00:00:07 Introduction and background of Pierre Khoury and shippeo.
00:01:25 Real-time visibility in supply chains and its importance.
00:02:59 Shippeo’s technology and how it works in practice.
00:04:00 The impact of real-time visibility on decision making in supply chains.
00:07:03 Technical challenges faced by shippeo and overcoming them.
00:10:01 Factoring various issues for estimated time of arrival algorithm.
00:11:07 The benefits of granular, real-time data for supply chain efficiency.
00:12:55 The need for granularity in supply chain data and identifying potential problems.
00:15:14 Breaking silos in the supply chain through shared information.
00:16:30 Shipping company’s approach and benefits for stakeholders.
00:17:19 Challenges of predicting and optimizing without real-time visibility.
00:18:35 Future of real-time visibility and tactical corrective actions.
00:19:56 Real-time visibility as a top supply chain topic and value creation through data sharing.


In this interview, Kieran Chandler speaks with Joannes Vermorel, founder of Lokad, and Pierre Khoury, CEO of Shippeo, about the importance of real-time visibility in supply chains. Vermorel stresses the need for optimization based on accurate measurements, while Khoury highlights Shippeo’s technology that aggregates data and GPS points to determine estimated times of arrival. Real-time visibility allows better decision-making and can help mitigate potential issues. Despite technical challenges, both Vermorel and Khoury agree on the value of real-time data for tracking assets and avoiding costly disruptions. Shippeo’s data-sharing aims to break down silos and improve efficiency throughout the supply chain.

Extended Summary

In this interview, host Kieran Chandler speaks with Joannes Vermorel, the founder of Lokad, a software company specializing in supply chain optimization, and Pierre Khoury, the CEO and co-founder of Shippeo, a European leader in supply chain visibility. They discuss the importance of real-time visibility in supply chains, the challenges faced, and how technology is being used to solve these issues.

Pierre Khoury begins by sharing his background as an engineer, his experience in finance, and the founding of Shippeo four years ago. The company’s aim was to bring real-time visibility to road transportation, which has been a black box for the past 40 years. Shippeo now has 70 employees and is a leader in tracking road transportation across Europe.

Joannes Vermorel explains the cardinal rule of optimization: you cannot optimize what you do not measure. Supply chains are difficult to optimize because they are highly distributed and spread across continents. Historically, people have tracked static items, such as what is in warehouses. However, supply chains involve tracking moving items like trucks and shipments. Real-time visibility is crucial to accessing this information without waiting days for data to reach headquarters, allowing for better optimization.

Pierre Khoury details how Shippeo’s technology works in practice. The company has developed more than 150 connectors to various systems, such as telematics and dispatch systems for carriers, as well as mobile apps for smaller carriers. Shippeo aggregates data and GPS points every three to five minutes, regardless of the source, and then uses predictive analytics to determine estimated times of arrival.

Joannes Vermorel emphasizes the importance of real-time visibility from a supply chain optimization perspective. Knowing the status of shipments allows for better decision-making, such as whether to purchase more or change suppliers. For example, if a supplier is expected to deliver products in ten days but the shipment is delayed, real-time tracking through Shippeo can help identify the issue and allow for mitigation strategies, such as placing an emergency order with a different supplier to avoid production delays.

Real-time visibility in supply chains is essential for better optimization and decision-making. Shippeo’s technology is helping to solve the long-standing problem of tracking moving items in road transportation, enabling companies to make more informed choices and mitigate potential issues.

Joannes Vermorel explains that the primary challenge they face is the limited information they receive from transactional systems like ERPs, WMS, and order management systems. These systems do not provide information about what happens after a purchase order has been sent, creating uncertainty in the supply chain. He believes that while probabilistic approaches will still exist, the uncertainty can be greatly reduced with real-time data from companies like Shippeo.

