00:00:04 Antony’s aerospace career introduction.
00:01:57 Joannes’ aerospace forecasting method.
00:04:56 Evolution of probabilistic forecasts.
00:05:28 Safety stock method limitations.
00:06:52 Changing market and inventory solutions.
00:08:00 Adopting local approaches and education.
00:10:20 Problem prioritization and business insights.
00:12:27 Aerospace industry’s unique supply chain.
00:14:36 Local approach benefits in inventory.
00:15:57 Client benefits from new strategies.
00:18:29 Time allocation: customers vs. internal tasks.
00:19:56 Pattern recognition in supply chain.
00:21:17 Aerospace supply chain evolution.
00:24:00 Need for quantitative modernization.
00:24:48 Potential for automation in warehouses.
00:26:00 Better software integration necessity.
00:27:06 Future of predictive maintenance.
00:29:19 Skills for future aerospace professionals.


In the Lokad TV episode, host Kieran Chandler speaks with Antony Nardozza, Head of Inventory Solutions at OEM Services, and Joannes Vermorel, Lokad’s founder, about supply chain management in the aerospace industry. Nardozza, with his vast experience, and Vermorel, with his forecasting innovations, have collaborated to optimize the complex aerospace maintenance process. Emphasizing aligning quantitative models with business objectives, they’ve evolved from safety stock to a more dynamic inventory management approach. The pair acknowledges future supply chain management will incorporate predictive maintenance and data-intensive operations. Consequently, practitioners will need to expand their skill sets in data science, engineering, and supply chain management to meet these upcoming changes.

Extended Summary

In this episode of Lokad TV, host Kieran Chandler interviews Antony Nardozza, Head of Inventory Solutions at OEM Services, and Joannes Vermorel, founder of Lokad. They discuss the supply chain challenges faced in the aerospace industry and Lokad’s unique approach to forecasting.

Antony Nardozza has nearly 20 years of experience in the aerospace industry, having started his career with aircraft manufacturer Airbus. He has worked in various roles, including flight testing and customer services, and has experience with multiple aircraft types such as the A380 and Embraer E-Jets. Nardozza currently leads inventory solutions at OEM Services, overseeing all aspects of supply chain and inventory management for the company.

Joannes Vermorel has developed a pioneering method for forecasting in the aerospace industry, with significant input from Nardozza. One of the key insights of this method is the importance of aligning quantitative models with business objectives. This means that any numerical recipe or optimization method should account for the complexity of the aerospace maintenance process and capture the essence of the business, rather than merely providing numerically correct but business-wise inaccurate results.

Initially, the pair started with safety stock applied to aerospace, but soon realized that it was not an effective solution for many situations. They sought to develop a process and numerical recipe that would support decision-making in a way that made sense from a business perspective. A primary concern was ensuring that airplanes could be kept flying as much as possible, and translating this insight into a workable process.

The aerospace maintenance process is inherently complex, with numerous unknowns such as delays and required quantities of parts. Even with advancements in sensor technology and the ability to gather real-time data, it is impossible to account for every variable. As such, the forecasting method developed by Vermorel and Nardozza aims to provide a numerical recipe that assists with specific classes of decisions, such as determining how many parts need to be kept in stock at a given location or deciding whether to sell a particular part to a third party and at what price.

Before collaborating with Lokad, OEM Services was already managing their purchasing decisions using safety stock practices. However, as the market evolved and service providers began consolidating services from different airlines, cost efficiency became a significant concern. OEM Services realized that they could improve service levels and reduce the amount of unused inventory on their shelves. This led them to partner with Lokad to optimize their inventory investments for better service levels.

One of the challenges faced by OEM Services was adapting to Lokad’s approach, which gave their staff more autonomy in making decisions. Initially, employees found it difficult to take responsibility for their decisions, as they were used to relying on predetermined stock levels. The new approach provided a broader view of service levels, costs, and decision-making implications, which required a shift in mindset.

Another challenge was prioritizing purchasing decisions and inventory sizing. Different stakeholders within the organization had to adapt to the new system, which could lead to feelings of disappointment or satisfaction depending on the allocation of resources. Nardozza emphasized the importance of accompanying the solution with the necessary organizational changes.

To determine where to start and which problems to prioritize, Lokad relied on the business insights of their clients. The fact that a business was already operational indicated that they had some level of knowledge and insight into the industry. Lokad focused on understanding the client’s business rather than the data at first, as the data could be influenced by the complexity of IT implementations.

