00:00:08 Challenges in managing inventory for a network of grocery stores.
00:00:35 Introduction of Richard Lubienski, expert in supply chain and retail.
00:01:30 The difficulties of forecasting fresh food and managing waste.
00:03:00 Importance of tracking remaining product life to optimize inventory.
00:07:58 Discrepancies between advanced software claims and actual implementation.
00:10:11 Discussing challenges in optimizing product freshness in grocery stores.
00:12:24 Importance of network optimization in grocery stores.
00:15:06 The impact of promotions, price, and product substitution on supply chain management.
00:18:36 Introduction of All Futures, a company focusing on supply chain optimization.
00:19:15 Exploring the vision of All Futures and its approach to tackling supply chain challenges.
00:20:05 Discussing the benefits of considering all future possibilities in supply chain management.
00:21:33 Importance of bananas in supermarkets and how probability curves can optimize supply chain decisions.
00:23:01 Managing fresh supply chain, constraints, and the role of events like Christmas and Easter.
00:24:07 Critique of traditional retail supply chain strategy and reliance on armies of clerks.
00:27:10 Exploring better ways to make decisions in supply chains, focusing on prioritization and network perspective.
00:29:55 Discussing SKU statistics and how supply chain costs can be reduced.
00:31:18 The challenges of joint optimization due to organizational silos.
00:33:43 The need for inspired leadership to implement change in supply chain management.
00:35:46 The market’s future in the next 5-10 years and the complexity of omni-channels.
00:38:01 Incorporating economics and customer loyalty in supply chain decision making.
Summary
In an interview, Joannes Vermorel, Lokad founder, and Richard Lubienski, All Futures Managing Director, discuss supply chain optimization for perishable goods. They highlight the challenge of balancing overstocking and understocking to maintain profitability and customer satisfaction. Both emphasize the importance of product freshness, network optimization, and data-driven approaches. They also address the need for breaking down departmental silos and incorporating economic considerations and customer loyalty into supply chain strategies. While advanced analytics and machine learning can enhance supply chain management, they stress the importance of maintaining a human touch and adapting to external factors like changing consumer preferences and environmental concerns.
Extended Summary
In this interview, Joannes Vermorel, the founder of Lokad, a supply chain optimization software company, speaks with Richard Lubienski, Managing Director of All Futures, who has over 30 years of experience in the supply chain industry, with a focus on retail and fresh food supply chains in Australia. The discussion covers the challenges faced in managing inventory for perishable goods at a large scale and some key aspects for success.
The primary challenge in managing fresh food inventory is perishability. High service levels can be achieved by overstocking, but this can lead to waste and significant financial losses. On the other hand, insufficient stock results in poor customer satisfaction and missed sales opportunities. Forecasting fresh food inventory is complicated by the fact that it is not all fast-moving, and there are products with varying rates of sale.
Walking the tightrope between overstocking and understocking is crucial to maintaining profitability and customer satisfaction in fresh food retail. Lubienski emphasizes the importance of keeping goods moving from source to customer while ensuring freshness.
They discussed the challenges of inventory optimization in supply chain management. Vermorel believes that many software tools in the market are “blind” to the importance of the freshness of products in inventory. He explains that having the exact information about product freshness can help optimize the stock level in stores and improve sales. However, optimizing stock levels becomes more complex when dealing with a network of stores, as weather and seasons can affect product supply. Lubienski provides an example of the challenges of managing fresh produce in Australia, which is a continent-sized country with a domestic supply chain. He notes that replenishment decisions have to be made in advance and must be done as close to the last possible moment to ensure that every product is placed in a location where it will sell, rather than go to waste. Vermorel and Lubienski agree that network optimization is an ongoing challenge, but one that can be tackled through the use of advanced algorithms and a data-driven approach.
They discuss the challenges and frictions that come with supply chain optimization. They touch on issues such as variable customer demand, substitutions, and pricing in fresh produce. Lubienski shares his vision for All Futures, which he believes offers a better way to solve supply chain issues. He discusses the importance of probability curves in forecasting and how they can differentiate the value of products. Lubienski also critiques the current supply chain strategy in mainstream retail, which he believes has not evolved much since the late 80s and early 90s, and calls for a better way of decision making that prioritizes aggregated decisions and flows. He highlights the potential benefits of such a solution for the network.
