Demand Forecasting Investments for Retailers Pushing Through the Pandemic

The urgency of accurately forecasting consumer demand has been shoved into the spotlight by COVID-19, and both retailers and CPGs are powering through the pandemic with a renewed focus on intelligent analytics. Implementing a short-term forecast is fundamental in understanding and predicting changing consumer demand patterns, with real-time, data-driven supply chains critical to success. In “Short Term Demand Sensing: New Techniques to Predict Changing Demand Patterns,” a webinar presented by RIS News and CGT and sponsored by SAS, we take a deep dive into the need for speed in forecasting, what COVID-19 has taught us about the value of statistical forecasting, how machine learning is becoming table stakes for success — and how corporate culture can make or break even the best tech stack. The full transcript and presentation slides from the conversation are below. Albert Guffanti: Good morning, everyone. I'm Albert Guffanti, group publisher of CGT and RIS. I'm thrilled that you've joined our highly anticipated webinar, “Short Term Demand Sensing: New Techniques to Predict Changing Demand Patterns.” Today's webinar will dive deep into the topic of supply chain and real-time data. We'll examine the topic through the lens of the COVID-19 crisis and we'll ask questions like: “What has COVID-19 taught us about real-time data in the supply? “Why is speed so important for demand forecasting?” and “How does corporate culture affect the ability to achieve a real time business intelligence supply chain platform?” I also want to mention that while the materials in this webinar have been reviewed by the editorial staff, the views of the speakers and their organizations are their own and do not necessarily reflect the opinions of CGT, RIS or its owner EnsembleIQ. I'm thrilled and excited to be joined today by Charlie Chase, Executive Industry Consultant at SAS, and Orkun Ozturk, head of IBP Demand and Supply Chain Planning Center of Excellence at Mondelez International. Gentleman, thank you so much for joining the webinar today. Before we dive into the conversation, Charlie and Orkun will set the stage with some of their own insights. And then we'll dive into some more conversational Q&A type of formatting. But, until we get there, Charlie, why don't you kick things off and share some of your insights for us? Charlie Chase: Great. Thank you. As you know, over the past 20 years the approach to demand planning at most retailers and CPG companies has remained stagnant despite the increasing availability of data, advanced analytics and technology. Today, many companies continue to rely on the traditional practice of using historical shipping data and its statistical plans theories data deployed by their ERP systems along with a whole lot of Excel spreadsheets developed demand forecasting and demand plans. It is labor intensive, consensus driven planning process requiring hundreds of manual overrides before it reaches supply planning for final execution by manufacturing and distribution. This traditional approach to demand forecasting has been able to capture many of today's complex market dynamics and it's plagued with biased human error, waste and inefficiency. As a result, the demand planning process has been unable to capture marketing dynamics including price changes, promotions, as well as changing demand patters and external factors that are being caused by COVID-19. In fact, legacy ERP systems that create historical forecasts really weren't designed for this new digital economy. It could only model patterns associated with trends in seasonality. They don't account for the effects of sales promotions, other related factors. In fact, most companies are forecasting shipments, which is a supply signal. Some are forecasting sales orders, which is a replenishment signal. And we all know that the true demand signal is POS and syndicated scanner data for both retailers and CPG companies. Furthermore, CPG companies have been focused primarily on the symptom rather than the root cause. Using buffer stock to offset poor forecasting with a focus on supply planning. In fact, most demand planners are actually planning constrained supply plan, not a demand plan at all. They're too far removed from the consumer, reporting upstream to supply chain management. As a result, they can't take advantage of downstream data and information regarding consumer demand to detect changes in consumer purchasing patterns. This requires working closer with sales and marketing, integrate marketing program and spending which influences demand. As a result, retailers and CPG companies supply chains lack demand visibility from customer to the consumer. This leads to manual intervention and increases buffer stock to respond to fluctuation in demand and supply. There is some good news though and recent developments are underway using intelligent automation supported by machine learning as well as real time digital consumer intelligence and downstream data to predict change in consumer demand patterns more accurately. These new approaches will be the new normal for retailers and CPG companies to successfully emerge from this current pandemic. Over the past six years, I've been following CGT’s annual Analytics Studies. These independent third-party studies focus on the feedback from executives like CEOs, senior vice presidents, VPs and directors. We've found this information validates what we're seeing with our customers in retail and CPG. Here's a snapshot of the CGT 2020 Analytics Study. The interesting findings are that retailers and consumer goods suppliers for the first time, in the last six analytics studies that I've been reading, are in agreement on the top three areas of focus over the next year as a result of the