 All right, welcome back everyone. Hopefully you had a good breakout session and what we wanted you to do in the breakout section just to recap and you can still enter in your solutions now is come up with up to three potential things you should do differently for running a large scale network design problem. The situation I gave you is that you're operating in Europe, you're manufacturing in a handful of different countries and you're distributing to all and so what should you do to try to come up with a plan that you think will be valid for five to ten years out and so the reason why I came up with this is because one it's this is something that's on everyone's mind who's designing a supply chain right now and there's some actual cases so for example Whirlpool which manufactures white good appliances like a washing machine's dryers things like that prior to Brexit the where Britain left the EU they had manufacturing in both in England in Great Britain as well as in France and southern part of France for both their dryers and their washers serving all of Europe and then because of Brexit they made decisions and they decided that the plant they're going to make that's in England will only produce product for England so we'll get a little smaller and then they closed the plant that was in France and moved it to loads Poland and so that will provide the rest of the UK so I mean excuse me the rest of the EU so what they did is they separated out now they didn't say I don't know how Whirlpool came up with this analysis but that's one of the results that came out for how they're changing their manufacturing network so Sergio what what do we get from some of the groups okay and please keep typing in your options okay so have a some a answer from a bar am I here so option one he says having a minimum capacity in relevant locations that are from 230 having additional capacity in every plan to buffer for uncertainty to achieve changing a product across facilities to manage uncertainty that option one no so so so essentially what we're saying here is add extra capacity so so the question so I think the big thing is hopefully people got the idea out of it you're going to have to run scenarios it's not going to be a one-time solution and in and so the real question is how would what type of scenario should I consider right and how would I what would I change for each of those potential scenarios and so what he put for option what he or she put for option one is what is all about it's almost a solution yeah it's all about a finale extra capacity so the whole idea of adding extra capacity and extra inventory is buffer that that could make sense but look what that would do for the that's one scenario where everyone has extra capacity but how would that help you model the uncertainty that would help you buffer if you think there's going to be demand uncertainty but the uncertainty here is different than what we talked about in our classes it's not just demand uncertainty which is what I really focused on when I did the New England root beer distributors case the real issue here is currency fluctuations or tariff changes between countries so going from from country a to country b what normally costs a thousand dollars a load might double and how will that change things so I think what you did for your first option is a possible solution but it's not a not something you would incorporate into your methodology you've already you're dictating a solution there but okay what was the second okay the second one is having production closer to seaports and a post okay so that could be a scenario where you move them closer but then so you just give them more options but again I think you're focusing a lot on solutions not methodologies to come up with a solution okay and the last one is explore contract partners manufacturing options today to address uncertainty yeah so that that's good but you should probably try to get a little finer grain but you could so you're presenting other options maybe we use a third party to do distribution where in a new country where we used to have operations because it's it's kept separate now no that's fair so what would you do so okay so in case if we talk about the methodology I think I will try to somehow estimate the potential new cost of the delivery transportation okay so for example UK or in this case England would be isolated mm-hmm so as you mentioned that the cost are gonna be higher you assume that's what I'm guessing that's what they did where the cost of distributing to Germany whatever suddenly increase so therefore makes sense not to distribute from there okay okay and I think that higher cost will penalize and will look for another option to know the places of a of you yeah and also I think the lead time mm-hmm so in this case you should be a larger than the initial and I think after that we a somehow to the estimate yeah because the lead time when we did some of the New England root beer stuff it was all about how far I can deliver from 1dc because we had a like we had to deliver within one day or two day or something like that you can think of the same thing here it was seamless if anyone's traveled in the EU the borders are frictionless but if suddenly they become with friction and it might take more time to clear custom so instead of an you know it's six a one-day trip it might turn into a one and a half or two-day so that might lead potentially as those tariffs as those time commitments go up to more DC's and it might lead to an ultimate strategy of every country has its own DC which is way back to pre-U days whenever country used to be operated independently from the others yep interesting okay let we have another potential a solution in this case is modeling risk management in the model with scores and acceptable overall risk score as a constraint as AC constraint so using Monte Carlo simulation on distribution of probability of disruptions so yeah that's a that's sounds