Discussion related to supply chain optimization from a network perspective. Vermorel and Lubienski discussed the importance of prioritizing correctly in supply chain departments, which can often be quite siloed between forecast teams and logistics teams. One group may focus on service level and forecast accuracy, while the other is heavily cost-focused. However, if departments can work together to reshape tasks across the network, there is a huge opportunity to reduce supply chain costs. For example, a typical supermarket may have 20 to 30,000 SKUs, but 70% of those products will sell less than one unit a day and replenish less than one carton per week. By smoothing the task across the network, supply chain costs can be reduced, creating a new horizon for the supply chain industry.
However, Vermorel and Lubienski noted that the silos within companies can make it challenging to consider this sort of joint optimization, where multiple problems that belong to multiple departments are brought together and reoptimized with computers to get something superior to anything that could be done in isolation for every segment. Inspired leadership is needed to shake up the traditional way of working within departments. The change requires a program manager who can manage the transition states between the current way of doing things and where they want to be in the future. While it may be challenging, the prize of optimizing supply chain costs, shareholder value, safety, environment, and waste reduction is worth pursuing.
The discussion was the future of the food retail network in the next five to ten years and the challenges the industry faces. They covered topics such as the linear supply chain concept, economic considerations in supply chain management, customer loyalty, the pace of change in the industry, and the academic field of supply chain management.
Lubienski expressed concerns about the linear supply chain concept, finding it confusing and inadequate in addressing the uncertainties, inflation, and the growing prevalence of omnichannels. He proposed that incorporating economic considerations into supply chain decision-making and optimization would lead to better outcomes. By using customer loyalty as a factor in the economic equation, companies can make more informed decisions about their supply chain strategies.
The conversation highlighted the slow pace of change within the food retail industry, particularly in the area of supply chain management. Lubienski noted that progress in the academic field of supply chain management has been limited, which may contribute to the slow adoption of new ideas and approaches. He suggested that as mainstream software companies begin to embrace these concepts, the industry as a whole will be more likely to accept and implement them.
Throughout the discussion, both Vermorel and Lubienski emphasized the importance of innovation and the need for the food retail industry to adapt to the changing landscape. They acknowledged that overcoming the challenges of the linear supply chain concept and integrating economic considerations into supply chain management could lead to more efficient and effective supply chain strategies. Additionally, the incorporation of customer loyalty into decision-making processes could improve overall business performance.
As the interview progressed, Vermorel and Lubienski touched on the role of technology and software solutions in the evolution of the food retail industry. They discussed the potential benefits of employing advanced analytics and machine learning techniques to optimize supply chain management, including demand forecasting, inventory control, and transportation planning.
However, they also recognized the need for a balance between adopting new technologies and maintaining a human touch in the industry. Vermorel and Lubienski agreed that while technological advancements can provide significant benefits, it is essential for companies to remain focused on customer satisfaction and personal relationships.
The discussion also explored the potential impact of external factors on the food retail industry, such as changing consumer preferences, environmental concerns, and geopolitical tensions. Vermorel and Lubienski acknowledged that these factors could present both challenges and opportunities for businesses in the sector, underscoring the importance of flexibility and adaptability in supply chain management.
Full Transcript
Joannes Vermorel: Managing inventory in a single grocery store is difficult, especially if you want to keep your customers happy while minimizing the amount of waste. Doing that at scale with a network of grocery stores is even more difficult. And if you want the epic scale version of the challenge, you try to do that competitively in a country that has the size of a continent. So today, it’s a pleasure to welcome Richard Lubienski, who has been doing precisely that for a decade, I believe. We are in Paris, and you are joining us from Australia. It’s a pleasure to have you, Richard. Could you tell us a little bit about yourself?
Richard Lubienski: Firstly, thank you for having me here. It’s an honor to be here in Paris with Lokad. My background is that I’ve spent approximately 30 years working in supply chain, in manufacturing, but really the last 20 years or so in retail, and the last 10 in fresh supply chain with supermarket chains in Australia. I’ve had the privilege to both run the forecasting replenishment operations, meaning flowing the stock from suppliers to stores, as well as running the programs to develop systems to do that as automatically as possible.