pandemic. And those are demand forecasting, consumer insights and inventory planning. Although analytics maturity levels have increased, they're still lagging behind in demand forecasting and inventory management when it comes to investigative and predictive analytics. Another key finding, both retailers and CPG companies are implementing AI machine learning in areas such as demand forecasting and planning, merchandise planning, supply chain planning and others. And we're getting a lot more requests from our customers asking us how they can leverage machine learning to better predict short-term changes in consumer demand patterns as well as providing intelligent analytics driven by machine learning to reduce manual repetitive work while providing real time consumer intelligence. As retailers and CPG companies continue to pour energy and focus on the ongoing COVID-19, they are emerging from the pandemic with a renewed focus on analytics, at least that's what we're seeing. And based on this analytics survey, there are key emerging findings, six in all. We already talked about the first one where they both plan to focus on demand forecasting, consumer insights, inventory planning. However, having a clearly articulated analytics strategy and putting than right staff together, whether it's an analytics department or a center of excellence, is really important, particularly for change management in acceptance purposes. In fact, we are seeing more and more companies investing in data scientists, as well as creating new analytics departments headed by a chief analytics officer so it's not uncommon to now see CAOs. However, still today only 36% of retailers are sharing their POS transactional data with CPG companies and only 25% are sharing promotional data. And only 40% of CPG companies and 13% of retailers are currently allocating more than 10% of their IT budgets to analytics. The good news is that CPG companies currently spend more on analytics than retailers. However, that's not good news for retailers, only for the CPG companies. And over the next two months, retailers and CPG companies plan to upgrade their current digitalization dashboarding capabilities. In addition, retailers are looking to enhance data security and in store analytics while CPG companies intend to focus on enterprise data and reporting tools, data warehousing storage. These overarching capabilities are not surprising given the current situation related to COVID-19. So, how do we get back on track with all these results from this wake-up call from COVID-19? So, there are some things we can take immediate action on. First, use point of sales syndicated scanner data to build up epidemiological data, social media, local economic and mobility factors to predict changing demand patterns, particularly in a short term. Investing in COVID-19 tracking and digitalization and dashboarding capability, like you see on the right, to get real time, day by day information or hourly information for that matter. Implement short-term demand sensing using advanced analytics and machine learning to create not only a one to eight week weekly forecasts, but now we have the ability to create one to eight week daily forecasts. Consider deploying something we refer to as assisted demand plan, which provides demand planners with a digital assistant using machine learning to improve their forecast. Machine learning actually to the rest of the demand planner, in what direction, by how much and what products to actually adjust based on the statistical demand forecast. This reduces their time significantly and reduces the number of manual adjustments. And then consider embedding your demand planning team downstream in sales and marketing organization closer to the customer and the consumer, providing real time consumer intelligence that they can integrate into their demand forecasting process. I'm not saying have them report to sales and marketing. I'm saying embed them down there so that they're closer to the consumer and the customer and they can work closer with sales and marketing. That's what I did when I was a director of demand planning and forecast. And finally, there's some long terms actions you can take. Invest in a data lake, demand signal repository or a cloud to standardize and normalize transactional data, point of sales syndicated scanner, sales promotions, epidemiological data, economic information, as well as other data sources. Also, hire data scientists who can apply more advanced analytics and machine learning to the process and also invest in cloud ready, I should say, or cloud native rapid demand response intelligent demand planning technology. If you want more information on this type of implementation, there is a link at the top of the COVID-19 dashboard that you can go to for more information. Now, I'm going to turn this back over to Albert. Guffanti: Thanks very much, Charlie. I'm glad that you left with some things to think about in terms of immediate and longer term actions. I actually wonder how many of our audience members today are implementing some of these suggestions. Now, I'd like to turn it over to Orkun who, working at Mondelez, has been on the front lines of the pandemic and the crisis and has some really valuable real-world experience to share. So Orkun, looking forward to your insights as well. Orkun Ozturk: Thank you. So, what are some of our learnings as went through these unprecedented times? Well, one thing for sure, priorities and patterns changed in a sizable manner and one might think also pretty quickly. Translating this into the snacks industry where we are, we have experienced some really key initial points. The first one was directly linked to priorities really. People started buying to make sure they don't run out of food, rather than indulging themselves. So, they buy