Pierre Khoury discusses the technical challenges they face, such as connecting over 150 different systems, ensuring a stable and consistent data flow for millions of GPS points, managing change in the slow-to-adapt transportation industry, and maintaining high-quality data pipelines. He highlights the importance of considering factors like traffic, driver breaks, loading and unloading times, and weather to create accurate estimated times of arrival.

Both Vermorel and Khoury agree on the high value of real-time data in supply chains, as it allows companies to track expensive physical assets and avoid costly delays or disruptions. Vermorel emphasizes that even if the assets themselves are not expensive, their availability or unavailability can have a significant impact on the supply chain.

Khoury provides an example of how Shippeo’s real-time data helped a customer during the Yellow Jackets protests in France, allowing them to reposition their trucks and avoid costly helicopter or plane deliveries. He also mentions that the level of granularity in their data, with updates every three to five minutes, can be beneficial for all parties involved in the supply chain, including warehouses, customers, transport teams, and customer service teams.

Vermorel explains that problems frequently occur at the boundaries of supply chains, where delays can occur due to various reasons, such as goods not being picked up on time or miscommunication between different systems. He emphasizes the importance of fine-grained lead times, which can provide valuable information for optimizing supply chain processes. Vermorel notes that while real-time data is important, the analysis does not necessarily need to remain at the minute level; for example, network-wide optimization might involve thinking in terms of days rather than minutes.

Khoury adds that supply chains are inherently composed of several actors, who often have a siloed approach to information. By sharing real-time information on transportation, Shippeo aims to break down these silos and improve overall efficiency. He provides an example of how up-to-the-minute information can help reduce waiting times at warehouses, leading to lower transportation costs and better productivity for all parties involved.

When asked about Shippeo’s main customers, Khoury explains that their primary focus is on shippers, as they are the ones who benefit most from the solution and can drive its adoption throughout the supply chain. The system is provided free of charge to carriers and other stakeholders to encourage widespread participation.

Vermorel highlights the importance of real-time visibility in supply chain optimization, explaining that it can help eliminate much of the uncertainty that complicates decision-making. Probabilistic approaches can be useful for dealing with uncertainty, but reducing it as much as possible is still highly desirable. He envisions Lokad being able to offer more efficient supply chain decisions at various levels, from long-term purchase orders to tactical, emergency responses.

Khoury concludes by emphasizing the growing interest in real-time visibility in supply chains, with Gartner ranking it as a top topic. He believes that sharing information with other systems can create additional value and foster a collaborative ecosystem that benefits everyone involved in the supply chain.

Full Transcript

Kieran Chandler: Today, I’m delighted to say we’re joined by Pierre Khoury, the CEO and co-founder of Shippeo, who’s going to tell us a little bit about both the positives and also some of the challenges of real-time visibility in supply chains. Pierre, thanks very much for joining us.

Pierre Khoury: Thank you, Kieran. I’m very happy to be here.

Kieran Chandler: Nice. So if you could just perhaps start us off by telling us a little bit about your background and also just telling us how Shippeo came about.

Pierre Khoury: Yeah, so by training, I was an engineer, and I went through finance where I co-founded a private equity fund. Then, four years ago, we started Shippeo with the aim to have real-time visibility over transportation, especially road transportation, which has been a black box for the last 40 years. So now, Shippeo is a 70-employee company, and we are the leader in Europe for tracking road transportation all over the continent.

Kieran Chandler: And, as always, it wouldn’t be LokadTV without Joannes Vermorel. So, Joannes, perhaps you could just tell us a little bit more about what is real-time visibility with regards to supply chains.

Joannes Vermorel: The cardinal rule of optimization is that you cannot optimize what you do not measure. So if something remains opaque to you, then you have no hope ever to optimize anything. In supply chains, it’s very difficult because it’s highly distributed, with things spread over literally continents. Historically, people started to track what did not move, like what you have in warehouses. As long as things are static, it was not easy, but it started several decades ago. Now, the problem is that supply chains are all about the things that are moving, like trucks and shipments, and there it becomes very difficult. Real-time visibility is about accessing this information without having to wait for days, so that all the data can flow back to headquarters. It’s a challenge because if you do not do that well, you lose the opportunity to optimize just because it arrived too late. So, real-time visibility is very important in this aspect.