really good it's very hard to come up with that distribution of disruption so what are the disruptions we're talking about currency fluctuations right because the idea is let's say France votes out of the EU if Penn wins the election what is the what does that distribution look like the pundits got the Brexit election and the US election wrong just wrong so do you how do you what probability disruption do you have for that I understand if I'm looking for like the probability of a hurricane or some weather or something that where you can look at over history these things political changes are not amenable to standard disruption kind of thing so that that's a harder one to do that's a go-to for a lot of people will be model it you assume a distribution it's a binary event and so you'd have to almost look and see what would what would happen there because some of the other things could happen they could implement a bat you know a value-added tax how does that change things so it's not so easy to say we'll just do supply chain disruptions and model with a probability that's why it's more of an art now and so I think what this would have end up doing is increase the number of scenarios that you have and you could try different levels because other things I've done in the past for things like this we tried to find something for example back in the early 2000s the price of fuel was much higher than it is now there's actually a glut of oil across the globe right now but when it was very high we tried to figure out and this was a project we did actually in Spain we tried to figure out what would your DC look like with the current price of fuel your your whole distribution network and then what would the price of fuel have to do where you would change what your network looked like and so the one way to do that is to repeated scenarios where you keep changing the price because you know your arc rather would increase in cost and at some point you go from say five DCs to three and it might change or five to seven because usually if you increase the arc cost then you're going to have more DCs closer to your markets right so we wanted to find out where that break point is because then that gives you an idea do I need to worry about this or not so one thing you could do let's say we isolate this down to the whirlpool problem and we could say okay if I'm producing in England and I'm shipping into the EU and the current cost is this what would it have to be where suddenly it would not make sense to ship and where I would ship that capacity to some other plant and you find where that break even is and if it's 10 more than what it is now then you want to worry about it if it takes you know a 5x then you know not to worry about it so the big underlying thing for this is to do multiple scenarios and not just to see point solutions but try to see if you can derive as best as possible where there's a break point where suddenly your network changes because some underlying factor changes and some of those would be the transit times which is a reflection of the customs currency fluctuations tax differential changes so you can do a series of runs to try one at a time what the impacts would be and if you have more time to try to change two things at once but generally it's easier told everything constant and change one thing and then have everything constant change another thing um as you get into changing multiple things at the same time then it gets a little tricky but if you think there's interactions then it does make sense so that was a long answer but the there is no single solution to this to this problem the big takeaway is you need to do multiple scenarios that are smart not just try it with 10 percent more 10 less you have to think about the underlying cost drivers there would the political cost drivers that would change your network design excellent thank you crazy we have a one question about the currency sure it's but found appreciated after Brexit with my still just the five experts from the UK yeah that's a good question um it it appreciate didn't come back a little bit I don't know how much it's come back because there's questions now it's such a fluid situation so a lot of times you want to it's a valid point you'd say what would the pound have to go down to where it actually makes sense to keep it in there and distribute it that it's a valid point so these models what's cool about them you can have all these different assumptions and test different ones so I could see a simple two by two where you look at the differential in the currencies right have a pound if it keeps depreciating that it might make sense to produce more there um but then the customs and any kind of the trade costs any barriers or additional costs for distributing you'd also have to look at at the inbound cost for the raw materials and see does it still make sense to bring that in to to anyone does that change things because you've got two you got coming in for the raw materials going out as a finished product and you have to look at how those um costs map up against a value evaluation for the the pound itself okay thank you very much and you you're talking about these these scenarios and we have a question and we call these sensitivity analysis this generation of yes scenarios is the same as well the doing scenarios gets you at sensitivity analysis right how sensitive is my model solution to certain things right and so I can increase um so like in the example I gave with the price of fuel what we did there was just keep increasing the price of fuel which naturally increased the cost of of distribution and so at some point you hit something and the network change there's always going to be some dead zone where it has no effect changing the price of the holding cost by one percent might not have any effect or it might be right at that break point where it totally changes things it's always helpful to know where your parameters are because