Vermorel shared an example of the challenges faced in understanding the client’s business. In the aerospace industry, retrofits are a unique supply chain pattern, initiated by the Original Equipment Manufacturers (OEMs) to replace components for safety or improvement reasons. Vermorel admitted that he was initially unaware of this concept, but understanding such industry-specific patterns is crucial for supply chain optimization.

One of the first points of discussion was the unpredictable nature of aircraft part movements, especially with newer models such as the A380, and how the traditional methods of inventory management often failed to account for these fluctuations. This led to the introduction of Lokad’s approach, which allowed for more adaptable inventory management, tailoring to changes in demand and supply chain behavior over time.

Nardozza emphasized the benefit of “letting their stock breathe” with Lokad, which meant inventory could be adjusted according to the changing needs of an aging aircraft. This ability to follow the shifts in demand and supply chain behavior prevented service level decreases and allowed for more dynamic, efficient inventory management.

Vermorel elaborated on the benefits Lokad brought to the table, including profitability and improved operational efficiency. He highlighted the shift from viewing supply chain operations as an operational expense, towards capitalizing on it. Vermorel pointed out how Lokad’s automation reduced the need for mundane human decision-making, freeing up more time for strategic analysis and value-added tasks.

Vermorel acknowledges the potential of software integration to significantly reduce mundane clerical tasks and improve overall efficiency. He envisions a future where supply chain practitioners are more intertwined with software operations.

Nardozza speaks on the future of aerospace technology, highlighting the growing importance of predictive maintenance. This approach, in recent years, has been gaining significant traction in the industry due to its ability to anticipate part failures, improve customer operations, and enhance overall component maintenance. The anticipation of failures allows for more efficient inventory management and can drastically alter the overall supply chain strategy.

Furthermore, Nardozza mentions the necessity for new skills and expertise in this evolving landscape. As the industry moves towards more data-intensive operations, practitioners will need to develop skills in data science, engineering, and supply chain management, fostering collaboration and communication between these areas.

Full Transcript

Kieran Chandler: Hey, this week on Lokad TV, I’m delighted to say we’re joined by Antony Nardozza. Antony is the head of inventory solutions at OEM Services and has over 10 years of experience in the aerospace industry. Today we’re going to be talking about some of the supply chain challenges that he’s faced throughout his career and also his experiences working with Lokad’s unique take on forecasting. Antony, thanks very much for joining us today. Perhaps a nice place to start would be if you could tell us a bit about your background.

Antony Nardozza: Sure, thanks for giving me the opportunity to discuss this topic together. I spent almost 20 years in the aerospace industry, starting with aircraft manufacturer Airbus. I worked on flight tests and certification, so already being on the aircraft and knowing the life and daily operation of line maintenance was a very good field experience to get to know the service environment. Then, I turned to serving customers in the customer services department within Airbus, where we were focused on the entry into service of the super jumbo A380. After having done a few entries into service, I joined the customer side of this industry with a company named Spairliners, which is a joint venture of Lufthansa and Air France. We supported the A330 and later, the Embraer E-Jets. Now, I’ve joined the equipment part, the Original Equipment Manufacturer, with OEM Services where I’m now leading the inventory solutions, covering everything related to supply chain and inventory topics for this company.

Kieran Chandler: Great, and as always, we’re joined by Joannes Vermorel. Joannes, today we’re going to be talking a little bit about the pioneering method you developed for the aerospace industry to do with forecasting. What exactly did this method consist of?

Joannes Vermorel: First, Antony has been a big part in helping us shape this method and rethink it. I think one of the many insights was to have a quantitative modelization that would be aligned with the business objectives you’re trying to solve. This means that whatever you want to optimize, you’re going to end up with a numerical recipe, and you have to approximate tons of things because, for example, the maintenance of aircraft is impossibly complex. There are tons of unknowns, and you cannot factor in all the variables. Even in the distant future with all the sensors we have on board, we’ll be able to have a much better view of the real-time and near-future state of any aircraft and its components. However, when you operate with an aerospace supply chain, you have tons of unknowns, both in delays and quantities.

You need to end up with a numerical recipe that helps you take certain classes of decisions, like how many parts you need to keep in stock at a specific location, or whether you can actually dispose of a part or resell it to a third party at a specific price. There are many classes of decisions, and whatever your numerical recipe is to assist your decision, you have to make sure it captures the essence of the business. The numerical recipe shouldn’t give you results that are numerically correct but business-wise completely wrong.

