00:00:08 Introduction and welcoming Nicholas Vandeput.
00:00:34 Nicholas’s current work and teaching at university.
00:01:00 Discussing Nicholas’s book “Inventory Optimization Models and Simulations.”
00:03:51 The book’s target audience and simplification approach.
00:05:15 The shift from IT departments to practitioners and supply chain maturity.
00:08:02 Python’s expressiveness and advantages over other programming languages.
00:10:19 Emphasis on Python in the book and the benefits of hands-on learning.
00:11:47 Popularity of Python and its contenders, such as Java and JavaScript.
00:13:51 Transition from Python 2 to Python 3 and its impact on academia.
00:15:10 Coping with uncertainty in Python and potential for language improvements.
00:16:03 Bargain price for trying new inventory optimization software.
00:16:18 Common confusions in inventory optimization and safety stock formula.
00:17:32 The importance of review period in safety stock calculations.
00:18:34 The need for better engineers in supply chain management.
00:19:57 Comparing supply chain problems with more complicated fields like microelectronics.
00:21:02 Importance of problem formulation and attracting brilliant minds.
00:22:32 Limitations of academic models and their application in the real world.
00:23:01 Moving from mathematical models to simulations for better accuracy.
00:24:01 Limitations of safety stock formula and the need for simulation.
00:25:07 Conclusion and mention of Nicolas’s book on inventory optimization.


In the interview, Kieran Chandler hosts a discussion with Joannes Vermorel, Lokad founder, and Nicolas Vandeput, author of “Inventory Optimization Models and Simulations.” They address the simplification and accessibility of programming in supply chain optimization using Python. Vandeput’s book offers simplified inventory optimization models, while Vermorel emphasizes the ease of implementing these models. They discuss the evolution of supply chain optimization, the importance of programming expressiveness, and Python’s advantages. Both experts acknowledge the limitations of traditional mathematical models and emphasize the need for alternative methods like simulations to handle real-world challenges in supply chain management.

Extended Summary

In the interview, Kieran Chandler, the host, introduces Joannes Vermorel, the founder of Lokad, and Nicolas Vandeput, the author of “Inventory Optimization Models and Simulations.” The discussion revolves around the simplification and accessibility of programming in supply chain optimization, with the use of tools like Python.

Vandeput shares that his book aims to simplify inventory optimization models and make them accessible to supply chain practitioners. The book focuses on providing simple numerical recipes in Python to address real-world situations, rather than diving deep into idealized supply chains. Vermorel highlights that these recipes can be implemented in just a few lines of code, demystifying what large software vendors offer and questioning the value they provide.

The book is designed for supply chain professionals who struggle with inventory management, aiming to help them understand and optimize their stock levels. Vandeput emphasizes the importance of making assumptions and understanding the limitations of the models used, rather than seeking perfection.

Vermorel explains that supply chain maturity has evolved over the decades, with companies initially struggling to establish a digital counterpart to their supply chains. After decades of progress, businesses can now manage their stock levels digitally without any intelligence, enabling them to focus on optimization.

The shift in mindset, with supply chain practitioners becoming more involved in programming, is attributed to the increasing accessibility of tools like Python and the simplification of processes. This change has allowed practitioners to take more control over their supply chain optimization, rather than relying solely on IT departments.

They discuss the evolution of supply chain optimization, the importance of programming expressiveness, and the advantages of using Python.

Vermorel highlights that it has taken four decades for ERP systems to reach maturity, which has allowed supply chain optimization to become a viable option. The conversation moves towards the importance of programmatic expressiveness, as it allows organizations to adapt to real-world changes and unpredictable events. Vermorel explains that Excel provides a certain level of expressiveness, but its limitations stem from the way logic is organized in spreadsheets. Python, on the other hand, offers a more abstract and expressive solution, making it ideal for supply chain optimization.

Vandeput then explains why Python is his language of choice for the book. He points out that Python is one of the most popular programming languages globally, with an abundance of resources available online. This means that users can find answers to their questions easily, making it less likely for them to get stuck. Additionally, Python’s simplicity makes it more approachable, and users can understand the code by reading it.

The book emphasizes Python for two reasons. First, Vandeput believes that practicing is crucial for learning, so he includes many do-it-yourself sections in Excel and Python, allowing readers to gain hands-on experience. Second, Python enables users to scale solutions for supply chain optimization, moving beyond solving problems for individual items and addressing broader supply chain issues.

