JP

Thank you to those of you who were able to provide some input into this discussion, it is really helpful to help us make sure that we're going to make it as good use of the sixty minutes as possible. Alan, Kane and I got our heads together last Friday to go through the responses we had and plan the agenda for today: firstly, to focus on the sources of volatility; what it is that we're trying to predict, the obstacles that people are facing in being able to get a better handle on that volatility and the extent to which predictive planning tools and technologies like machine learning and AI are really able to deal with those things or to what extent there may be more hype than there is reality. 

We'll get into all of that, and if we have time, we'll touch on the ESG piece as well, which I think is increasingly becoming part of the mix that people can't afford to ignore anymore. Hopefully that meets with all of your expectations. I will invite Alan and Kane to say more about themselves in due course. I just wanted to start with you guys first of all. Just starting on this point around volatility, it seems that to me at least, there are kind of maybe three buckets of volatility. There's what you might call as extrinsic shocks, so pandemics, black swans, that kind of stuff. At the other end of the spectrum, there's what you might describe as kind of intrinsic volatility…the stuff that maybe really ought to be predictable. I'm talking about demand, supply, capacity volatility which, for a number of reasons, are not predictable, or at least not stable. There's a kind of third group which kind of is in between, which I could maybe describe as trends. A few people have cited recession, economic instability, other global changes in trade patterns, those trends that are emerging over time. I'm going to start with H, if that's all right. H, for no other reason than we had an interesting discussion just the other day about this and that you had a particularly interesting pandemic, I wonder if I could ask you to kind of reflect on your experience of those different sources of volatility.

H

Okay. For the last two years, I've been working for a German sushi business with interests all over Europe. I fell into that after the start of COVID so what they saw as a business was because their kind of style of selling is through supermarkets, they saw lots of people move to buy their product because they couldn't go to restaurants…they wanted a treat. As part of their weekly shopping, they'd grab a box of sushi. We saw sales go up by 15, 20% as a consequence. What we tried to do was replicate that on all our European businesses by having a mad dash, really, to expand very quickly. What we then saw as some of you will probably be familiar with is massive increases in raw material costs. Sushi, as you probably know  it's fish and rice and a lot of that fish comes from Japan. 

The rice either comes from Italy or Japan or China. The costs involved in transport, because many of the containers needed to move that stuff from where it was to where it was needed were stuck in China or they were stuck in the Suez canal when that ship got stuck. What I saw over the period of two years was massive changes in demand, massive changes in costs, all of which made trying to determine where you needed your inventory, where your sales are going to go etc. an absolute nightmare. I think, JP, that's what you were pointing to in terms of recent history, what my experiences have been. 

JP

If you could just maybe just touch upon the kind of obstacles that you faced in terms of being able to deal with those challenges. I know it's a big question, but just to kind of give us a sense of where your focus was when you're trying to deal with that volatility. 

H

So initially it was all around cost. You've projected you can only sell sushi for so much and so you've got a cost constraint that you need to buy the product at. There comes a point where you just physically need something to maintain your brand, to keep it going. We were having to move very quickly to find alternative sources of rice, salmon, tuna and other ingredients. I was quite new to it. The relationships were quite new in the countries that I was operating in. What we weren't able to do was fall back on good, solid relationships. You're scrabbling around trying to find salmon or tuna or whatever it was. You haven't got the rapport of having dealt with someone for years, which is a position I'm more used to, so you can't really call on those relationships. 

What I ended up doing was having a very disparate supply chain, where I had, rather than having one, two, three suppliers of these things, we had many, which in that world was interesting because it's a high risk offering. What you need to have confidence in is the quality in the case of salmon and tuna, where you're effectively serving it raw, it's going to cut the mustard in terms of taste, safety, all those sorts of things. Normally that kind of due diligence takes time to do, but when you're trying to maintain supply, you're having to do that at speed. You have to throw resource and therefore cost at it. 

Kane

H, sorry, that must have given your technical department a nightmare because the reliability, the freshness and everything else, you have multiple suppliers all landing at once. That must have been a nightmare. 