When we started working with Antony, we started with safety stock applied to aerospace, and the starting point was realizing it doesn’t work. Maybe for consumables, it might work a bit, but for many situations, it was desperately not working. I believe that, thanks to Antony, we uncovered some of the key insights that were really driving the business. In a very obvious way, you want to keep the airplane flying as much as you can. So how do we translate this kind of insight into a process and a numerical recipe.

Kieran Chandler: So, we are moving forward with the descendants of those early insights, which were probabilistic forecasts and a few other things. Before we get into the probabilistic forecasting side of things, we mentioned safety stock. What was going on before you were approached by Lokad?

Antony Nardozza: Safety stock has existed for a long time, and it was already in place before we collaborated with Lokad. It was working, but the market has changed over time. In the past, every company would project their own inventory, creating their own stock, and the side effects of this approach were marginal in regard to the total cost of managing an airline. There wasn’t enough incentive to optimize this approach and go beyond the traditional safety stock, which was sufficient in terms of support efficiency but a nightmare in terms of cost efficiency.

The integration of service providers, who collect services from different airlines or different delays from various airlines, and the centralization of the responsibility of service delivery have forced businesses to be more efficient. This was the first trigger for us to move in this direction – we needed to understand the problem and try to solve it. We observed that the service level we achieved was not what we expected, and we had plenty of material on the shelf that was not moving. This was the starting point of our reflection and the question we raised with Lokad. We wanted to find out how we could turn our investments in inventory into the best service level possible.

Kieran Chandler: Let’s talk about the internal changes you had to make within your organization to adapt to Lokad’s approach. Was it easy to convince people?

Antony Nardozza: Convincing people is never easy at first. The system that Lokad offers is simple and logical, but it requires time to educate people and bring them to the right level of knowledge about these types of solutions. Another issue is that it gives them more autonomy in decision-making, which can sometimes make them uncomfortable. In the past, when they were given a single piece of advice, such as a stock level, there was no question about whether to proceed with it. They could simply blame the calculation or the person who gave them the advice if things went wrong.

However, when they have a wider view of the service level, the costs involved, and the implications of each decision, they can sometimes feel disturbed. This new approach means they are free to make the right decision or fail, and this can create some discomfort. Some people might want to step back to the previous version, where they didn’t have to worry about making their own decisions. But our new approach provides a full view of the possible future, depending on the decisions we make together.

The other major change is the way we think about purchasing components and sizing inventory.

Kieran Chandler: How do you give priority because depending on the size of the aircraft, type, full fleet, customer type, you can optimize differently? Some people who are responsible for different parts might feel disappointed or happier because they get more money. I saw all these kind of changes which are important to bring with the solution. Joannes, is this a good point for you to jump in? There’s so much variability, so many different problems, and so many things to take into account with these decisions. How did you know where to start? How do you know which problems to even prioritize?

Joannes Vermorel: I mean, that’s where you rely on the business insights of the client. The good news is that it’s very rare to meet a business that is already working where they don’t have those insights. Decisions are already being taken every single day. Somehow, even if it’s someone who is just saying, “I believe in this,” it means that someone, somewhere has knowledge that is of interest. The proof is in the fact that the company exists and it works. We are lucky, and the insight is to focus more on understanding the business rather than understanding the data at first. The data can reflect more of the complexity of the IT implementation, which is a bit accidental. For example, the company may use a famous German ERP that supports a large variety of verticals, but your client is not in fresh food or light manufacturing; they are servicing aircraft parts. They are using this gigantically capable software, but for a tiny subset of the whole thing. There is a danger when you look just at the raw data of getting lost in things that are completely irrelevant. Sometimes, you might even have a concept that is missing, something you cannot even comprehend because it’s something you haven’t even imagined.

For example, I discovered, I was very ignorant when Antony told me about retrofits. I was thinking, “What the hell are those retrofits?” If you don’t know, the idea is that in aerospace, safety comes first. If an OEM has the slightest doubt about the potential security or improvement of any component, they can recall every single unit that has ever been sold and send a new one to replace the old one because the new one is superior for a variety of reasons. Then you have a very specific supply chain pattern where there’s a big wave of parts that move through the supply chain, but it’s not initiated by the clients but by the OEMs themselves. It’s a very unusual pattern where it’s like a forced push from the OEM.

