 Hello, learners, and welcome to the Second Life Event of S1X, Supply Chain Fundamentals. I'm Ima Borrella. I'm a postdoctoral associate here at the Center for Transportation and Logistics at MIT. And I am the co-lead of S1X, as you already know. And I'm really happy to be here today with Dr. Larry Lapidi. Hi, Larry. Hi, Emma. Thank you for inviting me. Thank you for being here. Dr. Larry is an expert in forecasting, and he's an OP. And I would like to briefly introduce his bio, because he's a very long and interesting bio. I'll try to do my best to summarize all his achievements. He holds a PhD in operations research from the University of Pennsylvania's Wharton School. He's an electrical engineer with a master's from MIT. And he currently works as a lecturer at the University of Massachusetts. He's also a research affiliate here at CTL in MIT. And he led a really interesting project, Supply Chain 2020, that focused on understanding the future of supply chain management and led the demand management research group at MIT. His experience in business is also very wide. He has more than 30 years of experience in supply chain and marketing, experiencing consulting in the high tech sector, and also as a market analyst. And he has been teaching while he was also working in business on a part-time basis. He's been part and helped several professional society officer positions, including a few years serving as a board member and treasurer of the supply chain council, president of the Boston chapter, the Institute of Management Sciences, sorry, now informs. And most importantly, he got a great award. He's the recipient of the Lifetime Achievement in Business, Forecasting, and Planning Award, bestowed by the Institute of Business, Forecasting, and Planning in 2012. Congratulations for that. And he has written numerous articles, including writing an ongoing column in the Journal of Business Forecasting since 1997 and supply chain management review since 2008. So great value. Thank you so much for being here, Larry. And so we're going to learn a lot from you today. And the agenda for today's life event is the following. First, we will discuss a little bit of forecasting in a business environment, because you've learned a lot about forecasting. But Larry will share with us his insights on how it is that you develop a forecasting function in a company. Then he will explain why he wrote the article. We will discuss today this article. Don't just measure forecast errors, but use them, I guess. So he will introduce the article. After that, we will go to the breakout sessions, where you can discuss with your peers about two questions we will propose you. After the discussion rooms, we will come back here and we will discuss with Dr. Lapiti on these two questions. And also, you will share your insights with us so we can comment with you. And after that, we will wrap up and close the life event. So Larry, the floor is yours. OK, thank you. Nice introduction. I'm just going to put my glasses on, because I may have to read some of my notes. But I'm glad thank you very much for having me here. I'm talking about a topic I've been dealing with for over 30 years. It's business forecasting. When we go to school, when I went to school, we learned about forecasting, but it was all about statistical methods. And those are very important. But in a business environment, you also have to worry about the process of forecasting. So if you look at forecasting, it's a business process. It's got to be done. It's got to be done routinely. And it's got to be done, I believe, by a very professional business forecasting organization focused on forecasting. So I want to talk first a little bit about forecasting in the business environment. And over the last 30 years, the big trend has been into something called one number forecasting. One number forecasting does not mean this is one number, but it means this is one set of numbers that the whole organization is planning on around. So we have to know what's going on in the future in terms of revenue. That's the thing that drives all the operations in business. So the marketing people have to plan in their area for the revenue we're expecting to get and create and shape the salespeople as well. Supply chain, we have to get all the supply chain resources in place in time to meet customer demand. The finance people have to worry, do we meet our financial performance targets? So everyone is interested in that business forecast which says, how much business are we going to get from customers? What is going to be the revenue? Because we're trying to get to one number accountability and commitment from everyone that that's the set of numbers that are going to be the revenue numbers. And then the people are going to plan around it to make sure we meet the profitability targets as well over time. Okay, so that's kind of what we mean by one number forecasting. Now, what does a forecast to do? Well, basically we build models and we quantify. That's what we do. We take all the marketing and sales activities and plans going forward, maybe six months to 18 months out. And we look at what's going on, it has gone on in the business historically and then we look at potentially what might happen in the future competitively and to the business environment and we develop a set of models that quantify the numbers. And what we normally are working on numbers that drive the rest of the organization. And we'll talk a bit about that when we talk about the pyramid. So that is really what we're doing. But if you remember, the people who are responsible for creating and shaping demand in a company is marketing and sales. So a key thing is to work very well with the marketing sales organization to understand what they're trying to do to impact the future for revenues themselves. And the key thing is also we are looking at unconstrained demand because we want to meet unconstrained demand and we don't want supply to constrain that. So we can achieve revenue targets because supply doesn't have it. That's why