Vermorel agrees with Vandeput’s points on Python’s popularity and simplicity, but he also acknowledges some of the language’s limitations. He suggests that other programming languages, like JavaScript and Java, are not as suitable for supply chain optimization due to the extensive software engineering skills required to work with them.

The conversation revolves around the merits of Python as a programming language and common misconceptions in inventory optimization.

Vermorel argues that Python is well-suited for supply chain optimization due to its concise nature and ease of use, especially for newcomers. He notes that Python’s evolution from its creation in the 1990s to the present has made it increasingly popular and effective for addressing academic and industry needs. Despite its advantages, Vermorel mentions that there is still room for improvement, specifically in handling uncertainty. However, he asserts that Python is a cost-effective solution compared to other expensive options in the market.

Vandeput, on the other hand, delves into the common misconceptions within the industry regarding inventory optimization. He points out that many practitioners often confuse lead time with transportation time and overlook the importance of the review period when calculating safety stock. Vandeput emphasizes that the review period must be considered in addition to lead time, and suggests that reducing this period can lead to a reduction in safety stock.

Both Vermorel and Vandeput acknowledge the prevalent confusion in the industry and the need for better-educated professionals in the field of supply chain management. They stress the importance of understanding the nuances of inventory optimization and utilizing appropriate tools and techniques to achieve better results.

The discussion touched on the challenges faced in the supply chain industry and the need for attracting more talented individuals to the field.

Vermorel noted that the complexity of problems in the supply chain industry is often less than that in other domains, such as microelectronics. However, he emphasized the importance of attracting more brilliant minds to the field to help solve the challenges faced. He praised Vandeput’s book for making the supply chain field more appealing to intelligent, enthusiastic individuals who can become genuinely interested in tackling such problems.

Vandeput discussed the limitations of traditional mathematical models in supply chain management, which are often based on simplifications of reality. He explained that some models may work well enough for certain scenarios, but when they fail to perform adequately, other approaches such as simulations may be needed.

Vandeput cited the example of safety stock formulas, which assume normally distributed lead times. In reality, suppliers may be on time most of the time, but when they are late, they may be significantly late. Traditional mathematical models struggle with this kind of situation, highlighting the need for alternative methods such as simulations.

The conversation focuses on the limitations of traditional mathematical models, such as safety stock formulas that assume normally distributed lead times. Both experts emphasize that real-world scenarios often deviate from these assumptions, creating challenges that necessitate alternative methods like simulations for more accurate supply chain management.

In conclusion, the interview highlighted the importance of attracting brilliant minds to the supply chain industry, the limitations of traditional mathematical models, and the potential benefits of using simulations as an alternative to overcome these limitations.

Full Transcript

Kieran Chandler: Today on lokad tv, we’re delighted to welcome back Nicolas Vandeput who’s going to discuss with us just how simple it can be and what we can learn from his new book. So, Nicolas, thanks very much for joining us again. Today, as always, we like to know a little bit about our guests and what they’ve been up to. So, what have you been up to since we last saw you on the show?

Nicolas Vandeput: Hello again, Kieran. Well, actually, as always, I’m busy working on creating some nice models for companies on inventory optimization and forecasting. And, well, I can also say I’ve been writing some books and teaching at the university, as you know that’s my big hobby now.

Kieran Chandler: Okay, lovely. And today, we’re going to be discussing one of those books, Joannes. It’s called “Inventory Optimization Models and Simulations.” So, what is it that’s different about Nicolas’s book?

Joannes Vermorel: I think, you know, Nicolas in this book is embracing, I would say, frontally one thing that I believe to be a cornerstone of modern supply chains. And when I say supply chain, I mean in the sense of supply chain optimization, not just management in the sense of just accounting for things. You need to have a programmatic expressiveness if you want to have something that has any chance to deal with real-world supply chains. And in this book, I believe that one of the things that is very interesting is that instead of going super deep on idealized supply chains, where you say, “Oh, let’s have a mathematical proof of optimality for this or that or that for an idealized supply chain that just will never exist,” and where if you add a bit of real-world ingredient into the supply chain, all the mathematical frameworks just fall apart. Nicolas is doing something that I believe to be much more in the agile mindset, which is much more appropriate: just show how you can provide very simple and straightforward numerical recipes with Python.

