Demand Sensing and the Haves and the Have-Nots In Supply Chain Technologies

Even before the black swan of COVID, the supply chain was under stress from shifting consumer demands. As consumer goods companies determine how to adjust their strategies to these constantly connected expectations, the important role that demand sensing can play in separating the winners from the laggards has emerged. IDC’s Simon Ellis and EdgeVerve’s Suresh Bharadwaj joined a recent CGT webinar to dig into the growing role of demand sensing in successful supply chain operations. As part of this, Ellis shared exclusive info from IDC that plots an assessment of organizations’ digital maturity against revenue and profit performance to identify the gap between the companies that have embraced digital technologies and those that have not. Read on for the edited transcript and all of the presentation slides, including IDC’s findings. Lisa Johnston: Hello, everyone, and welcome to “Drive Omnichannel Growth by Unifying Demand Signals.” I'm Lisa Johnston, editor-in-chief at CGT. Today's consumer goods companies require a deep understanding of consumer needs in order to remain both future-ready and resilient. Given that many consumers today are digitally savvy and constantly connected, the most forward-thinking manufacturers and retailers know this means they have to meet consumers where they are at all times and understand just how they want to interact with the brand. What does that mean for the supply chain? It means a lot of things, but we're here to talk about the role of demand-sensing across direct, indirect, and online channels and how it's growing in importance. During this session, two industry experts will dig into some of the challenges that are occurring, how consumer goods companies are navigating them, and outlooks for what's ahead. Joining us are Simon Ellis, program VP of IDC manufacturing insights, and Suresh Bharadwaj, AVP and global product head of TradeEdge at EdgeVerve. Gentlemen, welcome, it's great to have you here. Can you each take a moment to introduce yourself? Simon Ellis: Hello everybody, Simon Ellis. I've been around manufacturing, consumer packaged goods and supply chain for a very long time. I've been with IDC now for 15 years leading the supply chain and manufacturing research. Prior to that, I was at Unilever for a little over two decades, across, frankly, just about every part of the manufacturing and supply chain organization. I look forward to the conversation. Suresh Bharadwaj: Suresh Bharadwaj, I'm a lead product manager with EdgeVerve. I've been with the company since its inception back in 2014, about eight years now, and I’ve been with the Infosys Group for about 15 years. Prior to this, I was an entrepreneur. I co-founded Manhattan Associates, where we built warehouse management systems. I've been a product guy my entire career, which spans a little more than three decades. I've been in the supply chain domain for most of the time. At EdgeVerve, we've been working with large CPG companies in the supply chain, helping them with all the challenges that they've had for years now and more recently, of course, in the last two to three years. Johnston: Simon, I'm going to ask you to kick things off. We don't want to start every event talking about the pandemic, but when it comes to the supply chain, you can't do that without talking about the pandemic. Can you talk about what's changed during the pandemic and where we find ourselves now? Ellis: I've managed to avoid actually getting COVID, but I have COVID fatigue because I'm tired of talking about it. It's myopic on some level because it hasn't just been that, there have been other things that have happened — COVID had nothing to do with the boat or ship getting stuck across the Suez Canal, nothing to do with what's happening in Russia and the Ukraine. It probably does have something to do with the inflation pressures we're dealing with and, at least, the specter of recession, but I think of it through the lens of disruption. Even prior to 2020, we were seeing a more disruptive supply chain environment. Here at IDC, as a market research company, we do surveys. We did a survey in early 2020, then another one a couple of months ago, it was a combination of both manufacturers and retailers, but the companies are telling us about this idea of lack of visibility, not being able to see changes in time to react effectively to them. A lack of digital tools, potentially, to transition to new business models. One of the things we've seen over the last couple of years is that the pandemic affected different industries to varying degrees. For example, if you were in the travel and transportation industry, you got hit really hard. If you were in the chocolate and candy business, you probably got hit hard. If you were in the pet food business, you probably had a banner year. We saw a lot of interesting dynamics. Although this is through the lens of COVID, a lot of the conversations that I read in the news and see in the press is supply, supply, supply. We've had some supply issues, but demand has been a real problem as well. In many of the companies that I talk to, they've had supply issues, transportation issues, but unpredictable demand, volatile demand, has been at the top of the list for many of them. Think about demand planning, broadly, as a pretty mature space. Most of the vendors who sell demand planning tools have the necessary stochastic, forecasting algorithms, all that sort of stuff. That's a mature space, but what is not so mature, and what in some ways is much more interesting, is this notion of demand sensing and how we can change how we think about forecasting and the ability to fulfill forecasts. Using granular levels of data, new data sets, putting them into