Joannes Vermorel: The obvious challenge when it comes to fresh food is perishability. Having a very high service level is relatively straightforward: you pile everything super high. Obviously, if it’s fresh, it’s a highly rotating, easy-to-predict product, so you can go with super accurate forecasts, and then everything will go according to plan, and you will have no stock out. And yet, everything will rotate just enough so that you have zero waste as well. So, was this your experience? Is this the way it’s done?
Richard Lubienski: It’s not exactly like that. In fact, it’s not really like that at all. Obviously, being fresh food, the key differentiator to the ambient groceries in the store is the perishability and the fact that if you overstock, you waste product, and that is extremely costly. So, being right in fresh or having the right decisions is almost like walking a tightrope, where too much loses you money and literally loses you extraordinary amounts of money. And obviously, not having availability is bad for sales and bad for customers in many ways as well.
You also can’t be precise about it, so forecasting fresh is not all fast-moving normal distributions and so on. There are slow-moving products, as merchants try to maintain an interesting range for customers, always trying new things. There are products that won’t sell a carton within their life, unfortunately, but there are also products that will sell 10 or 20 cartons within their life. So, you have it all. Having the right balance is definitely a challenge. Flowing the goods from source to customer, keeping it fresh, and really keeping it moving is the key challenge.
There are a few aspects to that, which are, in my view, absolutely essential to being successful. One of the things that is interesting is that when I got interested in this topic more than a decade ago, the textbook answer to this challenge, let’s say a safety stock analysis, entirely ignored the perishability aspect. In that sense, the textbook supply chain approach results in those problems.
Joannes Vermorel: Managing inventory in a single grocery store is difficult, especially if you want to keep your customers happy while minimizing the amount of waste. Doing that at scale with a network of grocery stores is even more difficult. And if you want the epic scale version of the challenge, you try to do that competitively in a country that has the size of a continent. So today, it’s a pleasure to welcome Richard Lubienski, who has been doing precisely that for a decade, I believe. We are in Paris, and you are joining us from Australia. It’s a pleasure to have you, Richard. Could you tell us a little bit about yourself?
Richard Lubienski: Firstly, thank you for having me here. It’s an honor to be here in Paris with Lokad. My background is that I’ve spent approximately 30 years working in supply chain, in manufacturing, but really the last 20 years or so in retail, and the last 10 in fresh supply chain with supermarket chains in Australia. I’ve had the privilege to both run the forecasting replenishment operations, meaning flowing the stock from suppliers to stores, as well as running the programs to develop systems to do that as automatically as possible.
Joannes Vermorel: The obvious challenge when it comes to fresh food is perishability. Having a very high service level is relatively straightforward: you pile everything super high. Obviously, if it’s fresh, it’s a highly rotating, easy-to-predict product, so you can go with super accurate forecasts, and then everything will go according to plan, and you will have no stock out. And yet, everything will rotate just enough so that you have zero waste as well. So, was this your experience? Is this the way it’s done?
Richard Lubienski: It’s not exactly like that. In fact, it’s not really like that at all. Obviously, being fresh food, the key differentiator to the ambient groceries in the store is the perishability and the fact that if you overstock, you waste product, and that is extremely costly. So, being right in fresh or having the right decisions is almost like walking a tightrope, where too much loses you money and literally loses you extraordinary amounts of money. And obviously, not having availability is bad for sales and bad for customers in many ways as well.
You also can’t be precise about it, so forecasting fresh is not all fast-moving normal distributions and so on. There are slow-moving products, as merchants try to maintain an interesting range for customers, always trying new things. There are products that won’t sell a carton within their life, unfortunately, but there are also products that will sell 10 or 20 cartons within their life. So, you have it all. Having the right balance is definitely a challenge. Flowing the goods from source to customer, keeping it fresh, and really keeping it moving is the key challenge.
There are a few aspects to that, which are, in my view, absolutely essential to being successful. One of the things that is interesting is that when I got interested in this topic more than a decade ago, the textbook answer to this challenge, let’s say a safety stock analysis, entirely ignored the perishability aspect. In that sense, the textbook supply chain approach results in those problems.