The beauty of the recipes is that it’s very hands-on and literally shows that most of the classic supply chain recipes can be implemented in like five lines of Python or so most of the time. And I believe that’s very interesting because it conveys the idea that if you want to do something very simple, it can be done in very simple ways. It doesn’t take a half a dozen software engineers to come up with those things. And I believe that as a side effect, it profoundly demystifies what large software vendors are actually pushing to the market. Because when you show that basically, you can do what they are saying they are doing, but just in a few lines of code, the question is: Is there any value to what is being proposed by those vendors? And I believe, to a large extent, no. But even beyond that, it shows that if you can have very small building blocks, you can suddenly combine them to inject your bits of real-world challenges into those recipes so that you have something that has at least a decent chance to be suitable.

Kieran Chandler: Solution for your supply chain. Okay, sounds certainly very interesting. Nicolas, this idea of just a few lines of code and you’re already getting results and it’s much more satisfying than spending hours and hours looking at coding. So, who is this book aimed at?

Nicolas Vandeput: One of my big obsessions is simplification. So when I write such a book, I’m trying to think, “Okay, if I manage the supply chain and I want to optimize inventory, how can I simplify everything to give a global picture and yet at the same time allow practitioners to do it themselves?” So I’m really writing this book for anyone in a supply chain right now that is thinking, “Oh my god, we have too much stock,” or “We have so much stock and at the same time, we don’t have the right service level,” or simply, “Well, I’m in charge of the stock, but I have no idea how much I need.” I really wrote this book to help those people, to basically do it themselves. There are so many “do it yourself” sections in this book that show, in the simplest way, how you can do it. And yet at the same time, I’m really trying to tell them, “Well, we have to make some assumptions, and as we follow some assumptions, we are not aiming for perfection that doesn’t exist. So we’re going to do some simulations and we’re going to see the limits of the models we use.” So in short, this book is really for anyone in supply chain today who’s thinking, “Okay, I need to get the inventory correct.”

Kieran Chandler: Great, and it’s a bit of a change of mindset, isn’t it, Johannes? Because historically, programming was something that was left to IT departments. Now it seems that there are more and more practitioners that are adding those strings to their bow. So why would you say there is this mindset shift?

Joannes Vermorel: I believe those things have been around for a long time. But in terms of supply chain maturity, for decades, companies were just struggling to have the digital counterparts of their supply chain, to have an ERP or WMS setup where you can just have the stock levels managed in a straightforward manner – no intelligence whatsoever, but just an accurate digital counterpart of your supply chain. It took a long time for us to get there. If we look at the fact that the first ERPs were deployed – they didn’t go by this name – by the end of the 70s, we have four decades of ERP history under our belt nowadays. So it took a long time to have sufficient digital mapping so that it would become a reasonable option to do all sorts of optimization. And it has been a slow move from Excel to Python, by the way. Excel gives you already a lot of programming capabilities, which are also, to some extent, illustrated in the book. So for me, it’s a continuum, not a complete disruption where you go from something to another. You need this programmatic expressiveness to cope with all the things that the real world is throwing at you.

Kieran Chandler: So Joannes, you were just talking about the need for software to be expressive to handle the unpredictable nature of supply chains. Can you elaborate on that a bit more?

Joannes Vermorel: Yes, I think that there will always be things that are completely random and unpredictable. It can be a Brexit, a Trump tariff, a pandemic, or suddenly the fact that the company has to deal with the distribution of a vaccine, which is going to turn your supply chain upside down. So there’s a lot of stuff that happens, and if you just have like a rigid piece of software, it’s just not going to be able to cope with all the things that happen in the real world. So you need to have something more expressive, and thus you need to have this programmatic expressiveness. Excel gives you that, but it comes with certain limitations that are, I would say, profoundly related to the way you organize your logic in a spreadsheet. Python gives you the next level; it’s a bit more abstract, but you get, I would say, the next level of expressiveness where you can have things like user-defined functions. You can have that in Excel through Visual Basic, but for all intents and purposes, Python is just kind of a superior flavor of VBA.

Kieran Chandler: Nicolas, let’s talk about maybe the continuation from Excel to Python, and there are plenty of other programming languages out there, things like SQL, C-sharp, and so on. Why is it that Python was your language of choice for this book? What does it provide that perhaps some of those other languages don’t?