JP

Thanks H. P, if I could come to you please because you said something a bit similar. Very different business, of course. Maybe you can just give us a sense of, looking forward, what are the sources of volatility that you're worried about? Not necessarily retrospectively, but the things that you would like to get a better handle on if you could. 

P

Obviously, throughout that period supply chain costs were horrendous as was the reliability. I've just made a note actually about security of supply chain or, at least, the stability of it allows you to run a more lean operation. The minute you introduce complexity and risk your buffer stock needs to go up in terms of being able to supply your customers. For us as a high street or as an omnichannel retailer, we saw stores close, we saw ecomm boom initially. Trying to manage that was very difficult and forecasting almost went out the window. You were guessing on a daily basis what was going to happen in terms of the volume. Post pandemic, we've seen volume go back to the high street and we've seen ecomm decline, interestingly. 

Of course, the cost of living has seen price increases and required us to monitor and manage our own costs as customer behaviour changed significantly. Ours being a relatively luxury product, you have to understand people are not going to be spending as much and as often. That's seen promotional activity change. Trying to forecast that and how that, not only changes stock, but changes my distribution, changes my intake and overall stockholding has been extremely interesting and still is. I've just come out of a budget meeting, actually, as we start to look ahead to the next financial year. It's funny that myself, the CFO, head of commercial, we've got the head of retail in there too and we're all coming up with completely different answers about what we think is going to happen next year and what we might want to do in supply chain. Really interesting conversations. 

I'm trying to take the business on a data driven journey. Our forecasting, as we talked about in a previous meeting JP is quite old school, it's very manual, it's spreadsheet driven. We're trying to look at AI and machine learning and so on. The argument back to me was yes but on what data? Is it going to try and run algorithms? Because the data of the last two years is so unique and indifferent, the answer you're probably going to get out of a machine is going to be nonsense anyway. If anybody's got a crystal ball, let me know and I'll have a go at it!. 

JP

Would be worth a bit, wouldn't it?! Thanks for that, P. I see quite a few nods in response to that. S, I wonder if I could come to you next because you mentioned particularly the potential effects of a recession, but you also highlighted the extent to which machine learning can be used to predict those, perhaps, inherently unpredictable things. What are the major challenges you're hoping to address?

R

So, good afternoon, everyone. In terms of my responsibility, my company is a multinational Japanese company, manufacturing beauty products. Related to what P just said, which is you put rubbish in, you get rubbish out; from any machine learning tool software, the more data you put in, the more it can give insights, but if what you're trying to put in is not accurate or incomplete, you're not going to get the answers you need. So, in terms what I have been doing in my team and my organisation is we have been trying to not go externally. We keep a lot of things internal in terms of the tools we are designing, because we still have a high rotation of individuals within the organisation, which means the pool of knowledge goes away. Any machine learning on any statistical run on the historical data, that history stays with those individuals who knew that X event happened at that time on that customer, at that price. That is individual knowledge. The thing that AI machine learning needs to do in the future is capture the information of such events and model them because if we then identify what that pattern is, and name it,, then you can then predict similar circumstances in future. 

For unpredictable events, the most used approach today is demand sensing. What indicators do we need to have as a capability to identify those patterns? What are we trying to look at specifically? Is that the volume, is it the price, it is the customer? Are you looking daily? Are you doing it weekly? Hourly? What is your indicator? That is the capability also that you need to have available to formulate the question before you need to get any answers? I think that's the capability that needs to be made available. What I need to gain, what I'm hoping to gain is to develop capabilities, develop the individuals with the support of a system, the support of an algorithm, but it only gives out if you know what you want first. That is the first priority. I don't know if that helps you understand, JP, what I'm inclining into. 

JP

Yeah, I think it does. In fact, we were having a quick discussion just before we jumped on the call earlier about machine learning. In theory, it can do anything that you want it to, but you need to define that. I'm sure the guys are going to pick up on that when we open up. I just want to bring a few more people in before we do though. L, you're on a very broad transformation of supply chain. What’s resonating with you? Are there any other aspects that you're also grappling with at the moment? 