a more dynamic demand-sensing capability is the way of the future. A lot of things happen just in time. Inventory proved to be not quite as good as we all thought it was going to be. Companies who relied on suppliers for, and used VMI and SMI technologies, maybe that didn't work so well. At the root of it is this volatility around demand. We have no precedent for a pandemic in the last 100 years; companies aren’t going back to 1918 to look at data. In a way, we've had to reset our view of demand, not to necessarily dust off the old demand-driven operations, demand-driven supply chain, but we have to think differently about demand and demand forecasting through the lens of sensing because that, ultimately, is the best way through. Maybe there's a better algorithm down the road, but that's not transformative; demand sensing is transformative. Johnston: We're going to get into demand sensing in a minute. Suresh, can you provide a little context or color to what Simon's saying, based on what you hear from your conversations with your customers? Bharadwaj: What I'm hearing is that this was a perfect storm that came together. As much as companies were trying to figure out how to move away from shipment-based forecasting to demand-based forecasting, things changed so rapidly. All of a sudden, the points of aggregation of demand, for example, which typically used to be a retail stores (which had fairly fixed locations so you could predict where to position inventory and how much to position in different geographies), suddenly was no longer valid because you and I could be sitting anywhere across the globe and be placing orders for much smaller sizes. It didn't mean that your overall revenues went up, but the demand signals got fragmented, significantly. It also was caused because new products were being introduced in the market. There was an explosion in the number of SKUs and there was a demand signal fragmentation happening as a result of that. Then, combining that with the size of these orders and the need to deliver to wherever your consumer may be, created a whole new set of challenges. This is what threw everyone off the cliff from a forecasting perspective. It's a perfect storm because never before had we seen an explosion in demand and a disruption on the supply side at the same time. You could have had a hurricane or some natural weather event that might have taken out a few factories or plants here and there, but nothing like this on a global scale. The question, especially when you look at folks in the fashion industry — apparel, etc., who have long supply chains and longer lead times — they placed orders for their products, at least 9-12 months, maybe even longer, ahead of their season. In many cases, those orders were still getting delivered, but then they had no place to go. For those who are completely out-of-stock and are making things overseas, it may not have been their tier one suppliers at the root cause of the problem — it could be a tier two or tier three. It was too late for them to know what was happening. Suddenly, getting to see no more about my supplier’s supplier has become the new ask for many customers on the supply side. How do I get into the tier two, tier three visibility? From a logistics perspective, real-time visibility is becoming the norm of today, getting to know where a product is. Of course, as consumers, you and I have tried to track shipments all the time, but now it's becoming more B2B with real-time visibility of tracking and shipments containers, or whatever have you. Across the board, demand, supply and logistics have seen significant impact. Ellis: To reinforce something important that Suresh said, the trend from 2015-2020 had been clearly e-commerce growth — online ordering, there was a clear growth trajectory. Companies had planned for that, but to the point around a perfect storm, almost overnight people were moving from working in offices to working from home, or shopping in-person to shopping online. If the growth had suddenly gone like that, we could be critical of companies who weren't prepared for that, but who is prepared for that? I've covered risk management in manufacturing and supply chain for a long time. The Black Swan events are hard to deal with, you don't expect them; this sort of “storm of the century'' comes every 100 years. However, when it comes, it takes you by surprise. Suresh, you're absolutely right, it was a perfect storm, a Black Swan event. We have to be careful about overreacting to these things, we don't have to throw everything out that we've always done. We don't want to keep 10 times as much inventory on-hand in the event there's another because you would never do that. My challenge, and the thing that I wrestle with every day, in terms of research is, how do we take a tempered approach to this? How do we do things that make sense, but not overreact, and not have knee-jerk reactions to things that don't make sense in the longer term? Johnston: It's very interesting because we talk a lot about finding the new normal, and then this is the new reality. Of course it's been a great acceleration, great change, but maybe not all of these changes are going to stick. I can see where the challenge is, trying to figure out which ones they are. Simon, you started talking about demand sensing, let’s dig into that a little more — the role that this is playing and what's different about it. What can be different moving forward? I know you have some information to share with us on managing risk. Ellis: This is again from the recent IDC supply chain survey. It's supply chain specific, not sort of factories and sort of operational stuff, although the observations probably still apply. Over many years, even predating 2020, I have talked a lot about this notion