Kieran Chandler: Okay, so that’s kind of the theory behind it. Pierre, could you tell us a little bit more about how it actually works in practice? How does the technology work at Shippeo?

Pierre Khoury: As I said, it was a challenge and unsolved for the last 40 years. So, we have organized ourselves to first solve the heterogeneity of carriers’ IT systems. We developed more than 150 connectors with various types of systems, like telematics dispatch systems for carriers and our mobile app for small carriers. We aggregate data and GPS points every three to five minutes, whatever the source is, and then we can do predictive analytics on that, especially estimated time of arrival.

Kieran Chandler: And Joannes, why is this of interest from a Lokad perspective? Why is this interesting?

Joannes Vermorel: For us, when we want to decide whether we want to purchase more or need to maybe purchase from a different supplier, in order to optimize this decision, it depends on what you have or what you most likely will have by a certain amount of time.

Kieran Chandler: So, for example, if you realize that a supplier can deliver products in ten days, but they are going to be late by five days, and thanks to Shippeo you already know that because the tracking tells you that the truck has barely moved for the last couple of days. What does that mean?

Joannes Vermorel: It means that maybe you can decide to mitigate what would become a production incident, or your manufacturing plant would stop just because you don’t have the raw materials, by placing an emergency order with a supplier that is maybe more expensive but is super close and can deliver what you need the next day. But if you want to have super agile emergency correction measures, you need to have this data. For us, the big challenge is that frequently all we have is data that comes from transactional systems like ERPs, WMSs, or order management systems, which do not give any information besides where things have been sent. Frequently, we don’t know more than the purchase order has been sent, and what happens after that, we don’t know. Ultimately, we will know that it has been received, but it will be very, very late.

Kieran Chandler: So, would that sort of replace the probabilistic approach we’re taking to lead times? Is that going to be completely eradicated if we had customers who are using Shippeo?

Joannes Vermorel: I think probabilistic approaches will remain because if you’re halfway through a delivery that is supposed to take a couple of days, when you are on day one, you still have some degree of uncertainty that remains for the last four days. But the question is that this uncertainty can be vastly reduced, especially as time goes on. The interesting thing about Shippeo is that if you have something that is supposed to take five days for transport, like crossing all of Europe, at day four, if your truck is where you expect it to be, the remaining uncertainty is almost nothing. In contrast, without Shippeo, we’re at day four, and we don’t know anything more than what we had at day zero because the truck is not there yet, that’s for sure, but whether there is a delay, we don’t know. So, we still have this very uncertain situation, whereas with Shippeo, at day four out of a delivery that is supposed to take five days, you have almost no uncertainty left, which is excellent and would lead you to better optimization just through reduced uncertainty.

Kieran Chandler: So it’s basically all about refining those probabilities. Pierre, could you tell me a little bit more about some of the technical challenges you face? I mean, when working in real-time, it can’t be very straightforward. What are some of the challenges in place there?

Pierre Khoury: Sure, I would say there are three main challenges. The first challenge, from a technical perspective, is to connect more than 150 different systems and gather that into a single canonical data model in real-time, and to have a stable and consistent data flow for millions of GPS points every day. The second challenge is about change management. Transportation is usually very slow to change, and it’s a big challenge to make it happen, to onboard carriers, users, train them, and achieve the results we want. The third challenge is about the quality of data and having the right rules to fetch the data and to have something consistent and of high quality.