I first encountered this when Antony was working on the data for the A380 during its early years. When an aircraft is new, you typically have more retrofits at the beginning of the lifecycle. This was the case for the A380 because it was still a very new plane at the time. That’s the sort of thing where you can fail dramatically if there is an important concept that you can’t even imagine. Just imagine what sort of optimization you can do on top of that if you’re completely missing the point.

Kieran Chandler: Let’s talk about the benefits after using the Lokad approach. What has changed from your point of view? What are the key benefits that you’ve seen?

Antony Nardozza: First, the ability to let our stock breathe.

Kieran Chandler: Which was something a bit difficult to do before because when you put stock on shelves, you just expect that one day you’re going to serve your customer. And then here, we had the opportunity to take some inventory out, some other in, and for good reason as well. The consumption is changing with time. An aircraft which is new is not having the same needs as an aircraft which is getting a bit older or very old. And so the need for parts, the shape of the inventory, is to be different. And if we don’t follow the change of the demand or, as we’re changing the supply chain behavior, some legs are going to be longer than others, and some are going to be shorter at the time. And all of this is changing, so if your inventory stays just fixed with all the same parts and shelf life, your service level will just mathematically decrease. And then you don’t have any control because you expect that things are stable. You want inventory that is tailored, we know your role is changing around you, so we need to breathe and follow this, yeah. We’re always changing. And how about you, Joannes? What sort of the key benefits that your clients have seen or told you about since working with Lokad?

Joannes Vermorel: I mean, we had interesting discussions with CFOs where it was profitable, significantly so, that’s good. But what I see, distancing myself a bit from the immediate profitability, is to gradually shift the interest of the teams more toward things that add more value in terms of strategic analysis. You see, as opposed to treating your engineers as operational expense, where to keep your supply chain operating, you consume X days of manpower every single day, that’s your ongoing operation. You start capitalizing on the work that is being done. It’s a transition where, instead of thinking that you have mundane decisions where you need a human to spend a lot of time to generate every single decision, which because you have a complex supply chain where things are moving all the time, it requires a lot of manpower all the time. You transition toward something where you have gradually automated this process, so that people don’t have to spend that much time on that, and so they can do something that has much more added value, which is to change the process itself and even challenge the process for which the recipe is used. So sometimes, you change the process, and then the recipe should be changed, and overall, you have something better.

An example in supply chain would be, okay, yes, you can be super tight in terms of inventory optimization, but actually, you are optimizing for certain lead times. And if you manage to change the organization with your suppliers, with your repair shops, etc., to reduce lead times, then you reduce the amount of inventory you need. And there are similar things. And again, the question is that I believe that most of the time, we didn’t invent those opportunities for improvement; teams were already aware of them. But the problem was that they were not having the time. And where I was personally very satisfied with seeing Lokad deployed is, over the years, to see those clients who visibly have more time to really think about those deeper strategic issues and to really ramp up into some sort of work that seems to be more fulfilling, even as a human.

Tackling the same thing every single day for your entire career is typically not very fulfilling as a human being. You want to have something where there is a sense of progress, where there is some kind of achievement you can capitalize on. So, besides just pure immediate financial return on investment.

Kieran Chandler: Welcome to the show. Today, we have Joannes Vermorel, the founder of Lokad, a software company that specializes in supply chain optimization, and Antony Nardozza, Head of Inventory Solutions at OEM Services, a great expert in the field of aerospace supply chain. We will discuss supply chain optimization and the evolving role of supply chain management in the aerospace industry.

Joannes Vermorel: A key aspect of supply chain optimization is to improve calculations and provide valuable feedback to customers. It’s important to focus on giving good advice and explaining how you do it. This is essential in order to differentiate your business from competitors.

Another important aspect is that optimization tools don’t have any emotions. For example, if you remember a part that has caused troubles repeatedly, you might still purchase it even if things have been going better for the past six to nine months. You need time to come back to reality, and optimization tools can help us adapt our vision more efficiently to what is actually happening.

Antony Nardozza: Yes, it’s interesting to think about human attachments to machine parts. Humans are not very good at perceiving randomness and tend to see patterns in everything. Emotional attachments can lead to us seeing patterns where there might only be chance.

Kieran Chandler: Antony, as Head of Operations at OEM Services, how have you seen the role of head of supply chain evolve, and how do you see it evolving in the future?