of supply chain resiliency. Yes, it has become the new overused term, but what does it mean? What are the things that we have to be able to do? To my point from before, how much of it is sensible, good business practice versus not overreacting. I think of it through the lens of two things: 1. Visibility We've been talking about supply chain visibility for a long time. Suresh, what you guys have been doing in business around networks, collaboration, and connectivity is one example of how visibility is far better today. You made the point about visibility into supplier's suppliers, which is poor for most manufacturers. Sometimes the problem is at a tier three supplier. I hear a lot about near-shoring and made in America, at least from a US-centric perspective. Let's be clear, it's not “Made in America,” it's assembled in America because a lot of that stuff is still coming from other places — it's not the supplier you have to worry about, though, it's the supplier's supplier's supplier — you can go on and on. That notion of visibility is almost disruption agnostic. It's something that companies should aspire to, regardless of the disruption of the day. 2. Agility This is a little trickier. There was a company I talked to in early 2020 with good visibility capabilities, but their supply chain is so rigid that they couldn't react to it. It’s like watching an unfolding disaster, but not being able to do anything about it. Agility is the flip side of that: how do I make sure I have the ability in the business to respond to what the visibility is telling me and respond to it in a reasonable way? I've debated over time whether to call it sensible agility or pragmatic agility because there are certain things you wouldn't do — you wouldn't double your factory network and reduce your OE to 50%, or increase your inventory by 10 times. How do we build-in these capabilities so that the business is more agile, but it's not a level of agility that'll drive you out of business because it's financially untenable? There’s this idea of demand sensing. A lot of demand sensing is about seeing the data and seeing what's happening. Then, the agility is whether or not there is the ability within the factory network or outsourcing network to flex as that demand visibility suggests. There are lots of ways to think about visibility and agility, but it's particularly pointed when we talk about demand and demand volatility. Johnston: Suresh, does this dovetail with what you're hearing from your customers? Bharadwaj: Absolutely. Until recently, it used to be that customers were looking to syndicated providers to bring that visibility. Even now, many of them still do. Typically, it used to take anywhere from 2-3 weeks before they would know how a promotion ran, or working with market shares they have in a certain brand, category, etc. All of those have started changing while we are thinking that it’s too long. Sometimes you don't run a promotion that long, so you can't do much about it unless you get that visibility while a promotion is still running. That moved the needle significantly with direct collaboration with the retailers, where more retailers saw the benefit of providing visibility to suppliers and sharing what was sold in a store yesterday, today. We are talking about 24 hours. It’s not real-time, but that's good enough because what would I do if it was real-time? There's always a difference between how fast things move in the digital world vs. the physical world — that's what this is all about. At some point it doesn't make sense for me to know every product that is getting checked out at every checkout counter, every minute, every second — it creates too much noise. At the same time, I cannot sit idly by with a quarterly plan and stick to it, when I know that the demand is coming from someplace other than where I predicted it. My ability to make short-term interventions and be able to take those actions and adapt to something happening. The thing that Simon was discussing was about forecasting is that it's always going to be wrong. The question is: What do I do about it? The best you can do is be prepared to react. It comes back to making short-term decisions. You'll never get right about where you're going to position or deploy inventory, but your ability to move it around or move where the demand is coming from, those require action. It goes beyond insights. This is taking us into the domain of execution. Planning is one thing, but how quickly you can react and execute to what you're seeing in the market today becomes even more important. That's what we are seeing. Johnston: Could we talk a bit about how these strategies and the priorities can vary depending on the different channels you're selling through? Bharadwaj: One of the things that we have seen, even for companies who are collaborating very well now with retail partners, is that a lot of the collaboration happens in silos. It depends on the kind of relationships, the key account managers with counterparts on the retail side. There are some which are very good, some are not so good, but what we have seen is that such data is not only created in silos, but also taking away key time from their day job. They're chasing partners for data as opposed to making plans or selling — this is very critical — this is one aspect of it, silos of data. The second is latency. Basically, sales have their own information and data sources and come up with a plan based on what they're able to get. The same thing is true for the supply chain and marketing. They all have their own sources of data and silos of data, which is why we need to bring them all together in a process to agree on something. That's the reconciliation process that typically happens, creating a single version of truth becomes