Kieran Chandler: Joannes, would you agree that there’s a need for this data? I mean, it seems like nowadays there’s so much data being collected. Do we really need this in supply chains? I mean, I could probably sneeze, and somebody somewhere would be collecting that data. Do we really need this in supply chains? This is highly valuable data. We’re not talking about random tweets; we’re talking about trucks, and one truck will typically have several hundred thousands worth of euros of merchandise. If you’re transporting electronics or anything that is non-trivial, assuming you’re not transporting earth or dirt or sand, what you’re moving across those supply chains is extremely valuable. So, does it make sense to collect this data?

Joannes Vermorel: Yes, it does make sense because you’re tracking assets that are worth literally tons of money. You are frequently in highly asymmetrical situations where maybe what you’re delivering is not that expensive, just small parts that are the price of the metal. But if you don’t have them, and you have a whole manufacturing plant that stops working, then you have hundreds of people who cannot work anymore, just because you’re on hold until you have those repair parts that are being delivered. The data is incredibly valuable because it can be attached to physical assets that are expensive, but also, even if they are not expensive by themselves, their availability or unavailability can have very high costs down the line for other elements of your supply chains.

Kieran Chandler: So, where do we draw the line in terms of the data that you’re actually collecting? It’s obviously good to know where a truck is at any given moment, but are you keeping an eye on traffic reports? Are you even taking into consideration the weather? Where do you draw the line?

Pierre Khoury: We need to factor all these issues to have a reliable estimated time of arrival, considering traffic, breaks for drivers, legal rest times, and the patterns of loading and unloading. We factor all this into our machine learning algorithm to have something reliable. I think we are building one of the best estimated time of arrival algorithms in Europe. Just to give an example of what Joannes said, one of our customers sent us an email to thank us in December because of the Yellow Jackets’ protests. Traffic was not running as usual, but they were able to anticipate that, reposition their trucks, and mitigate the risk within their production plan. They experienced zero effect of the protests, and on the opposite, they could have been forced to use helicopter or plane deliveries for their products at a very high cost. It’s a good example of the value we bring.

Kieran Chandler: The Yellow Jackets is definitely a very specific example. So, you mentioned you have a level of granularity of updates every three to five minutes of data, is that correct?

Pierre Khoury: Yes, that’s correct. The new boundary we are opening is real-time data, thanks to GPS points. We gather data every three to five minutes, depending on the carrier’s system. It’s very granular and allows us to see a lot of things, like loading times, unloading times, and problems that can be encountered on the road. This data is valuable for all parties because the idea is not to restrict access to this data, but to share it with the right rules and granted access to other parties, like the warehouse, the end customer, the transport team, and the customer service team. In the end, everyone can be more productive and efficient because they know what will happen. I’d like to mention Amazon because, in the end, all customers, even for B2B deliveries, want to have the same experience as with an Amazon B2C parcel.

Kieran Chandler: Thanks, Amazon is the number one thing we seem to mention every week. When we’re not talking about terms of minutes, we’re more talking about in terms of hours and more really days. Is this level of granularity a bit too much? Do we really need this level of granularity?

Joannes Vermorel: Yes, because frequently, especially for transport, we don’t need it for everything, but in transport, you have to estimate the amount of problems that can happen in the supply chain at the boundaries. For example, you can have a delivery that is made to a warehouse, but then the merchandise is only picked the next day or eight hours later. A lot of things can happen really at the boundaries. Most of the time, goods are not moving or are not transformed. So, those thin-grain lead times make sense, supply chain-wise. We are not trying to have sub-millisecond measurements like if you were trying to optimize packet distribution over the internet. The granularity is more like a couple of minutes that makes sense when you have a physical operation that needs to happen.

And indeed, once we have this information, we can detect a lot of things, but it doesn’t mean that all the analysis that will flow will remain at the minute level. For example, if you’re trying to optimize and compress the times that are end-to-end several weeks, you have measurements in minutes that will let you pinpoint tons of things, probably patterns that are slightly dysfunctional, especially at the boundaries of the systems. The main problem in supply chains is when something passes from one system to another, that can be software, a company, or teams. That’s where the gaps can happen, and then you have accidental delays. You can remove them, and typically, when you want to do network-wide optimization, you’re going to shift to something that is more like optimization, where you think in terms of days rather than minutes. But it really depends on the type of problem.