L

I think the whole piece on data quality is something we are massively facing into because we are absolutely going on a supply chain transformation journey and it's actually taken quite a while for us to get our house in order behind the scenes so that we can really make a go of the capabilities that we're trying to drive out. Things like demand forecasting, how we replenish, how we plan our networks. That certainly has been an interesting ride. It's also been quite interesting because of the scope, I guess within XXX we have not just food, but the GM side as well. We have the XXX business and actually still bringing together those worlds behind the scenes and what's the future of our network. 

There's definitely been an interesting focus on, I guess, the different products that we have, the attributes that they have, less so being channel specific of this is GM, this is food, but actually, how can a product flow based on what is it? Is it a long lead product? Is it short life? Long life? There's so many different ways in which we're now looking and chopping our data up in different ways so that we come out of most of those silos, of those channels and look holistically across our product range and what are the capabilities that we need to be able to deliver to all of those needs. 

I suppose we've seen throughout the pandemic and how everything closed off all into ecomm. Now people are coming back into stores again. There's definitely some customer behaviours that have changed, but there are some that feel like they're here to stay. I think we've really seen a lot with our supply chain and supporting our suppliers...they're facing all of the same things that we are. You're feeling that combined impact of all of the different disruptions happening in the supply chain. We're feeling that with our suppliers as much as us, then passing that on with bringing those products through our own network and to our customers. 

JP

B from XXX. I could see that you're identifying with a fair bit of the last few comments there. How's it looking from your side? 

B

I think I echo a lot of what has been said already. I think what's been different more recently than was previously is that the shocks that we've had to the system have been so large and so sustained. I think we're all used to supply chain disruptions, we are used to a big order spike or a change in the market or a supplier disruption or something. We all grew up as supply chain managers coping with those and managing that through the system. What's come through the pandemic and since is that some of the products that you wouldn't have thought would have been affected have been. You never had any buffer stock or relatively low buffer stocks, or you could always go to another supplier. Also some sources of supply have been hit for long periods and that's kind of really changed our thinking about what we need in terms of supply chain resilience. 

To give you an example, we make a lot of products that go in little plastic bottles and you could buy plastic bottles from ten different suppliers and it made no difference which one supplied it to you. During the pandemic, you couldn't get a plastic bottle for love nor money. They were all out of supply. 

You can't make anything. It's not that we can make a different product. If you haven't got anything to put your products in, you can't make anything. That's really changed our thought processes in what we keep in terms of strategic inventory and our relationships with suppliers to cope with that type of disruption. I think that's the biggest shock that we've got is just that…not that you've got to cope with difficulties, but the difficulties are just so big that you've really got to rethink how you manage your business to protect against that. 

JP

Thank you, B. A from XXX. I wanted to ask you the same question…what sources of volatility are the most concerning you?

A

Yeah, of course. We are in quite a big transitional phase internally anyway, which isn't helping the external factors. As a group, we've now decided to move into other products. Our biggest challenge, outside of the ones that everyone's mentioned so far around import and exporting costs and some of the volatilities of supply of components or product etc., my biggest challenge is more around the people than it is around the data. The data is difficult, but I have a business that has been traditionally seasonal for the last 14 years, and I've now upset it by bringing in SDA and opening up other channels, marketplaces and Ecommerce. 

That's not what these guys are used to. My biggest challenge is people, which is not what I was expecting, but data has been the biggest driver that's helped people understand the vision that I've got and the way that we're trying to move. From our perspective, I concur with everyone. We've had very similar issues with supply of component products. We couldn't sell in 2020 because we couldn't buy key components. Conversely, we were lucky enough to have had different products during October and November. That was a winner because it was a key hero product but it's adapting the supply chain and the people within the supply chain to move ugly freight and small freight. That's been my biggest challenge, trying to get people to come on that journey with me. 

JP

Okay, thank you for that. A. N, just before I open it up so anyone can jump in, I just wanted to ask about your situation and the challenges are that you're looking to tackle through greater predictability. 