it's important that supply chain also take the revenue forecast and use it to create the supply plans going forward for the next six months to 18 months out. And if you look at it from a forecast then what's another thing we do? And I always talk about the thing we work and look at in the past is variation in demand. And the reason why we look at variation in demand is because if simply put, if demand did not vary every no reason to have a forecaster. And the forecasters are important because we have to understand what impacts the business and why those demand variations have happened in the past so we can also explain what are the variations in the future that might happen. Okay, so that's what it's all about. Now, the thing that I like to talk about is the permit and we're gonna pop that up right now. Do you wanna pop up that permit? Okay. You have to get buy-in for your forecast. You have to basically, okay, do we have it yet? No, you got us, okay. Technical problems. Share. Okay, so these are a set of forecast and plan hierarchies. I think they're the most important thing to be doing anytime you've been developing a forecasting and or sales and operations planning process which we won't talk much about today but these are the hierarchies you have to create. And I describe it like you have to take the forecast and even the plans, if you would talk about that and you have to put it in multiple language. What are those languages? People in the company talk different languages. So if you'd like to take a look at the left-hand side of this pyramid, the pyramid called demand side views. How do people talk about the business in that area? So if you look at sales, many companies make mistakes and try to get the sales organization to agree to unit-based forecast. How many of an SKU am I going to sell? They really can't do it that way. They don't think that way. They think in terms of how much they sell to each account and dollars in the US, of course. In other countries, it's whatever currency you have. So when you talk to sales and you try to take your forecast and roll it up to them, this is all about aggregating the data. And we're aggregating the data to their language. And their language deals with dollars, money, which is really revenue. And they want to see it by accounts. Maybe they don't want to see it by every account or maybe they want to see it by their major accounts, right? And they can look at it and say, well, you're forecasting this for Walmart. And we don't think Walmart's going to do that. We think Walmart's going to do less than that. Or so they can respond to it when you present the forecast for Walmart. And also by regions. So the US rolls it up. You roll it up to the US level. You roll it up to Europe. You roll it up to Asia Pacific. And they have to look at it and say, that looks okay from a dollar perspective in terms of what we can sell. So you're in the language of sales. And that's the language. And then obviously it's at the company level. So basically the view as we aggregate the data from the lowest level has to be that view to them. And I always say to salespeople, speak Greek because we have an expression in the United States. To speak Greek, nobody understands you. And unfortunately in many companies, the sales organization has their own unique identity and not everybody on supply side can really communicate as well with them. But that's a different story. Okay, now how about the marketing people? Okay, I was a marketing person. So I know how marketing people think. And they think in terms of, they can think in terms of dollars and units as well. They don't have to think it's just dollars. They also think of units. They wanna see things by brand. And so if you show a brand manager at a company, what the revenue forecast is and they look at it and they say, yes, that's what I was expecting from my brand. I can commit and be accountable for making sure we generate that revenue in that brand. And each brand manager would wanna look at it and make sure that they're comfortable with the revenue target. You roll it up even further, brands are in groups. And so you get groups and basically you roll it up to product groups. And again, the person responsible for the product group also wants to see it and see whether or not they can be accountable and they commit to making that revenue number. And then you go to company. So that is the marketing view. That's the demand side view. Now we go to the middle one, which is the view we all love, supply chain, right? What is supply chain people? How do we think? Okay, so if you take a look at logistics, we talk about units in cases. We very rarely talk about revenue. We wanna see the forecast in terms of units in cases because we can do something about that. So we wanna know something about the geographies. Where are the ship to locations? How much are we gonna ship to the West Coast? How much are we gonna ship to the East Coast, United States? Maybe how much we're gonna ship out of Asia, Latin America? So the ship to location view is some view that we can then feed the transportation people who have enough transportation capacity to achieve that. Then we roll it to warehouses. How much revenue is gonna have to be sent out from the warehouse? We don't care about revenue, but we care about cases and units. How much is it gonna get sent out? And is this forecast make sense that we even have capacity to do it? Do we have the capacity to receive materials and product? We have capacity to load and ship it out. So they wanna see it by warehouse and they can say, okay, we have enough warehouse operation capacity or they don't, right? So that's the logistics view. Then the manufacturing view, they have production lines. Is this revenue forecast? And again, in terms of units, is this revenue forecast meaning that we're using 80% capacity of the line that might be okay, it's all right. But if you're gonna put 150% on a production line, I only agree to it if I get overtime where I can do another shift, all right? So basically the manufacturing