really important. As a strategy, it's important for companies to look at creating this as a central repository or central source of truth for everyone to read off the same page. That means you need individual functions to give up what they're doing and trust the central force that's going to drive this collaboration and then disseminate the same information to everyone, so that they all can say we can make plans out of the same data foundation, if you will. I've talked about near real-time latency, again, if you're getting data at different intervals, different granularities, how do you bring them all to the same common foundation in terms of the frequency of getting that granular visibility, both on a geographical scale, whether it's at a store level, skill level, or even on a time scale. If you really look at it, 80% of the data that you need to do this planning is outside your organization. There is no getting away from that, this is something that successful companies have managed better in terms of creating a single version of truth for the entire organization to consume, and then offering this as data as a service internally within their enterprise. That's something that we have seen yield a lot of success. Johnston: What's separating the winners from the laggards? Simon, a lot of what we're talking about as you pointed out, aren't new challenges, this is something that's been going on for a long time. This idea of trying to achieve a single source of truth isn't necessarily something new companies are trying to do, but how can demand sensing help companies? What's the important role that it's playing that helps companies do something different right now? Ellis: As Suresh mentioned, a majority of the data that a company or consumer goods manufacturer needs is outside of its four walls. As outsourcing grows, what's within the four walls shrinks as well. I published a piece seven or eight years ago and made the argument that companies had this forecasting finish. They were too focused on the forecast, the only thing true about any forecast is that it's going to be wrong and hopefully it's not completely wrong. There, immediately, you start having variances against the plan. I've argued for a while now that speed is the underappreciated capability — if you can do something faster, you can outperform. I mentioned the Ever Given before, stuck in the Suez Canal. After that happened, I talked to two companies, one of whom knew immediately what they had on the ship, what was backed up on the water, and what production they were expecting in the next two months that might be affected. They were able to make decisions very quickly. The latency that Suresh talked about was very low. Another company, two weeks in, still didn't really know. They were having to make gut-feel decisions. Everyone knows which company they’d rather be. Demand sensing at its decomposed level is data, timely and real-time where necessary and rapid assessment of what that means to the plan, as well as what that means to what I have to ship when and where. I don't want to be dismissive of traditional demand forecasting and the clever algorithms that companies have used to great success, but that's evolution-ish. We’re talking revolution-ish, or how do I use the data better? How do I collaborate on that data within and without my organization? Certainly there are many supply chains, many operational organizations within CPG companies that are well on the way on that journey, but there are others who are not. Suresh, you were talking about collaboration before. One of the questions we ask in the supply chain survey is around collaboration, the number three priority is “collaboration within my own business.” It's not collaborating upstream with suppliers and downstream with customers, it’s companies saying they still haven't worked through collaboration within their own business. As with any maturity curve, the majority of the companies are in the middle of the bell and they've got work to do. This idea of demand sensing is interesting because it focuses on all of the things that we're talking about: data, data accuracy, data timeliness, latency, connecting visibility to agility. While it's not the only thing that drives a resilient supply chain, it's one of the ways, maybe the best way, to try to get through some of this demand volatility that many companies are wrestling with. Johnston: You mentioned the two companies: the one that was able to tell where inventory was on the ship and the other that two weeks later, didn't know. There's obviously a gap between the haves and have-nots. Are you seeing this gap widening between the companies that are more digitally mature and those that are not, or is it narrowing? Ellis: This is something I've been thinking about for a long time. One of the things we did at IDC was to plot our assessment of digital maturity against revenue and profit performance. Now, this is all organizations, it's not just CPG. We make a judgment based on knowing the manufacturer or looking at publicly available information like annual reports, announcements, wins and losses, etc. to make a judgment about what constitutes digitally mature vs. digitally immature. Think digital vs. non-digital, we plot that against publicly available financial information and we see an increasing gap between companies who have embraced digital technologies and the performance in terms of revenues and profits, vs. those that have not. Now, I use the term correlation intentionally because causation is hard to prove. Do companies that make more money, have more resources to spend on digital technologies or do digital technologies drive...? It's a correlation exercise more than it is a causation exercise, but the increasing disparity is quite clear. It's always