N

I recognise everything that everyone said, and maybe I can add another one. The war was a big one for us last year, so we did have one site in the Ukraine and two in Russia, which we now subsequently don't have. That threw a very different dynamic into the business last year. 

Going forward, I’m expecting to have lots of debates about forecasting. I’m fairly new to this industry but most of my colleagues have been in it for 20-30 years. Nothing changes…we produce the same thing roughly day in, day out. But change is coming. A lot of companies are getting more, let's say, adventurous tapping into new markets so the portfolio is changing radically. My colleagues will say, well, it keeps changing a lot, but it's at what level you choose to forecast. We've got different dynamics growth, we've got, obviously, regulatory pressures, tax changes. Obviously they are different by country as well, lifestyle choices. You throw all that into the mix and it's like, how the h*** do you predict all that? As we kind of look at the higher levels and we aggregate the data up, there starts to be some kind of consistency and some stability there. 

So that's what we're looking at. I've still got that challenge of changing the mindsets internally. Who will own the forecast, who will facilitate the forecast? Currently, today, we are very much dependent on receiving forecasts from our customers and after about three months, they don't really pay much attention to it. When you get to capacity planning, inventory planning, do we need to maybe invest in assets? Do we need to reduce our footprint? All those decisions that we kind of were kind of thinking about two years ago and then the war happened, well, guess what? We started throwing money at the problem and we spent in excess of €20 million within the space of four weeks to try and get ourselves out of the hole. That's what I'm trying to avoid today. 

I think, B, you mentioned resilience a number of times. BCP [business continuity planning], which were more paper exercises with our customers in the past, are now not paper exercises. They want to see physical demonstration of it. That's where I need a decent forecast. I need to understand the network and what the future requirements are and what does it mean? What does it mean to supply chain, what does it mean to the bottom line? That's where I'm struggling because I don't have a decent forecast and I don't know what that means in terms of agility because traditionally the business just filled every asset. I come along and say, well, we never will be agile until we target decent utilisation. That's a big change for the exec team at the moment. 

JP

That's all, eh N?! Well, thanks for that. Thank you to everybody for giving us a really good picture there. I'm going to invite Alan and Kane to pick up any of those issues. We’ve covered the volatile situation, the need for resilience, the challenges around the data, the processes, the siloed aspects of how businesses work in practice. Which of those bits do you want to pick up and unpack? 

Alan

Well, I was going to say Kane, you're more like the strategy minded one. I'm more into the data a bit. Why don't you kick off and be controversial and say you disagree with everyone. Go on, I dare you!

Kane

I don't disagree with everyone. The problems and the realities are very real and they're there. What I have found quite interesting and forgive me, L, but please speak to me afterwards, I do apologise. You literally listed every conceivable way of what you're doing and how you're looking at projecting forward but you didn't look at the environment or changes to any of that. The environment or sustainability changes didn't get to factor into what appeared to be any decision making. Now there is another similar company to yourself, not the main one, that I've been talking to recently who are doing a similar transformation, who are adding sustainability data in to drive their decision making. So when they were in no man's land of ‘what am I doing?’, they decided actually what's the environmental impact? Maybe that will be the decision maker. They have found that's impacted financial changes but yet causing nightmares on supply changes because obviously the routing around different vehicles, methods, containers, et cetera, there's still a huge volatility around what method people are going to use for the future. 

Forecasting: how that impact is going to happen creates another issue. But I just found that quite interesting. So apologies for choosing you. 

L

No problem at all. Happy to provoke the discussion and I think it isn't something we have left out either. I think that's definitely on our minds. Really, as we enter the next 5-10 years, sustainability is going to be more and more prominent and we're going to have to be more and more inclusive of that in our decision making, either by legislation and requirements or just fundamentally doing the right thing for the environment. That's absolutely on our mind and what we're also considering is we might know the data that we have today and what we have but it's what other data could we be looking at and using to better inform our decision making to feed our systems to help us make those decisions. It's definitely not left out but you're right, it's an important topic to talk about. 