people have to be comfortable or we can produce what the revenue forecast says and it's all in terms of units. Okay, so that's the middle one. Those are the good guys, remember, because we're supply chain. Now we go to finance. We got two views of finance. One is the budgetary units. In other words, the finance people are responsible for budgets. Budget is nothing more than an allocation of resources we do once a year and we do it they wanna look at revenues and margins and say, okay, what are the margins and the revenues that are gonna result from this forecast? I wanna see it by operating unit to see whether or not we've got adequate resources to support it. I wanna see it by divisions and business units and to make sure we're supporting the divisions in that way. And then of course we wanna see it by weeks, quarters and years because I want to see if we're gonna meet financial targets. Normally we're on a annual performance targets for finance and so we wanna make sure that the forecast is consistent with what we're trying to achieve from a performance perspective. Okay, so that's the hierarchies and the last thing I wanna talk about is the models themselves, okay? How do you make these models that people will be leaving? Okay, because now you've got the looks but how did you get this numbers? If I don't agree with it, how do you convince me as a marketing person, a sales person or a financial person that it's a good forecast, okay? And here's what you gotta do. You gotta do first of all, the models have to look reasonable. I know when we're all in school, we learn very fancy mathematical equations and stuff like that. That doesn't fly in business. It has to be understandable by somebody. Even if it's a complex formula, it's still gotta be reasonably understandable in terms of what you're doing. So it has to have what I call face validity as well as validity. And so validity is accuracy. Face validity is, looks reasonable, okay? So that's the one thing about that. You also have to build it up. You have to build up what we call base business first. Turn business, that's typically a statistical forecast which takes the history and moves it forward but he promotes the data, gets all the promotions out and then new product sales out and basically gives us a good historical and to reject from, okay? So we have what's called base or turn business. On top of that, we add on business. We're gonna get from promotions that get run by marketing and sales. On top of that, we're gonna produce new products that we're gonna be putting right into market. We gotta put on top of that. And then we have to bring in some marketing intelligence which is a marketing intelligence from people in marketing and sales tells us what's going on out in the business, okay? So that's a, you build it up in four ways. Now, so that's a good thing because then you can tell people and make it clear what's going on in the pyramids. Second piece of this is you want to basically focus on facts, figures and assumptions. Now, what do I mean by that? Typically, you've gotta get buy-in and commitment and accountability from all the organizations. Let's say the salesperson says, I don't believe that number, I want you to change it. I'm not changing the number. Why? Because what I'm gonna explain to you is what the facts, figures and assumptions I got to get to it. Those facts, figures and assumptions led to a number because I have a model to get to the number. So if you wanted to dispute the number, you have to dispute either the facts are wrong, the assumptions are wrong or the figures are wrong. If you can't say they're wrong, you're not responsible for the model that was what I do, which is to take those facts, figures and assumptions and quantify it to a forecast. That's what I get paid to do. The salespeople can respond to it, but then you got, if you want to change it, and you have to disagree with one of the facts, figures and assumptions or you got to do something different, right? So that's basically the way you defend it in an organization that has to be defended, okay? But not the numbers themselves. Fact, figures and assumptions are behind. Okay, now. So that's the first thing I wanted to talk about was forecasting the business environment. Next, why did I write this article? Basically, it's very simply put. Forecasters are ashamed of their forecast errors. I'll be honest with you, we've done a lot of benchmarking these forecasting. You're ashamed of the numbers. 50% error on maybe a stock keeping unit on location is not bad. It's actually a good forecast because with all promotions and stuff that get created, it's not bad, okay? So essentially, I wrote this article to say, you got to learn from your errors is one thing and you also have to be able to tell people what the errors are so they can plan for risk management, okay? And that's what is in the article. So I think we're at the stage where you're gonna be addressing the two questions and then we'll talk about it, okay? Thank you, Larry. That was extremely interesting. I think it's gonna be very useful for our learners. So now we're gonna go to the breakout rooms. I hope all of you have read this article. You can download it from the section, live events, second live event in the course. So you don't have downloaded it, download it now and go to the breakout rooms. We will be discussing these two questions. Are you showing them Arthur? Thank you. So question one is how might you analyze your forecast errors in order to improve your forecasting? How will you use them to improve your forecasting? And the second question is in what ways might a company use forecast errors in order to manage risks? So is this idea that Dr. Lapidia just put out? I'll say you have these errors. Usually people are the same to them, but they're usually, they're really, really useful, really valuable. So how could you put them in use to improve your forecast and to manage risks? So please go to the breakout rooms, discuss these two questions and we will be back here in 15 minutes.