about how technology enables the business to improve? How does it allow the business to seize on opportunities that perhaps, with the lesser technology, they could not? How does it enable the business to address challenges in ways that if they didn't have the technology, they would be less effective at addressing those challenges? It's always translating “digital-ness” into business performance. As the world gets more disruptive, the spread between the haves and the have-nots gets bigger. When we get around to doing 2021 and 2022, because the world is increasingly disruptive, you'll see a continued departure of digital versus non-digital. It's interesting that this is a way to quantify, from a correlation perspective, that these investments are actually working. Johnston: Suresh, I want to get your reaction to what Simon's been sharing about this digital maturity with performance? Bharadwaj: Going back to how quickly you can react, they travel at different speeds. Digital means you get to know what's happening immediately, but then, what do you do with that? I can get all that data, but if I'm not able to do much with it, the data becomes stale pretty soon because there's new data that's coming in. It's not only about getting it, but what you do about it — that's where the difference comes in. Again, a lot of times it's not about sales, inventories and fulfillments, there may be other factors that are not directly related to this. In the early days of COVID, for example, correlating the number of COVID cases against where products were flying off the shelf, is something that companies that were digitally savvy were able to do. They were able to realize the out-of-stock was limited to just a few days vs. some who were out-of-stock for weeks on end. There are some events that we can't predict — in terms of weather events and so on — but the things where you could identify those events, that are tracking it, can start making use of the data and see where it’s going to go out-of-stock next. It was a brilliant move for this company. Ellis: That's an excellent point. Late last year, a consumer goods manufacturer was watching to see where infections were starting to grow in a region, then they would peak, and then the infections would start to go somewhere else, and they would peak. The company took the learnings from the prior New York City/New Jersey/Connecticut peak, including what consumer behaviors changed and what products were higher in demand. Aside from hand sanitizers and masks, people were working, living from home, being isolated, how did the demand changes affect production and inventory? It's not precise from a factory-run strategy perspective, but it's generally instructive in terms of where to position and how to position inventory for which products. Johnston: Simon, could you explain the definition of a digital organization relative to your charts? Ellis: We looked at any and all publicly available information, and it ends up being a judgment. Does a company announce that they're going to be deploying a new supply chain planning suite, that they’re implementing artificial intelligence and machine learning capabilities, or they're participating in a supply chain network. We look into publicly available annual reports to see what companies are saying about technology and business processes. Then, we have some internal knowledge based on conversations with many different factors. It's not an enormously precise process, but it's about collecting any and all information we can to make a judgment about what companies are a bit more digitally mature and which ones are less digitally mature. We revise that approach each year as more and different information is available to us. It's a huge undertaking because there were over 1,000 companies the last time I looked. We've got folks that are looking through and you'd be surprised how much publicly available information there is. Ultimately, it's our judgment about what constitutes digitally mature vs. digitally immature. Johnston: Suresh, can you talk a bit about sustainability and the role of visibility and sustainability, what you hear from your customers. We know that it's a growing priority, or at least, companies are saying it's a growing priority for them. I'm curious to get your take on how you're seeing them or where they need to leverage this increased visibility to advance your sustainability goals? Bharadwaj: Quickly reacting to what Simon said, I'm not sure that this has gone into your benchmark for a digital organization. It's not only about the organization itself, but also its value chain. How much investment these companies are making in value chain partners in digitizing the value chain because ultimately, your success is only going to be as good as the weakest link in the value chain. That is something that's outside the organization, but overall contributes to how digitally mature you are, as well. Coming back to visibility and sustainability, while we have talked endlessly about the supplier's supplier, then getting to know the status of my order, where it is, where the container or truck is, etc. is equally important. We hear this more and more in the boardroom. Therefore, it's not only about timely deliveries, but also the practices. This is a different kind of data collection — this is not necessarily sales and inventory, but as an organization, there are certain things that are within my control. How many resources I use within my organization, whether it's water power, the packaging material I use, the waste I create — all of that is within my organization. When you look at the impact on the climate and overall the ESG metrics that people in the companies want to start tracking, a lot of that also has to do with your partner ecosystem. What's