Joannes Vermorel: You’re actually on the business side, and you examine why you’ve got stockouts, and look at, day after day and weeks after weeks, why there is a gap. You can see very clearly if you put your charts of delivery, sales, waste, and replenishment in the store. All the ups and downs for your stocking store are very apparent. What’s happening? Can’t read the brochures, as you know, can’t read the versions. So, it’s not solved. I’m aware of people working on that right now, in fact.
Richard Lubienski: You know how we solved it in previous companies? It was probably complicated. It was a bit of a model, it wasn’t precise, but actually, now, nowadays, I know in Australia, Woolworths, for example, this is common knowledge, have QR codes rather than barcodes and contain an expiration date on that. You actually can’t take an out-of-date product through the self-checkout and buy it. It’ll say red light, which actually means that you don’t have to model nowadays. You don’t necessarily have to model what the life of the stock is and make estimates and so on; you could actually take a fact. I’ve sold one with an expiration date of the 28th or 27th because of customer behavior. I’m sure all of us will put our hand to the back and get a fresher one unless we know we’re eating it today.
Joannes Vermorel: That’s very interesting because if we start having the exact information about the freshness that is being picked by the customers, it gives an extra dimension in terms of modeling, I would say, the willingness to buy, willingness to pay. So, you have the products, the price point, and the freshness, which means that you can stop, you know, potentially, when you want to think about doing the right mix of the best price, best products, best freshness. It becomes possible to potentially start tackling the very interesting problem of putting all those things together because you don’t have to guess what people are actually doing in terms of cherry-picking, you know, the product that has the biggest lifetime, the biggest amount of days of freshness in it. Or on the contrary, if customers are kind of happy because they are just going to consume the products right away, so they don’t really mind if they still have one week or ten days’ worth of product life cycle.
Richard Lubienski: Yeah, the other end of that question, at the beginning of that question: Do you know what product life you’re putting into a store? Because typically, you’ll pick a carton and put it on and not do a transaction in your DC system that records the date. However, you generally know what pallet a carton has been picked from or very close to it, where the pallet that’s been in the pick face, and you have the data associated with the pallet level. I’m receiving either an expiration, best before date, or in terms of loose product or a packed-on date, and that’s a really good start for solving that problem.
Joannes Vermorel: And so, we have the end goal of actually optimizing, just as we discussed, the stock level in the store with all the freshness being one of the biggest, very specific concerns. I mean, there are freshness and lifecycle issues in other industries, even in aerospace, but obviously, this is much more dominant in grocery stores.
Joannes Vermorel: There are freshness and life cycle issues in other industries, even in aerospace, but obviously, this is much more dominant as a concern as far as grocery stores are concerned. But also, as soon as we start looking at the problem from a network perspective, there are other challenges. In your experience, what are the big challenges if you want to do that efficiently, I would say through the angle of the network? What are the extra gotchas when you start thinking of optimizing not one grocery store, but actually a network of grocery stores?
Richard Lubienski: That’s a really good point. So, within Australia, let’s talk about fresh produce for a moment. Australia is the size of Europe and has about five major cities, ranging from half a million people to maybe eight million people, with a 25 million population in total. There’s at least a thousand kilometers between each capital city, and not much in between. About 97% of the fresh produce is grown in Australia, not always in the same location because being a continent, seasons change and climates change, so that can move around as well. But what it means is we’re tied into a domestic supply, which means a couple of things. When weather is adverse and supply is tight, you can be short of a product that you’d ideally want and the customers want, but you just can’t get it, and it’s not practical to import it in the quantities that you would want.
The other side of that, as well as the shortage or cut scenarios, is the push scenario where you’ve brought stock into the distribution center (DC), but you must put it in stores because you cannot sell it from the DC. Every day it’s in the DC, it loses value, and you need to give enough life to stores to sell it in a reasonable time. So, you have to do this on a network basis. You generally have maybe a daily time to stores, but you may have two or three or even more days lead time for getting products into a DC. Your DC replenishment decisions have to be done in advance, and you’ll have many replenishments into store within the window of potentially one DC replenishment.
What you really need to do is postpone the decision to the last possible moment of what you’ve got in your shed and how you decide whether to put it or hold it. This is a subject that we’ve worked on with some good colleagues and come up with a deterministic algorithm to sort through that. We call it pushes and cuts, and the objective is pretty simple: you want to put every item in a place where it’s going to sell and not in a place where it’s going to waste, and you want to do this for every item and every decision. I would say we had a fair attempt at that, but you always move on and learn, and some of this was a few years ago now. You’d rethink how you solve that.