biscuits and bakery products before a chocolate bar maybe. Then, there was the very practical angle of how long you can store the products you buy as no one really knew and doesn't know even now how long this situation would last, at what rate you would consume what you bought. We tended towards long shelf life items. So the ones that we can store with ease let's say. But again, that was not all of it either. A product of ours, I'm sure you know, that has relatively shorter shelf life, the very iconic Philadelphia Cream Cheese became even more popular than normal, if you can believe it, in this period. And this was, we believe, linked to the changes in our lifestyle. Now, most of us who are here that are joining from home I'm guessing so most of us are working from home. We are cooking and eating more at home. Maybe tending towards lighter and healthier options, like sandwiches, literally meant more for the Philadelphia consumption. And one might throw in the occasional cheesecake in there as well that you need that good cream cheese for that. But, put this altogether, the changes of how life is for us made a big difference in an industry like ours, I'm sure with many industries as well. Now, tying all of this back to demand planning. The context we are in really pointed towards some critical forecast accuracy and bias watchouts. Then what this meant for how we operate within the supply chain was another topic or point of vary really for us people in supply chain. Clearly, it became a matter of picking up the demand signal as good as possible and passing it on to the manufacturing network as quickly as possible so thereby aiming for the maximum flexibility in manufacturing because you have to then start some lines, stop some others, balancing workforce, balancing materials. So, trying to be as agile as possible, in essence. Now going into the crisis, our thinking was, okay, this is where all the bets are off. There's no good historical data, not a clear path forward. Therefore, demand analytics might not get us what we need for this ability. This was our thinking. Rather, it's our people on the ground so our trade activation people, account managers, local teams that will be our eyes and ears. So, market intelligence will trump analytics, this is what we thought going in. But in reality, things were different I must say. After the three months we have gone through, and much was half and half, let's say about April, May and June, we came to see that statistical forecasting provided at least as much benefit as local market intelligence, if not more, for the markets that we use it in obviously. Now, I will not be able to claim that we for sure know the details of this very complex context. But, analytics models provided really good results or forecasts let's say that compared to actuals, and we're now trying to understand a bit more about this and from two angles really. One is what kind of quick changes in modeling, so tying back to what just Charlie mentioned before me, what kind of quick changes in modeling can help us in sudden events like this one. Today, it could be the pandemic. Hope that does not repeat again. But tomorrow, it could be another event that comes up all of a sudden. It could be a competitor losing manufacturing capability. It could be anything. So, we need to first understand then we need to build a plan for situations like this. Not only situations like this, but from a modeling perspective, be ready for it. And the second angle is even in the middle of such a variable environment, what are some of the elements that remain constant or that remains constant because we have come to notice that there are quite a few platforms that the pandemic didn't really make any difference in how they were bought, which was very interesting. So, you can look at this in many different ways, but the ones with a high degree of forecastability, even in the face of pandemics, is something that's really important for us to research further. Guffanti: Thank you very much, Orkun. Those insights and learnings were very valuable and they'll continue to be valuable, as you mentioned, as pandemics and other situations arise in the future. I'd like to transition this to more of a conversation-style discussion and I have some questions that I'd like to ask both of you. And I'll start with Orkun to build on what you've just talked about. I'd love to know specifically what has COVID-19 taught us or you about the use of real-time data in the supply chain? Ozturk: So, along the lines of what I just said, everything starts with a healthy demand signal in supply chain, everyone knows that. So, I just mentioned that when we went into COVID, we were expecting the impact we would get from real-time data from our teams actually. So, we were thinking, it's not going to be the real-time data, it's going to be our teams on the ground. But clearly, if we can have real-time data and we can see real-time data into our statistical models, that is in a way the best of both worlds. You leave human bias, target setting, et cetera, type impurities out of the forecasting ecosystem and you get the best information without bias, in forecast terms and in non-forecast terms. And you feed it into this machine that will give you, this model that will give you your best response option. And that will be a great input for the remaining parts of the supply chain. That's why it's very, very important to have real-time data and have it quickly. Guffanti: Thank you very much for that. Charlie, do you want to add your perspective to that question? What has COVID-19 taught us about the use of real-time data in the supply chain? Chase: Real-time data's probably the most important thing that we've learned as far as getting that information very quickly in what we refer to as real time. And the fact