Nicolas Vandeput: Well, overall, we see that Python has a few advantages. The first one is that it’s actually kind of well-known. I don’t know if you could say it’s the number one language in the world, but it’s at least close to it. It means that today, if you ask yourself a question about Python, you just Google it, you’re going to find an answer, and this is really convenient. If you find a programming language that’s much faster than Python, so you’ll say, “Okay, I’m going to use this one, it’s faster,” but you get some questions and you type those questions online, but you see no answer, you’re going to get stuck. With Python, it’s extremely difficult to get stuck because I can really tell you that someone somewhere already had the same question and it had been answered already. On the other hand, and I think this is extremely important, I’m really into simplification. Python is really simple. I always remember from my days as a quantitative analyst those kind of colleagues that use VBA. We all have many people have in mind this kind of huge VBA macro file in Excel that everyone is kind of afraid of, and you don’t dare to touch anything in the Excel file because if it breaks down, it’s impossible to fix. Python is nothing like that. Well, Python is much simpler, and you can basically read it and understand what it does just as you keep on reading the file. I had one or two readers writing me an email after reading the book, telling me, “Nicolas, I don’t know anything about Python, but I read your code and basically, I understand what you do in your code because everything is super clear.” So thanks to these two things, I think Python is really the language to use if you want to learn something new. Now, concerning the book, why am I putting so much emphasis on Python? Well, I see two reasons why I’m doing this. First, I believe that if you want to learn something, you

Kieran Chandler: So I’m curious, in your book, why did you choose to focus on Python for supply chain optimization?

Nicolas Vandeput: I think it’s much better if you can experience it by yourself and try it out by yourself. I’m pushing many do-it-yourself sections in the book, either in Excel or Python, because I want the reader to acquire new know-how and knowledge. I also want to demystify supply chain optimization by telling them it’s not so complicated. You just type a few lines of code and it’s going to work on your own computer. The main reason I’m pushing Python is that it’s easy to scale the solution to a supply chain. With Python, you can easily run a supply chain using simple assumptions and models.

Kieran Chandler: Would you agree with that, Joannes? We’ve obviously spoken a bit in the past about some of the limitations of Python.

Joannes Vermorel: In terms of popularity, there’s no question that Python is in the top 10 of the most widely used programming languages. Contenders would be JavaScript, Java, and a few others, but they are not good options for supply chains. It takes significant software engineering skills to do anything with those languages. Some of the good qualities of Java, for example, extensive support for object-oriented programming, can be a defect when it comes to easily onboarding new users. These features are not readily useful for supply chain optimization or supply chain modeling purposes. If you just throw that into the mix, you have a more complex programming language without any obvious upside, at least not for the first few months of any project.

Python was started in the ’90s, and it took almost three decades to achieve popularity. There was a massive migration from Python 2 to Python 3, which I believe was a takeover by academia. Python found its sweet spot in academia, and the big transition between Python 2 and Python 3 was to remove all the bad parts of Python. What emerged from that during the last decade was a language that was much more in line with what you need in academia and for supply chain optimization.

Kieran Chandler: So, let’s talk about programming languages. Joannes, what are your thoughts on Python?

Joannes Vermorel: Python is something where you have something that is very concise, where you don’t have too much mass with too many things floating around. The exact opposite of that would be probably something like C++. I’m not sure if there is anybody on earth that says, “I know all there is to know about C++,” because the specification of the language is so absolutely gigantic that I don’t think it’s even humanly possible to be familiar with all the parts of C++. The language is kind of insane when you think about it. So, all of that put together, you have Python that is really a sweet spot, to let people start readily. I believe that’s a good starting point to avoid a lot of the pitfalls of accidental complexity. Now, where I believe that it’s not actually the end game, though, and that’s by the way what we have been developing at Lokad. But obviously, this is not the topic of the book of Nicolas, so I’m not going to digress too much. But I believe that if you want to cope with uncertainty, there are plenty of things that can be done relatively simply in Python. But if you’re willing to go as far as modifying the language itself, it could be done in ways that are even simpler. But it goes beyond the book’s scope. For the purpose of the discussion, I think right now that if you have to choose between what you can get from Python versus most of the expensive options on the markets, it’s literally a bargain, and there are very little reasons not to at least give a serious try. Even if you fail, it will be a much cheaper failure compared to failing with an SAP of this world.