typically called scope three emission data, that's really not within your control. You want to go and look at, where are your partners sourcing their raw materials from? Again, are they using child labor? Maybe that's one example, but how much waste are they creating? How much power are they utilizing? A lot of that lies outside the organization. Once you build a framework or platform to start helping with your primary supply chain use cases, there's no reason you can't use the same platform and extend it to other use cases, like sustainability. That is the connection that I can make. Ellis: The “but” is you can choose who you do business with. Stuff that happens outside your four walls is largely out of your control, but you can control who you do business with. What's interesting is whether sustainability has climbed on that priority list. It's sustainability now, the second most important thing in terms of assessing a supplier. Is it the third? Is it the fourth? Cost is the first. I call myself a sustainability skeptic, not because I don't think it's important, I use this term a lot from posters to practice. It's fine to put stuff on glossy paper and posters and nice looking websites, but are we doing it? Are we operationalizing it? We did a supply chain resiliency MaturityScape a few months ago and did a benchmark as a consequence of that. One of the criteria was sustainability and the least mature is (I'm paraphrasing), “I have no idea.” The most mature is, "I'm operationalizing it. I'm building this into the way that I run the business." Ten years ago you saw a lot of chief lean officers in businesses, you don't see them anymore. The reason you don't see them anymore is that lean has become part of the way we run the business. We could argue we've over leaned and that's a different conversation — we see that with sustainability, the ESG programs and the reporting. How do we start to build this into the way we run the supply chain? Suresh, if you are in the Western part of the United States, which has been in a drought for apparently 1,200 years, water consumption is a huge problem. If you have a factory that's water intensive, it isn't sustainable. My only tempering observation is you can control who you do business with, but we have to start thinking about how to operationalize this stuff. It's not enough anymore just to report carbon and offsets and that sort of stuff and corporate social responsibility. We have to start truly operationalizing these things. Johnston: Sounds like we should do another separate webinar on that because there's lots of talk about, a lot of action needed. We're almost out of time. I want to ask each of you to end with some best practices. If you had to pick one or two, your best pieces of advice to companies who have a lot of heavy lifting to do, or who are just getting started on this and need to make a lot of progress, what would it be? Ellis: I used to get asked back in the days when we were going to conferences, and maybe that will happen again this fall, but I used to get asked all the time about what I thought would differentiate the best supply chains, the best operational companies from less capable competitors. It's about data. With all due respect to companies that sell warehouse management systems, it may not be the WMS, it's who's better at leveraging data. If you've got data and you've got the right level of analytics, you can connect that to immediate, midterm, and long-term actions, you will outperform companies that can't do that. You can build all the agility in the world, but if you don't understand what the data's telling you, if you don't understand where and how you should react, all that agility is probably academic. It's got to be about the data. If you are using the data better, understanding what data's important and what’s not, and you have the analytics and machine learning capabilities — to me, the clear differentiator for the future and companies that can do that better will outperform those that cannot out. Johnston: It’s almost getting back to basics, even though we're talking about advanced technologies. Suresh, how about you? What final words can you leave us with? Bharadwaj: As Simon mentioned, data is the key ingredient. You could drown yourself in a lot of data. Sometimes, you need to think about what you're wishing for and therefore trying to do all this yourself is not the best thing for the organization because that's not your core in a business. A lot of times there is so much data and then to sift through all of the noise and bring it to a level of what's consumable and what's not, is very important. Otherwise, you could spend hours and hours and days going through all of the data and not get out of it. For those who are specializing in doing all of the data engineering and then bubbling up the ones that you should be looking at. Then, starting to utilize that data is very important. I mentioned this within the context of an enterprise where individual functions, when they start trying to handle all the data themselves they create data silos and then they're all working off of a different data set, which may not be the same as what someone else has. As much as you want to drive the discipline and drive a single version of truth, single source of truth within the organization, try not to do it all yourself because you could be overwhelmed with the amount of data that you're going to receive. Johnston: Great thoughts to leave us with. Thank you. Simon, Suresh, thank you both for joining this webinar. We really appreciate you sharing your insight with the CGT audience. I'd like to thank EdgeVerve for sponsoring today's webinar, and I'd like to thank the attendees for giving us your time today. We hope you found it worthwhile.

Related Topics