that we are now in a digital economy, a lot of information is being captured on a regular basis by the minute based on devices and social media, devices that are embedded on IOT and social media. So someone asked me at the beginning of the COVID-19 pandemic, "Where should I go?" And I said, "The first thing you should be looking at is your data that you're capturing from POS, from your key accounts like Walmart, Kmart, other large grocery chains because that you're capturing on a daily basis.” So, you'll start to see almost immediately those changes in consumer demand patters and then follow that up and supplement it with syndicated scanner data." Now, a lot of people will say, "Well, the syndicated scanner data has a lot of latency." But, it doesn't any longer. Most companies are getting it within a week or two at the worst and some are getting it even quicker depending on what information you're buying from those syndicated scanner companies. Those are your two real demand signals and then take those two demand signals and look at your customer orders that are coming in and compare those and determine how those patterns are changing. You'll see almost immediately, within several days or a week, versus waiting for the next demand forecasting cycle. So, I think we're finding that digitalization is going to be key to capturing information and data. And also if you're selling a lot of products on Amazon, you would see that shift because Amazon is allowing you to access that information. For your products, you'll see the shift from brick and mortar to mobile direct to consumer sales through that one channel because basically that's really one of the key mobile channels. And most CPG companies now, somewhere between 2% and 10% of their business going through Amazon and as much as 30% to 40% of business through Walmart. So even if you only had those two particular digital sources of information and you were capturing that through your interfaces on a daily or even a weekly basis, you'd be able to see those demand patterns. And we can connect that information to shipment models in a two stage process that allows you to use that information, use the point of sale and syndicated information as a leading indicator in the models. So yes, we're learning a lot about digitalization and how that can impact the supply chain in real time. I hope that answers your question. Guffanti: Absolutely. Thank you. And then I'm going to ask a related question to that and it has to do with speed. We hear the word speed come up so often during the crisis. Orkun, can you articulate for us, from your perspective, why is speed so important when it comes to demand forecasting? Ozturk: Sure. And I will quote a former manager of mine on this one. He used to say, "Manufacturing is like a big ship. It accelerates slowly, but when it has speed it's also taking a lot of time to stop it." So to answer the question, it's because the earlier we feed the best data we have into the manufacturing network, the better response we get from it. So, having to balance labor, materials, etc., is not easy to start with. And last thing you would want is to get it late or get it wrong. So, specifically the speed here is speaking to the late part. You'd like to make sure that you have the best of the information. Even if there's a need for a change later on down the line, then still you have the advantage of having more time on your hands and that luxury will get a lot of businesses breathing room to do things differently if they wanted to or if there's a change in the environment basically coming up. So, this is why it's really, really important because the manufacturing network where you really get what you're going to be selling is not nearly as flexible as the rest of the chain. Guffanti: Thank you very much and I love the analogy of the ship. And I think it's right on, and I think it has a lot truth for our entire industry where many of our organizations, many brands have decades upon decades of processes and procedures in place. So, this kind of brings me to my next question, and Charlie, I'd love for you to answer it, and that is what needs to happen culturally for supply chain data to go from historical based cost savings tool to a purpose-driven, real time business intelligence platform? I think, like I said, a lot of our organizations are trying to break down the silos, are trying to all pull in the same direction, but really what needs to happen here? Chase: It's all about collaboration and using data and analytics to drive the supply chain versus bias and politics. It's really today most of the supply chain, particularly in demand forecasting and planning, it's really about target setting and backing into capacity constraints versus really looking at what the real information data is. In fact, they're leaving a lot of money on the table in many cases. They're only able to fill those orders that meets their capacity. If you have more orders than capacity, we try to cherry pick that information. But, it's really about collaboration, not just inside your organization. This is another reason why I keep talking about moving demand planners down into sales and marketing because they're responsible for demand generation. They're closest to the customer and the consumer and all the promotional activity and the programming information is managed by them. So, you would have direct access to that information. But, it's also about collaboration with your customers, the retailers. In fact, there was a recent study that just came out in June 2020 by Coresight Research, and they're finding that retailers and CPG companies who are collaborating with one another, they're just seeing significant improvements, not only in customer service, but also in revenue and meeting the demand. However, there's a lot of