impression that data will unlock some kind of business improvement, but that’s not the case. Without data, you will not even have the tools, and the reality will not even challenge you enough so that you can even build an actual and an intellectual model of what optimality really means for your business. To get proper ideas on how you should even think about your supply chain, you need to do this data crunching, and then once you’ve done that, you will have the tools to make judgment calls.
Kieran Chandler: So, Joannes, you’ve talked a lot about the quantitative approach that Lokad takes to supply chain optimization. But I imagine that there are still some qualitative judgment calls that need to be made. Can you tell us more about that?
Joannes Vermorel: Yes, absolutely. Even though we try to rely on quantitative data as much as possible, there are still judgment calls that are completely not quantitative. You see, that’s a quirk. And the illusion would be that you can bypass a judgment call to just keep it quantitative all the way.
Kieran Chandler: Okay, that makes sense. Can you give us an example of a situation where qualitative steps are necessary?
Joannes Vermorel: Sure, let me give you an example. Clients for example, right now in this period, ask us, “Tell us, you know, you’ve just deployed Lokad. We are fortunate we have still managed to close clients during the first half of 2020. And they look at the situation, they ask, ‘Prove us that we are doing better compared to last year.’” And obviously, considering the absolutely horrific economic situation we are in, I mean everything is worse, you know, they have less sales. If they are in the fashion business, most of the winter season didn’t, I mean they are left with tons of stock from the winter season. It’s quite bad, it’s quite dire in many situations. And I believe that even in this situation, we are able to improve things. But compared to what? Can we compare the pre-COVID situation to the post-COVID situation? Does it even make sense? How do you compare when you have, like, retail networks that have closed a third of their stores, or when basically e-commerce has doubled but the stores have lost half of their demand? It’s a different world, it’s not entirely different, but it’s quite different. And so when it comes to comparing the numbers, I would say you need to make judgment calls. You can present the number, you can reason about them, you can try to align them as much as you can, you can try to say, “Here are all the biases that we are we try to account for.” But the reality is that the numbers that you present in the end, you need some high-level judgment. You cannot just say, “Oh, it’s just compared to two numbers, and be done with it,” and tell me how many millions of euros or dollars you’ve saved, for example. What we had to face during this crisis was there were many clients where we had to basically take emergency measures so that their supply chains would have a clean shutdown. That was very complicated to orchestrate and completely unprecedented. And then we had to do the reverse thing, two months afterward, where we had to have like an emergency, I would say, restart of the supply chain. And then, how do you judge whether Lokad did perform correctly? And I believe that our quantitative approach makes it actually reasonably straightforward to make a judgment call. But make no mistakes, there is a judgment, a very educated judgment call involved, and it’s completely qualitative. For example, for the restart of the supply chain, did we have, you know, how many panic situations did we have? Did we have suddenly things that were accidental bottlenecks, or on the contrary, did we kind of keep things smooth and, I would say, and uneventful? You know, the uneventfulness property of your supply chain.
Kieran Chandler: It may sound like something that is typically undervalued, you know, a quality of a very good supply chain optimization process is that there is no remarkable event anymore. And it’s very hard to put, I would say, a numerical to put that in numbers. You know, when there is nothing remarkable that happened just because everything is operating smoothly without, you know, much fuss, just as it should. And that’s a very, very important quality of our, I would say, of a well-worn supply chain. Things are going very smoothly, and there is nothing really remarkable, you know. There is no massive stock, there are actually fewer, as little stock as you can possibly dream of, and there are very little, you know, stockouts again, as little as it kind of makes sense to have. And overall, it’s very uneventful, but just sometimes those clients not get a little bit frustrated because you’re focusing on this kind of interpretation of events instead of focusing on kind of the numerical results. Is Lokad not somewhat maybe overstepping their mark there?
Joannes Vermorel: Yes, I mean, it’s frustrating because it’s not easy, you know. You would like to have something that would be like easy, or just give me this number and be done with it. And when you look at the problem, and we come back and say, “Well, you know, we have at least 10 angles to this problem, and we need to balance them,” because supply chain is mostly about balancing conflicting economic forces, you know, cost of stock, cost of higher production, cost of higher, you know, more flexibility in terms of suppliers, cost of having shorter lead times versus having more stock and longer early times. I mean, all those things are kind of in conflict. So when we look at this problem, and people say, “Oh, we would like this problem that has at least 10 dimensions to be collapsed to just one number and be done with it,” we say, “Well, you’re missing the certainty, and the society is exactly what makes your business great.” And it’s not always possible to just collapse something that is very, very complex into a simplistic perspective.
Kieran Chandler: And I’m assuming you’re not saying that you should seek complexity for the sake of complexity, just to look like you know, because that’s something that there are many actors that in this industry are doing. They are just putting, I would say, an overemphasis on super technical solutions with way too many numbers just to look impressive and to look savvy, you know, to look like they know a lot more than you do.
Joannes Vermorel: That’s not the point. Usually, we say we have to spend a lot of time on the interpretation to even, for example, qualify whether a number is worth looking at, and that’s worth a discussion. Are we, you know, yes, it’s always possible for any single dashboard to add one more KPI, but is it really worth it? And, again, that’s a qualitative discussion about, you know, numerical results. Usually, those are the and the surprising thing is that whether a KPI is worth looking at depends on the numerical values exhibited by this KPI. And again, how does a KPI, how does the numerical value of a KPI qualify as being surprising? This is again not a numerical statement.
Kieran Chandler: You talked about numbers and layers of thinking. Could you explain this paradox that people often have about it being only about numbers?