Pierre Khoury: Just to complete what Joannes said, by definition, supply chain is a chain, meaning that there are several actors within the chain, and these actors today have a very siloed approach and information. What we do is share the information on transportation to have all the actors having the right information and break these silos to optimize. And yes, sometimes, information up-to-the-minute is important. For instance, to know that within a warehouse, there are two hours of waiting at loading time and two hours of waiting at deliveries. With the information about where a truck is and when it will arrive, you can reduce that by up to 50% and have fewer waiting times, meaning fewer transportation costs, fewer stock costs, and everyone is more productive. It’s the same for carriers because they’re not waiting for nothing. We think this kind of example is a good example of a win-win-win situation for everybody.

Kieran Chandler: Let’s talk about these different actors along the chain. Who should the emphasis be on? Who should be Shippeo’s main customer? Should it be the retailer who’s actually ordering the stock, or should it actually be the shipping companies themselves?

Pierre Khoury: Our model is based on the shipper approach. We think that it’s the actor who will reap one of the main benefits and would be pushing the solution towards change. So, our customer is the shipper, and the system is free for all the stakeholders, including the carrier. The system is 100% free for the carriers, and we try.

Kieran Chandler: To involve as many actors as possible. Okay, great. And Joannes, we’ll start sort of concluding things a bit today. Other than just sort of lead times, why is there so much interest in looking at this? I mean, how can you see the two tools working so well together?

Joannes Vermorel: Right now, we are frequently making, I would say, wild guesses concerning the state that we are trying to optimize. Yes, we are measuring things as much as possible by retrieving the stocks that we have, but the reality is that all those things that are on order, where you have quantities on order or in between locations, very frequently, it’s just opaque. It needlessly complicates because it forces us to have guesstimates of the state of the system. Not only does that complicate the modeling, but it also makes the optimization less efficient. The more uncertainty, the more you need a probabilistic approach to survive numerically speaking. However, that doesn’t mean that uncertainty is desirable. Especially when you have the option to remove almost all of the uncertainty by having something that is still probabilistic but much more compressed, it makes everything easier. There are entire classes of corrective actions, super tactical corrective actions, that are only possible if you do that. Frequently, right now, there could be corrective actions that we could suggest in theory to clients with very little effort, considering that we have already done the work to integrate data from their ERPs and many other sources. But lacking the real-time status on the shipments, we cannot perform those optimizations and generate corrective suggestions. In the future, I see more and more capacity for Lokad to deliver optimized supply chain decisions at many levels, some for a purchase order in Asia, and some being very tactical, such as emergency replenishment needed by placing an order right now to a nearby supplier that is much more expensive for a small quantity that will help you survive the delay.

Kieran Chandler: Okay, we’ll leave the last word to Pierre, as our guest. What is the key lesson you want people to take away for real-time visibility within supply chains?

Pierre Khoury: I want to say that I think real-time visibility is one of the hottest topics in supply chains these days. It’s not just me saying that; Gartner has ranked it as a top one topic. This talk is very interesting to see how we can create value while combining with others. Our mission at Shippeo is really to focus on real-time visibility, aggregating data, and giving predictive insights on that. But we see that if we share information with other systems, for sure, approved by the shipper, it creates another level of value. This kind of ecosystem is what we like to bring value to the customers and to the world supply chain.

Kieran Chandler: Okay, we’re going to have to leave it there, but thanks for your time this morning.

Pierre Khoury: Thank you so much.

Kieran Chandler: That’s everything for this week. Thanks very much for tuning in, and we’ll be back again next week with another episode. Until then, thanks for watching.