Typically, in the large retail networks, at least in Europe, I’m not 100% sure about Australia, they complicate the life of the supply chain folks by…
Joannes Vermorel: Being very active on the promotional side, there are lots of pushes and slowdowns. We already have many variables that are not properly under control. Customers do whatever they want, and producers don’t have entire control over their harvest. The weather is also unpredictable. On top of that, we have all the promotional activity layered on top, and we have to orchestrate the flows. In your experience, what sort of challenge or friction comes with that in this vertical?
Richard Lubienski: That’s a good question. Promotions are clearly an obvious factor, and price is a key input into demand. Regular price in fresh produce, at least in Australia, is not a fixed price for the year. The cost price can change every week, depending on availability. Our approach involves selling in markets as well as to supermarkets, which can be an ongoing, weekly discussion. This obviously changes demand.
Other aspects within fresh produce include substitutability. There’s some obvious substitutability, like one apple versus another apple or one mango versus a different variety, but also for the customer, it’s about choosing a fruit snack, like an apple versus a banana versus a punnet of berries, where price comes into play. Substitutions can be more obvious or subtle, like choosing between bok choy, pak choy, or choy sum for a stir fry. You have to incorporate those into your knowledge and look at them at both an aggregated and individual level. Balancing out shortages and excesses within a group like that is pretty important.
Joannes Vermorel: You’ve been working at Coles for a decade, managing these changes, and more recently, you started a company with a very appealing name to me: All Futures. Could you tell us a little bit about the vision you have for this venture and what you’re trying to do?
Richard Lubienski: All Futures is not an accident in the name. It’s partly inspired by reading your vision about three years ago when I first came across you and found it very inspiring. All Futures relates to the fact that our forecasts aren’t just median forecasts. I’ve worked in a world where we had teams of forecasters moving lines on charts, looking at history and trying to be right.
All Futures represents the idea that anyone who’s sat in that seat moving the line knows it’s just their best guess. A more accurate view is that it’s not about selling exactly 700 kilos of bananas in a shop on Saturday, but rather, a probability curve is a far better representation of what’s going to happen.
Joannes Vermorel: And that may be the mean, and you know, maybe I’ve got no chance of, you know, practically no chance of selling zero, although you can do because you can have a flood or the football game can take over the entire car park. So those things are actually real, and on the high end, but that is immediately a far better representation of what your future looks like. So, hence All Futures, if you really want to work that supply chain for that and across not just one product but all products, then combining that together into one solve, into an optimal solve, is a brilliant junction that is clearly a better foundation to start with.
Richard Lubienski: My second thought, by the way, after I first came across this idea was that none of the software I know is going to deal with that concept at all, not just at the forecast level but then, what do you do with it? And I think I wrote you three years ago and said I now need to take six months off and have a really good think about how this is. But what still excites me, and it’s going to excite me for the next decade at least, is that this foundation is a solve, it is the start of the solution for so many of these issues. And when you can find an idea that can be leveraged so widely, that’s really beautiful and uncommon. It’s much, you know, it’s elegant from an engineering background, it’s an elegant solution, much better than solving this problem, this problem, this problem, hoping that you add them all together and you haven’t got a monster.
So for example, having that probability curve, and I’ll talk about bananas again, it differentiates the value of the first carton of bananas I sent to that store and the hundredth carton. This one will sell, this one is more than likely going to be carried over and is there for display purposes, that high value, lower value. If you add economic drivers over the top of that, you can now compare the value between products, so the hundredth carton of bananas versus the first carton of cucumbers. Well, bananas are more important than cucumbers from the sales point of view, number one selling SKU in every supermarket in the world. A hundredth one, which isn’t going to sell, versus the first one, this one is more important. So if I have a truck that can take a hundred cartons, the right decision is not to put 100 cartons of bananas on it, it’s probably to put maybe 97 and three of cucumbers. That’s what’s going to sell and keep the customer happy.