Joannes Vermorel: Yeah, a qualitative statement, and that’s where I would say there is this paradox that people think when you say it’s only about numbers, it’s not. It’s about having its numbers with more layers of thinking on top of it.
Kieran Chandler: I see. And you mentioned the idea of these layers of thinking growing and becoming an iterative process. So, for a supply chain practitioner, how often should they be revisiting these decisions? I mean, they can’t waste their energy constantly reviewing. Are they looking qualitatively at the right things?
Joannes Vermorel: I would say they should be constantly challenging that, and the quantitative review of the system is pointless. When I see companies who say, “Oh, we are very, very number-driven, you know, we have people who manually review our forecast every single week,” I say, “Why do you need to manually review your forecast every single week? That’s a complete waste of time.” They say, “But the forecasts aren’t good otherwise.” If the forecasts aren’t good otherwise, maybe you should fix the forecasting engine that produces forecasts in the first place so that you don’t have to manually fix your forecast all the time. And be done with it. They say, “Oh, but we have tried for years, and we never managed to use that with our vendor.” Then I would say, “Well, pick another vendor.” Look, for example… But I digress. Back to the thing. You see this quantitative repeat work of doing quantitative review. I believe it’s a complete waste of time on the contrary. Revisiting always the question of what does it mean? What customer demand means? How can I actually better serve my clients? What does it mean to serve them better? Is there an angle that I have not seen? And those are questions that are worth revisiting all the time. And by the way, I believe that is one of the pillars of the Amazon thinking, is this idea of Day One. If you read the memos that Jeff Bezos gives to the entire world, you would see that one of the key ideas is basically one of the tenets of the Amazon way of doing business is it’s always Day One. Meaning that you should always try to look at the problems as if it was Day One, as if you’re starting fresh and you think, “How can we revisit from everything that we’ve done that has been done as if it was like the first day we are doing it?” So that we can have a completely fresh perspective. And I believe that that’s the exact opposite of people iterating over the forecast, you know when they quantitatively tweak their forecast. This is a quantitative perspective where they are going to do the same thing over and over, and it’s not the good perspective. The good perspective is something that is much more qualitative where you revisit and you ask yourself, “Okay, let’s restart from zero. What is a client asking yourself those very fundamental question and maybe you will come to the same answer, but if you, by any chance, come to a different answer…”
Kieran Chandler: Sorry, you might end up realizing that maybe you need a completely different answer. If you revisit what is a client and then you realize that you didn’t have the right mindset about what is a client, then maybe all the solutions that you’ve engineered to better serve what you thought was your client are probably completely off compared to what they should be. That’s a very tricky process. Interestingly, part of the paradox is that usually, if you ask yourself those questions in a vacuum, chances are that you will not come up with any new ideas.
Joannes Vermorel: And part of the paradox is that by looking at the numbers, they can give you the inspiration to get the spark that gives you a better way to look quantitatively at your business. That’s exactly the sort of things that, for example, we have managed to do at Lokad with those probabilistic forecasts. When we started looking at those probabilities, we realized there are so many things we had not even realized. We have not even started to scratch the surface on the problem, and we only came to this realization once we started to look at the probabilities, while in the past we were just looking at the forecast assuming that there is just one number for the future.
Kieran Chandler: Previously, you kind of spoke about the idea of falling in love with the problem and not just with the solution. So you’re saying that by taking more of a qualitative approach, that means it improves the way you approach the problem?
Joannes Vermorel: Absolutely. There is nothing more qualitative than a problem definition. It’s not the numbers that define the problems. But again, back to the point, numbers can really help you think better about the problem. When you have a problem statement, you need to make choices. Everything is not equally relevant for your company. Aesthetics is very important for many brands, and for luxury brands, aesthetics could be everything. If you’re selling mining equipment, well, maybe not so much. Does that mean that your equipment can look terrible? No, I would say, in a very kind of nerdy mechanical way, mining equipment can actually look good.
Kieran Chandler: Pretty cool, but clearly it’s not the same emphasis. You see, that’s why you have a problem statement. You also have to make a statement about what is not relevant to your company. You know, you need to make a parsimony of things that you pick your battles wisely. And again, when you do those numbers analysis, they can help you make the right choice of battle for your company so that you can have this supply chain optimization that picks the things where you can really have better results, ultimately better dollars. But beware, between the day you start looking at the problem and six months to one year down the road, you will not be looking at the problem the same way. And thus, it becomes incredibly difficult to do dollar-to-dollar comparison because your perspective on the problem has changed. Okay, if we start sort of concluding today, and definitely the difference between this qualitative and quantitative kind of paradox, it’s a little bit subtle. So why is it something that is important to discuss?
Joannes Vermorel: I believe it’s important because I would say, I believe, and we have 10 years of business to support this belief at Lokad, that quantitative methods bring better results for supply chain. But the crux is that better quantitative methods unlock new ways of thinking about the supply chain, which is something completely qualitative. And you see, there is this back and forth: you have better numbers, you can think of the problem differently. And thus, I would say for supply chain directors, I would say get familiar with those numerical methods just because they will expand your horizon on the way you can even think about your supply chain. That’s maybe counter-intuitive, but it’s a back-and-forth. And go to quantitative methods not for the numbers, but for the insights those methods will open on plenty of other things about your supply chain.
Kieran Chandler: Okay, brilliant. It’s been a bit of a mouthful, but I think we made the whole episode without getting the two mixed up. So that’s everything for this week. Thanks very much for tuning in, and we’ll see you again next time. Thanks for watching.