then you need to make a quick decision do I care should I care about this or is it something that's really important or is it something that you know what the value would have to increase by 10x for it to change my underlying structure and that's that's the trick that's the art to try to get one to identify what those drivers are and then two how do I incorporate them into my model because all the model does is fixed invariable costs on nodes and orcs that's it right that's all that it really does so you have to try to use the take the political and map it in to the things I have available for me in the model okay thank you and we have another question from from head does this problem meet with what many manufacturers had to deal with in operation in Hunton and facing the July 1, 1977 they changed in governments yeah that's you know I I'm not sure exactly what they did in 97 I think there is there's less manufacturing in Hong Kong more trade for that but I know a lot of decisions are being made there when Great Britain they had to hand over back to the Chinese government for the for Hong Kong and it's still evolving because remember nothing was supposed to change and now there are changes happening so I'm sure they did for scenarios maybe for different industries I think it affects the finance industry and shipping industry more than pure manufacturing because I think by that time a lot of manufacturing left Hong Kong for cheaper labor locations but absolutely people there's a lot of uncertainty about what would happen as soon as the the government changed same thing happens in Latin America all the time right because there's a lot of uncertainty look what's happening in Venezuela General Motors just pulled out of Venezuela as a whole because the state took over appropriated its plans so this stuff the political stuff it's so hard to predict and that's why in basic classes we usually say okay look at demand fluctuation because the political stuff it's tough and as you'll see if you take sc3x a lot of the decisions of where to locate plants and DCs and things it's driven by taxes and so we'll talk about a case where the pharmaceutical industry there's a reason why most of it moves through Puerto Rico in the United States it makes no supply chain sense but it makes incredible tax sense because of incentives that are there there's a reason why Ireland is known for a lot of medical devices it's a tax incentive so with these political things changes a tax incentive can be changed overnight and so if you build your supply chain on stilts and something when it still gets chopped off then you've got to rebalance yourself so that's why the political uncertainty seems like it's it's not as important it is what drives a lot of the financial incentives and taxes that underlie most supply chain network designs at a global level and so your trick is part of the art is how do I incorporate these softish kind of things that I don't know with certainty into the modeling which has very hard levers that I can change fixed and variable cost nodes and arcs and then some kind of constraints yeah so that's a challenge we have another question would it be fair to say that we need to model against the outcomes transit customs and not believe the event the election in this case that's that's a really good point so um in sc3x I talked about scenario planning and this would fit into some of that and the beautiful thing about scenario planning um what we talked about is instead of trying to predict the future and I call it like boxing where you try to do a point forecast everything we did in sc1x when we did exponential smoothing or causal analysis to come with a point forecast like boxing where scenario planning it's you say here's an outcome how would I adjust to it and so when we talk about scenarios here for network design we're saying the world is going to look like this what would your supply chain look like in a world like this when customs takes twice as long the trade bearers of a certain amount so what we're doing is we're saying there is a situation how would you have the optimal design in that world um and so what this allows us to do is move away from trying to predict different events to understanding their effect so that's the challenge for an event an event could be brexit it could be the same effect that brexit cause could have been caused by a war could have been caused by a certain pandemic where you can't have things coming in from england to the rest of the coast think of the the zombie apocalypse if it happened right but anything like this there are many many many events that will have the same effect on your supply chain and so one way to boil the ocean and not try to predict every possible event that could happen is to think back what are the main effects on my network and for the case with whirlpool it was where to manufacture and where to distribute from so the idea was if if england gets isolated then it can't produce out what would my that that's an effect there might be many events that cause that so i think what the question was was can i focus on the effects and not the events because it's impossible to predict and forecast events from happening it is easier to prepare for effects that are caused by multiple potential events did i answer the question yeah i think so okay thank you thank you please we hope we still have time for it for a couple of questions can you read it right yep we don't have any all right any final comments Chris no i think i've talked enough okay excellent thank you so thanks again guys for joining this just a couple of reminders the verification upgrade deadline is made the third okay we have two due dates for data assignments the due date for week one data assignment is April 26th that's tomorrow and the due date for week two week two data assignment is made there so you have one extra week for this okay so enjoy the material and see you next time all right see you guys bye