cultural things that are restricting them from that collaboration. First of all, the biggest challenge within a company itself is its own corporate culture. Why are we changing? Why are we moving to digital technology? Why are we applying machine learning? Why are we moving all our technology to cloud? These are the kind of things that disrupt the organization in a way that blocks them from moving forward and that's the same with retail. They found in this research study that the biggest challenge are cultural challenge the CPG companies and retailers are having because they don't trust one another. In fact, research of retailers said when they do work together, they've seen through improved collaboration, that actually leads to almost a 6% to 10% improvement in business planning as well as revenue and customer service improvement. So, I think that our biggest challenge is our own internal corporate culture as well as our corporate culture that we have in relation to our relationships with the retailers. I'm talking from CPG perspective because that's where I spent most of my career prior to coming to SAS. So, there is a lot of unnecessary stress on the system due to the politics and internal things. And also, I think another key area is sharing not only business strategies, but also sharing metrics. So, one of the reasons why S&OP fails is because first of all sales and marketing very rarely shows up to the meetings so it's very operations planning focused and it's focused only on those things that are related to supply, reducing costs, be more efficient. There's no talk about revenue generation and things like that. And they're using two different data sets. One's using POS and syndicated scanner, the other one's using shipments and sales orders and they're not sharing metrics, common performance metrics, horizontally across the organization. So, that's a key. I think that sharing metrics within your organization and then sharing metrics with your key accounts and customers and collaborating not only across your own company internally, but collaborating with those customers are very important. And that's the only way we're going to break through these cultural things that have been in place for the last 30 years that I've been involved in consumer packaged goods. I'm not sure if that answers the question. Guffanti: It certainly does. Thank you very much. We actually have been getting a real good volume of really great questions from our audience. I'm actually going to transition right now, in the interest of time, to some of these questions that have been coming in because they're actually better questions than I've asked. So, Orkun, I'm going to ask you, there was a question that came in from a very large CPG manufacturer and the question is, in the U.S. at least, COVID-19 spikes occur locally, first in the Northeast and now in the Southeast and South. How does your supply chain respond to this complexity? Can you answer that for us? Ozturk: Sure. From what I hear, it could be even more complex in Europe because what has happened is we had variability, not only geographically, but because certain places were hit more. Southern Europe, for example, they were impacted more with COVID than northern Europe and central Europe was one of the last ones actually. So, we did have this geographical variability, but on top of it, the behavior of consumers on what they were buying was a little bit different. So, there was an overall let's say inclination, but it wasn't that easy, it wasn't that, I have to say, predictable or easy to predict. It wasn't any different than what I just said a couple minutes ago. We tried to understand the best possible and then if we have any buffer, the smallest amount of buffer we have, we wanted to use it on the manufacturing flexibility piece. We would like to keep the timing, keep the time buffer let's say, just to make sure that we produce the right product. Get it through to the manufacturing network as soon as possible and then make sure that we prioritize the right products and we knew that we would be de-prioritizing and losing revenue on some others. But in situations like this, if you fail to prioritize, you definitely cannot win. So, this is how we have done it, we have created a plan to understand what's best to respond and we made sure that our manufacturing network was at their highest level of flexibility to provide that goods. Guffanti: Charlie, do you have any comments on that? Chase: You could actually capture local regional information, economic information that's available now. Most of it's free from the Federal Reserve, St. Louis Reserve Bank. We used a lot of that, they had actually global information, as well as taking information from a COVID-19 dashboard, like the one we showed today, that is capturing epidemiological data by region within countries. So, you could look at the southeast of the U.S. versus the Northeast. And we're finding that data actually could be modeled along with your internal shipment data and along with the POS data, the economic data as well as the epidemiological data and economic data. In fact, we worked with a European retail grocery chain that's a direct to consumer and when we ran the models, just to test them, we started out just looking at their demand history and forecasting it by region within that country and the error's maybe around 40%. Then, we added Google trends data for just that region and it improved it. And then we added epidemiological data for just that region and then economic data and we ended up improving the accuracy up to over 90% by just adding those other factors for a short-term one- to eight-week forecast. It's all about understanding the data and actually capturing that data within the local region and that data is available in many cases. Guffanti: I'm