So that’s a really powerful way to firstly look at pushes and cuts, it’s a natural solution for that, and also as a start into managing constraints, which is the other key feature of your life. I think managing fresh supply chain is, it’s not so much BAU with no holidays or anything like that. The volume that, of fresh in particular, that points towards events like Christmas, Easter, in particular where you have complete shut, no sales days, is extraordinary. The customer demand spikes in the extreme, you know, maybe three times the normal volume that you’re trying to flow through.
Joannes Vermorel: And again, with this All Futures adventure, the interesting thing is that you’re trying to shake and challenge a bit the status quo. I mean, my perception is that mainstream retail, general retail, is very much rigidified around a supply chain strategy which is divide…
Joannes Vermorel: And it seems that many supply and demand planners are managing anywhere from one to a couple hundred products or SKUs, depending on how they slice and dice the SKUs. It’s intriguing that this industry seems to have solidified a setup that originated in the late 80s and early 90s, almost three decades ago. People have spent their entire careers looking at a few hundred lines every single day, revisiting the same spreadsheet, or potentially a webpage now, because the system can be online. Fundamentally, they are doing pretty much the same thing as they were doing with a spreadsheet, even if they have a web-based app to do it. My question is, what was your perception about the added value of having an army of clerks revisiting guesses all year long, each clerk being responsible for their own small scope? The entire industry seems to operate by piling up armies of clerks to try to optimize these networks of grocery stores.
Richard Lubienski: The first thing I’ll say is that it’s pretty obvious the scale of a major national supermarket chain involves tens or even hundreds of millions of dollars a day in sales. The desire to have availability and satisfy customer requests is intense, and what you can gain or lose from this process still outweighs the cost of employing hundreds of people to work on it. Demand planners do learn and start to understand the subtle drivers that create demand for a product when they look at it day in, day out. But the second thing it says is that your software doesn’t work particularly well. It’s not an unusual experience for a lot of people who have worked in manufacturing and supply chain for my lifetime that even textbooks are ahead of software. The world has changed, and it’s time to revisit how these solutions are made.
Joannes Vermorel: I can understand why that approach has been taken, but there must be a better way. A better way is not about being precise, but about having a good, aggregated set of decisions and flows. When you start looking at this decision from a prioritized perspective, which is very much the Lokad way of tackling the problem, what do you see as other potential changes or improvements?
Joannes Vermorel: We’ve discussed the potential benefits for the network as well. You just gave the example of cherry-picking the cartons for the most pressing products. But at some point, you realize if you prioritize correctly, even a low priority product becomes competitive against the remaining carton of something that you’ve already pushed to the store. Are there other elements that can be considered when we start looking at the problem from a network perspective?
Richard Lubienski: Yeah, even as a network, supply chain departments, in my experience, can still be quite siloed, even within a company. You have a team that does forecasting that uses tools to define the tasks that need doing, and another group, Logistics, that executes the tasks. One can have a set of KPIs about service level and forecast accuracy, whether you like those things or not. But the other one is heavily cost-focused. Trucks, warehouses, thousands of people, tens of thousands of people can be involved and will have a cost KPI. Their opportunity to cut costs and be efficient can only go so far. It can be limited if they do a perfect job.
This is even more so in the world where you use a third-party logistics company, where their job is to do what they’re told as efficiently and effectively as possible. So, your supply chain costs can get to this level. However, if you bring the tasks into play and can reshape them, for example, smooth the tasks across the network, the opportunity to reduce supply chain costs is huge.
I’ll expand on that a little bit. If you took the center of the store, the ambient grocery, a typical supermarket may have 20 to 30,000 SKUs across the whole supermarket. Seventy percent of those will sell less than one unit a day and replenish less than one carton per week on average. Looking at those statistics, I would estimate that in a forecasting system or a reorder point system, maybe 50% of the volume that you would ship, and we’re talking hundreds of thousands of cartons on a daily basis, doesn’t have to go on the day that it’s triggered. It could be moved earlier, which isn’t so much of a challenging concept, other than for capacity in stores. But it could be delayed a day or even a couple of days, with little or negligible impact, or an impact that is negligible in comparison to the supply chain cost to actually ship it on peak days of the week.
If you can do that, it opens up a whole business case of opportunities for supply chain costs. Also, nowadays, we’re seeing a lot of investment in automation, customer fulfillment centers, and supermarket replenishment. The expensive assets being invested in could have a longer life or be smaller to deliver the same demand. I think that’s an attractive proposition, but you have to bring the cost to serve together with the replenishment decision-making. If you bring those together, the whole supply chain industry has a whole new horizon to go after, and that’s very interesting.