Antony Nardozza: The role of head of supply chain has definitely changed over the 20 years I’ve spent in the industry. In the future, I believe tools will be more supportive of our decisions. The aerospace industry, in particular, is moving from a project logic to a service logic. We are seeing a shift from purchasing assets to renting services. Supply chain management will likely become more focused on providing a platform of services. The challenge for supply chain management in current companies is to adapt to this change from product to service orientation.

Joannes Vermorel: I believe that the role will increasingly require strong engineering competency in managing and optimizing supply chains, particularly in aerospace. Quantitative skills will become more important. Of course, leadership and team management will still be essential. However, being able to think critically about optimization, even if you don’t do it yourself, is crucial. You can have solution providers to support you, but you need to be able to understand and engage with the optimization process.

Kieran Chandler: Would quantitatively modernize your objectives, you know, what are you trying to achieve? How many AOG incidents have you avoided this year, and how much did this cost extra? I see that as something that will be increasingly a problem of not having this appetite for these sort of things. Plus, also another angle that I see is that, as an effect of this industry transitioning toward services, it means that many companies are in a way building some kind of platform software platform and becoming service providers, but you know, with a lot of automation involved, which involves software. So, those companies also have an angle for the quantitative aspect, but there is also an angle for the software aspect.

Joannes Vermorel: Because much, I would say, you have the mechanical automation that is not actually available due to the fact that many things are either too large, too small, too diverse. As well, you can’t actually automate your warehouse like an e-commerce would do with robots. You can partially, I mean, you can put a very large number of items. If I remember correctly, the number is between sixty and seventy percent of the parts fit in a box like 50 to 40 meters. So, there is a way to do it, but not for everything.

Antony Nardozza: That’s pretty correct. I mean, an engine won’t be taken at home. We can obviously see this, but there is a part that can be done. Again, it’s a question of volume and concentration of activities to get a certain scale effect that allows you to do that. Where I see a potential for significant automation is for very mundane clerical tasks in the aerospace industry, especially because you have all the traceability. So, there is a lot of operation sheets and having all the component maintenance manuals, how this thing is moving along the path. We still have a lot of situations where I’ve seen things being printed, scanned, reprinted, rescanned, etc. And I believe that it will evolve with better software integration so that there is less manual glue in the middle, which means that, probably, again, for the supply chain practitioners, they will interact with more software and better designed software. But I see that the appetite for numbers and software will probably grow over the coming years, even if there are many other qualities that will remain present, like the capacity to manage a lot of people that are on the ground, in many different places.

Kieran Chandler: Okay, and a last word maybe from you, Joannes. What sort of technology is there out there that we can see in the future of the aerospace industry? In recent years, predictive maintenance has really burst onto the scene in the aeronautics industry. Everyone is starting to adopt this approach, as it’s seen as the future of supply chain management. The ability to predict when a part will fail, with a fairly accurate estimation, will improve operations for the customer. They will be able to prepare their shop to receive parts, react better to part failures on aircraft, and make improvements to their components. How do you think this will change the nature of supply chain management in aeronautics?

Joannes Vermorel: Predictive maintenance will completely change the nature of the aeronautics supply chain. The way inventory is located and managed will be different, as we will have a view of something we had no idea about before. The new generation of aircraft will deliver tons of data that will give us the ability to drive this better. However, it doesn’t remove the need to evaluate and estimate your stock and inventory value, as we discussed earlier. These are different time periods we are looking at. The investment in parts is a mid and long-term decision, while predictive maintenance ranges from a few hours to maybe one or two hundred cycles. These are completely different scales we’re talking about.

Antony Nardozza: Predictive maintenance will feed the other aspects of supply chain management, allowing us to locate parts closer to the point of need. Supply chain optimization and inventory management will provide the framework in which we’re going to operate. This will be the future of supply chain management and will require a change of skills in the people doing it. They will need to be experts in their domain, able to talk to engineers about components, have supply chain expertise, and also have skills related to data science, which involves handling these tons of data. All these people need to be able to work and communicate together, so we don’t have to look for a single person who can do all of these tasks, but we need people working in two or three groups.

Kieran Chandler: Great. I’m afraid we’re going to have to wrap it up there. Antony, thanks very much for coming in to visit us, and Joannes, thank you for your time.

Joannes Vermorel: It was a pleasure. Thank you for having us.

Kieran Chandler: Thank you very much for watching, and we’ll see you again next time. Goodbye!