going to shift a little bit. We have actually gotten a number of questions that relate to the point of sale. So Charlie, I'm going to ask you another, we've been asked a few times from the audience around this. How does one collect real-time data from the point of sale or from other promotions? Chase: Well, usually most companies, usually their top 20 key accounts represent, at least when I was working in CPG, represented about 80% of their annual volume. So, it's basically 20 different customers and usually those customers are pretty willing to give you that POS data, just for your products. And they'll also give you pricing information for your products. And the first company to actually do that, that we worked with back in the '90s ... Now my age is starting to show. I was actually a protégé. I was only about 15 or 16 at the time. Walmart came to us and said, "We no longer are going to forecast your data." And then we said, "Okay." And we had 36 SKUs in Walmart at that time, it was small. They probably have a lot more now. We said, "Well, send us the data and we'll forecast and we'll you forecasting." The interesting part is they sent us our shipping data. And we said, "No, we already have this data. I want access to the point of sale data." And we had retail links and he said, "Sure, just download it through retail link and you'll get the pricing information." And what we did is we applied our promotion programming to that and our everyday low price to that and other factors and we were able to predict our demand within those Walmart stores by region. And we just sent them a forecast and told them just take that forecast, run it against your distribution system and then put an order and send it to us. So, that information was available. We were able to download data from Target, from other large retailers that we were working with. So, I think if you reach out to those retailers, and I'm sure your marketing team probably already does because this is another reason why demand planners need to be in sales marketing. The marketing team has the trade marketing folks that actually work with these big key accounts and they can get that information for you in real time. Because if you're too far upstream removed from the customer and consumers, you're missing out on all this really rich information that's being captured more and more today digitally and in real time than ever before. And that information is just sitting there and it can be used for forecasting. Guffanti: There's another question that's been asked in a couple different ways as well and this one again comes from a large CPG manufacturer. And Orkun, I'm going to have you handle this. Basically, in March we saw consumption go crazy. How do you keep the models from overreacting to one-time events like the initial COVID poles? And there have been other questions around historical data, will they be impacted by the blip that is the pandemic? We hope it's a blip. Ozturk: So like I said, there isn't a lot of things that we fully understand. But, there is quite a bit that we do actually. So, as I also said before, this was a one time, the first time that we have seen something like this. But, what we tried doing was for those platforms that we have seen a huge spike, we tried to understand, with our consumer insights colleagues, where this might be coming from. Is it pantry loading, as they say? Or, is it really consumption? Or, somewhere in between? Basically, after some thinking, thought experimentation around that one, we came to the idea that it's somewhere in between actually. So then, we started looking into our models and trying to model it in such a fashion to say that, "Look, some of this is just a peak. It's going to go up; it's going to come down." Like, bread subs is something that you can store. You can eat it as sort of a bread. It has a really long shelf life. But, we didn't believe that the bread subs were gaining popularity. We thought that there would be some popularity left after COVID is gone, but this would be mostly a pantry loading. We did probably for 50-60 of our platforms, and going back to the topic of variability that were coming from different geographies really, we kind of created a blue print of what variability would be coming from which source and what we think is going to be remaining. So, we basically fed these kind of things into our models and said that, "This is the best information that's going in." And like I said, for the places where we use that analytics, the results were really good. So in summary, it's a bit of a brainstorming in the beginning — putting our heads together and then making the models or adapting the models for future consumption accordingly. Guffanti: This one's for Charlie: How important is real-time data and short-term forecasting in categories with longer supply chain lead times, like apparel? This all applies for food companies and consumables, but for those with longer lead times, how important is it? Chase: Well, I think it's even more important because the longer the lead time, more importance is reflected on that short-term forecast and understanding changing consumer demand patterns. Again, it's very difficult, as you know, to turn the faucet on, as we say, versus turning the faucet off. You can always shut down. But in the apparel industry, it is very different than fast moving consumer goods, like particularly grocery products and the fact that they order a lot of their materials six months to a year in advance. The good news though is they might be able to redirect those different materials into other products that are in high demand. You would hope that some of the materials and colors that they've ordered for that season they have the ability to convert those into different types of products that are in demand based on what