Kieran Chandler: Nicolas, as well as looking at some of the models in your book, your book also looks at some of the confusions when it comes to inventory optimization. What are the common confusions that we should be aware of in the industry?

Nicolas Vandeput: Yeah, with my experience as a consultant, discussing with many practitioners on how they deal with inventory optimization, and when you look at, well, you know, this safety stock formula that you see everywhere, even on Wikipedia, you see people going on Wikipedia, typing safety stock, looking at the formula, and then typing it into Excel. What we see there is that–and this is correct when you want to assess how much safety stock you need–you’re going to take a look at the lead time, so basically how long does it take to do a replenishment. While people’s first confusion is that I see that many practitioners confuse the whole concept of lead time and transportation time. It might just take two days or one day for a truck to go from one warehouse to another, but it might take three weeks of planning because you need to find a truck, you need to find a driver, and you need to have the picking time and all these kinds of things that it pipes up. And in the end, you take two weeks, but there is still one thing that people always tend to forget. Inspiration is that you shouldn’t just take a look at the lead time, but you should also take a look at the review period.

Nicolas Vandeput: Let me give you an example. If you make an order every week or every month, that’s an addition of what I call in the group the risk period. So basically, the period against which you have to cover your needs is not just the lead time but also the review period.

Kieran Chandler: Okay, Nicolas ,in your book “Inventory Optimization Models and Simulations,” you talked about the importance of taking into account the review period. Can you elaborate on that?

Nicolas Vandeput: Yeah, so the review period is basically the amount of time that you need to be protected, increased by the amount of the review period. So if you only do replenishment every week and you have a lead time of three weeks, the risk period is actually four weeks. So you need to be protected over what could possibly happen over four weeks. What I see is that most of the people, most of the software, tends to totally use and bypass this review period and will just focus on these three weeks of lead time. So I’m trying, reading the book, to emphasize the fact that, well, you absolutely need to take into account this review period when something safety stuck. But also, I’m pushing people to realize that actually, if they can reduce this review period, if they can reduce the frozen period for production and so on, they basically are allowed to reduce safety stock, which I think it’s kind of a good win for supply chain.

Kieran Chandler: Okay. And would you agree with that, Joannes? Would you say that there’s a lot of confusion out there, and that’s why people are so happy to pay for sort of inventory software?

Joannes Vermorel: Yes, I mean, again, I believe, you know, my belief is that supply chains desperately need, I would say, better classes of engineers. And, in one of my first lectures, I had this introductory joke that, you know, if you have a lot of energy, you go to sales. If you’re reliable, you go to production. And if you’re lacking, you know, all sorts of qualities, you end up in supply chain. But the thing is, you know, if I compare certain domains in supply chains, the sort of confusions that you encounter are not, I would say, the product of exceedingly brilliant minds. And I’m sorry to be, maybe a bit tough to the audience, but if you look at, let’s say, for example, what is being done with lithography for microelectronics and the sort of problems that those people are solving, you know, they are mind-blowingly complicated. It’s literally, you have everything. You have quantum physics with complicated math. And frankly, it has way more complicated than what is being the sort of problems that are being tackled in supply chain. You have physical problems where you have all sorts of difficulty. The technology is literally you have so many super complicated pieces of tech that you have to bring together. I mean, for example, what sort of companies like ASML are doing in the Netherlands, I mean, it’s almost magic because it’s it’s it’s it’s just, I would say, it’s a sort of achievement of mankind when you have like the most brilliant minds put together. But the challenge is that if you want to have smarter people coming to supply chain so that way we can attract very brilliant minds that are not going to be confused by problems as dumb as just the definition of the applicable period for your inventory replenishment, you know, a system that you’re about to model, we need to have to form the problem in a way that those people can exercise your intelligence. You know, because obviously, if if you have if your entire discipline is about you know, turning dumb buttons on dumb pieces of software, then don’t get too surprised if in the end, the only employees you get are not the most brilliant ones.