Joannes Vermorel: That’s typically a case that you’ve outlined here, which is joint optimization. So you’re optimizing the quality of service versus inventory write-off cost in the store plus the transportation costs that are also part of the picture. You’re doing multiple optimizations and potentially all of that can be enriched further with pricing optimization, controlling exactly the amount of decay of discounts for products that are nearing the expiration date. My own experience is that, although these elements are fairly obvious when you think about it, the typical problem that companies face is that their own division of labor has already created silos that even prevent people from looking at this sort of optimization. The silos define strictly isolated areas where the rest do not exist. So people who do the transportation do not see what is happening exactly in the stores. Every single division has its own silo. Maybe with All Futures, how do you see this sort of conduct of change in these large companies? Because the primary bottleneck that I see is that the organization in place is typically very averse to even considering, in the first place, this sort of joint optimization where you take multiple problems that belong to multiple departments, bring them together, and reoptimize with computers, obviously, across the board to get something that is superior to anything that could be done in isolation for every segment.
Richard Lubienski: I think the concept is more challenging actually for the traditional forecasting and replenishment departments than it is for logistics. Logistics has everything to gain from smoother operation, safer operation when it’s smoother, fewer peaks, and constraints managed and respected as opposed to receiving an unconstrained and unprioritized pick list that actually you can’t squeeze through at Christmas or at least they have to push the boundaries, and that’s not good. The change requires inspired leadership, frankly, from the top down. Departments of any sort are a bit defensive of their own structures, their own people, and their own way of working. Retail is an intensely pressured environment as well, so to start contemplating a complete shake-up of what you do on a day-to-day basis, seven days a week, is a tough personal challenge. But it’s an opportunity that’s well worth going for. The transition states between what you do and your desired state are something that needs careful management and attention as a program manager. It’s quite often easy to see where we are today and the vision of tomorrow, but actually, how do we get between those two? Because we’re not just going to switch the whole company over and risk their entire sales. You will start off at departments and at a certain level and watch it carefully. The prize is there. The shareholder value is there, the safety aspect is there, the cost, the environment is there, the waste is there. All the pressures are in the right places, actually, other than the technology direction being positive, which is going in the right direction.
Joannes Vermorel: To conclude this interview, where do you see the market going for the next five to ten years? When I say the market, I mean large food retail networks with regard to their optimization in a landscape that is becoming even more complex with omni channels. They have their stores but they have more and more additional channels, plus all of that in an environment where there is a bit of inflation with an extra variable of uncertainty to make sure that it’s very fuzzy. So, where do you see the market going in this environment?
Richard Lubienski: I think the concept of linear supply chains, where I just go from one tier to the other in a straight line to a customer, is already outdated. You try and add a customer fulfillment center into a supermarket network, servicing online orders. Is the stock coming from the supplier, or is it coming from our distribution center where it’s already lost some life? All of that is not going to work with linear algorithms. We need a different lens on the decisions, incorporating economics.
In the past, I avoided using the word optimization, knowing that it wasn’t true. Improvement, yes; optimization, no. In replenishment and supply chain decision-making, if you don’t bring in economics, it’s impossible to optimize. What are you optimizing? What’s your objective function? If you can bring the subtleties of customer loyalty and so on into that economic equation, even better.
These concepts are very big and complex at a supermarket level, but at a smaller level, take a pharmacy that takes one tote of a product from its distribution center every day, and it costs thirty dollars to send it there. If it’s got one bottle of shampoo in it, that’s not a good decision to send. We should have waited or filled it up with the best other products, depending on the economics of each product and the chance of selling it. It makes total sense.
The industry is slow to change, slow to adopt, and I don’t think academia particularly helps, still not pushing the boundaries of the science. Unfortunately, that’s still a reality, so it’s going to take some inspired leaders to push that. It’ll happen over time, and as more mainstream software companies adopt some of that thinking, there will be more written, and it’ll get into people’s minds. How fast? I don’t know, but hopefully in my career.
Joannes Vermorel: Thank you very much, Richard. This is it for today. Thank you for staying with us, and see you next time.