they're seeing through their direct to consumer sales because as you know, a lot of the apparel companies, particularly the big box retailers, are seeing a lot of online purchases for apparel products. So, that's a good indicator in short and real time that's those changing demand patterns. I would keep a close eye on that and then consider how they can take advantage of that information to redirect some of those materials and colors to other products that are in demand. Maybe people are buying more jeans than sweaters. Well, that's probably not a good example, but other products that have similar materials and colors that they could transition. But yeah, it's really difficult when you're purchasing materials to make products six months and a year in advance. So, that's another reason why their forecast needs to be very accurate. I would say the longer lead times you have, the more accurate your forecast needs to be. Guffanti: Kind of like the ship analogy that Orkun painted before. Chase: Yeah. We have our own analogy in the U.S. It's like an oil tanker going down the Mississippi River and when you get to New Orleans, there's a bend in the river, and in order to turn around at that bend, it takes a lot of work and about 20 different tug boats. Guffanti: I'm going to leave us with one last sort of big question that we can perhaps leave our audience with. And that is what can consumer brands do now to be prepared for the next pandemic, the next crisis, the next situation that we almost know will eventually happen? So, what can we all do to prepare and what should we think about as we now start entering hopefully a recovery mode? What are some of the things you want to leave the audience with? Charlie, I'll have you respond first and then Orkun. Chase: Well, I think digitalization is the most important thing. I think the one thing we've learned through this pandemic is that those companies who invested heavily in digitalization of their supply chain are the ones who are reaping the most benefits at this point. And collaboration with their key accounts and customers, CPG companies need to do that and be closer. We need social media, direct to consumer information, point of sale data, syndicated scanner data all play a key role in helping us identify those changing demand patterns very quickly. Investing in the cloud. We're seeing a lot of companies moving to cloud native products that they can leverage quickly, put up and running fast, spin those up and down as needed, run different tests on products as quickly, without having to invest in technology that takes weeks and months to install. We can now run an application within a few days and spin it up and connect it to the data within a matter of days and they're up and running. I think that those are key things, digitalization, moving to the cloud and collaborating not only internally across your organization, particularly with sales and marketing who are responsible for demand generation, but also with your key accounts and customers so that you can start to collaborate and create strategies. Because as we said, it's easy to turn off the faucet of the supply chain, but it's much more difficult to fire it up if you start to see a change in demand patterns for products that are not normally high demand products. And particularly when they're increasing at not 10% or 15%, but at 200% to 300%. I think digitalization is the key. Today we can actually manage the information right from the shelf using robotic technology, using shelf digital technology and send signals not only back to merchandising and planning, but back to the warehouse and then from the warehouse back to the supplier in real time. That's the autonomous supply chain, it's here. It's just a matter of whether or not you want to invest in it. All the technology's available today to be completely digital from consumer to shelf to supplier. Guffanti: Some really good food for thought there. Orkun, can you leave us with some of your wisdom in terms of what you're thinking about now as you're coming out of this crisis? Ozturk: Absolutely. And I will start with echoing Charlie here, so digitalization and companies have found this out. Not only mine, but talking in the industry, at least in the food and snacks industry, it's a common theme nowadays. Digitalization is key. It's fast. It's error free. It gives you the opportunity to experiment. So, it doesn't only give you a way out, but you also can see what different scenarios can yield in which different result. So, it's really great. So that's the key. Now, I think the companies will have to have BCP blueprints. I call them business continuity. They need to, now that we have data points, we need to, as companies, build blueprints. What could happen in this kind of situations? And we need to expand on this data point that we have. And digitalization is the tool to make that happen. Guffanti: Fantastic. Thank you so much. I'd like to really thank Charlie and Orkun for giving us such great subject matter expertise today. I think I speak on behalf of the industry when I say thank you for helping guide us through these unprecedented times. And finally, I'd also like to thank all of our attendees who showed up today and devoted some of your very valuable time to learn about the topic at hand. I hope you found it worthwhile and I look forward to having you on future webinars for CGT and RIS. Be safe and be well. Thank you. This has also been published on CGT.

Deep summer discounts are a slippery slope for retailers and brands

Consumers will delete retailers’ apps if value isn’t quickly communicated

Retail Minute: Instacart rolls out generative AI-powered search

You May Also Like

Retail Store Openings and New Concept Trends

Shinola and Filson Modernize Store Operations With Omnichannel Cloud Platform