Kieran Chandler: Um, so I believe that, and that’s also something that I like about this book is that, you know, I think the book of Nicolas is that it gives something where, if you’re smart, if you’re young, if you have some degree of enthusiasm, then you can get really genuinely interested by the sort of problems that you will tackle. And not only you will, by reading the book, you know, you can actually get better at solving those problems, which is a very good buttress for supply chain as well. It’s the sort of things that can make the world feel more attractive to brilliant minds who want to exercise their minds on interesting problems, in the first place. And so what I suspect is that, in terms of confusions, you know, the sort of things that confuse people now, and what the sort of things that I hope will confuse people in two decades from now, will be radically different, especially if we bring, you know, I would say a lot more talent to this industry. Okay, if we start sort of wrapping things up a little bit. Nicolas, in your book, Johanna sort of mentioned that some of the models that are used have their very much kind of applications, maybe from an academic perspective, but maybe in the real world, they have their kind of limitations. Would you say that some of these limitations can be overcome and how can they be overcome? And how can they end up being used in kind of the real world?

Nicolas Vandeput: Well, you always have to understand that somehow a model is a simplification of reality, right? So, from the base when you start a model, you have to understand that you have to leave some things aside. So, the real question is, okay, if I take a model that works, let’s say, 1995, 98% of the time, is it good enough for me? Someone would argue yes it is or no it isn’t. Now, if you do a model and you see that it only works, let’s say, 70% or 60%, then you really realize that okay, the usual mathematical model is not good enough. I need to move on to something else. And this is what I show in the last part of the book, saying that at some point, a mathematical model will not be good enough. I mean, will not be accurate enough, will not be tractable meaning.

Joannes Vermorel: And that’s something that, actually, you know, we see with the clients of Lokad as well, is that, basically, the models that we have, the mathematical models that we have, are good enough to identify what we call the big wins. That is to say, the things that, if you fix them, you know, you have a massive, massive impact on the supply chain, you know, profitability or whatever it is that you are trying to optimize. And this is where, I would say, you know, we bring a lot of value to our clients. However, there are also a lot of details where, you know, the model is good, but it’s not good enough, and this is where, I would say, you know, having some kind of industrialization, where you can really test and tweak and adjust and, you know, basically do, you know, testing in the real world. I mean, this is really the core of supply chain, right? It’s a mix of analytics and operations, where, you know, you need to have the analytics to make sure that you don’t do, you know, stupid stuff, but then, you know, the operations, the testing, the

Kieran Chandler: So, I wanted to ask you guys about inventory optimization. It seems that it’s an area where people are often hesitant to use models because they’re afraid that they’re going to get too complicated. So, you might go to a simulation, and I’m quite sure that Jonas has so much to say about how to do simulation in the supply chain.

Joannes Vermorel: Well, from this first part of the book, I’m also showing, “Okay, those are the usual models for the supply chain. Let’s do some simulation to see if they work, yes or no.” And I really see there that, for example, one of the things that we usually do in the supply chain is taking this, again, this usual safety stock formula, the one that’s on Wikipedia. And this formula can deal with random need time, meaning that from time to time, your supplier is late, so you should have some more safety stock, right? When actually you look at the formula to assess how much safety stock you need based on random need time, you see that there is a whole assumption that says, “Well, lead time, I normally distribute it.” So basically, it’s a well-behaved curve that your supplier, from time to time, it’s too late, but never so much late. Actually, in practice, I believe that most of the suppliers, most of the time, are on time. But then when they’re late, they’re quite late, right? So I would say 80% of the cases they’re on time, and then 20% of cases, whereas it can be one or two weeks late. What actually is there is no mathematical formula that is able to deal with this, right? So if you have such a kind of supplier, you will not go anywhere with this safety stock formula. You really need to go back to simulation, and this is where you have the limit of using some easy mathematical model and the limit of, well, we should go one step beyond and maybe start to use simulation to solve this.

Nicolas Vandeput: Yes, and I completely agree with what Joannes is saying. I think there’s a limit to what you can do with mathematical models in terms of supply chain optimization. And simulation is a very powerful tool to help you understand the impact of the different parameters on your supply chain, and to optimize it. And in fact, that’s what we’ve been doing for many years, and what Joannes is doing as well.

Kieran Chandler: Okay, brilliant. We’re gonna have to leave it there, but thank you both for your time. If you’re interested in reading Nicolas’s book on inventory optimization models and simulations, we’ll put a link in the description below. Otherwise, we’ll see you again in the next episode, and thanks for watching. Bye.

Joannes Vermorel: Thank you.

Nicolas Vandeput: Thank you.