Kane

It just makes the data even bigger because you've now got to start adding the volumetrics of kilowatt hours, et cetera, or whatever the measurement is alongside in that data collection. 

Where do I put it, where is the calculation going to come from? AI is there and there are tools and platforms out there that basically slot it alongside and leave it separated. I do feel for everybody because the data over the last three odd years is everywhere. Where do you find your baseline? Where is the point of truth, where do you start? Do you start during the pandemic? Do you go back two years prior and try to find a trend or a flatline that makes any sense? 

I think the algorithms are going to struggle with that, but I'll hand over to Alan to put some more sense around the data.

Alan

So, hi, everyone. Really great to hear your stories and those insights. I think the first thing I'm going to do is apologise. I'm an accountant by training, so I've focused mostly on finance before moving into supply chain, really for the last 20 years. Companies haven't traditionally been very good at working with their own internal data. If you just think about how budgeting and planning cycles happen, it starts off in finance, it feels like it gets thrown over a fence into operations and then it comes back. We're all aware of these sorts of silos as everyone has picked up on. 

It's not just internal data anymore, there’s a lot of external data. It's not just how things relate to one another within your business, like integrated business planning…it's really understanding and contextualising where you are in the world and how the world's influencing you. 

A lot of data…you've got to process it, you've got to reconcile it, you've got to have physical ways to go and store that as well. You're going to need databases that maybe didn't exist a couple of years ago, but I would always come back to the point I think S raised, is that you've got to really ask yourself what it is that you're trying to achieve. Don't just go down the route of bashing your head against a bunch of data. Understand what it is that it can help with the business, how its it going to help your people better within the business? How is it going to accelerate decision making? How is it going to inform, better decisions, better quality decisions, but also faster decisions? 

I think if I had to sum up, one issue that I'm coming across when people talk about data is that they can get to understand what the data is telling them but it's taken them so long that, by the time they've done it, the planet's gone round the sun another time and everything has changed. 

I'm seeing things that some companies are doing really well. I'm seeing a trend where a lot of companies will try to do a lot of this internally, so they get on board data scientists. That's great because it's very sensitive information that we're dealing with. I mean, to give you a feel, we've just entered into a proof with one of the world's largest mobile electricity manufacturers. They go and make kit that goes and charges up rock concert stadiums and stuff like that. The data we were getting was so sensitive and so voluminous that just the NDA took three and a half months to work out. This is the complexity when even just trying to understand how you mobilise data from one company to an external body takes so long. You've got a very complicated situation internally.

There's a lot of vendors out there. A lot of these vendors have got great experience. Do engage with external companies to go and have a look at this. As I said, I'd stick to S’s point at the beginning of all of this. What are you actually trying to achieve? Because you can spend an awful lot of time just going down rabbit holes, frankly. If you understand what it is you're after, streamline the data as much as possible, it is there to be done. I'd start off with the why and the people. That would always be my advice. 

JP

Thank you. 

Alan

No worries. As I said, I'd love to have fantastic answers. We've had great outcomes with companies and we can share these with you. I'm very much, as Kane was saying earlier, very much in listening mode to what you're sharing. So it's been fascinating. So thank you. 

JP

Thank you guys. I'm going to take a step back, really, and invite you to ask questions of each other or to pick up on any of the points that we've covered so far. I'm conscious that I haven't managed to involve everybody in the discussion yet, so please feel free to jump in either raising your hands or just unmuting and getting involved. 

F

I just wanted to add a bit more. We've been looking at our distribution operation specifically rather than the whole of the supply chain. We've been banding the VUCA word round quite a bit and I think everyone understands the concept, but we're trying to take it a step further and saying we need to move from measurements and things. That is based on business as usual and people giving feedback as to why we didn't perform in a particular period of time because of some external influence to saying, okay, well, how can we recover more quickly from those external influences? Asking the people in the distribution operations to say, well, what are the types of things that could happen and what would you need to do to be more resilient to those external influences? Of course, you get into the cost implications of having redundant equipment or redundant people. 

We're trying to look at a framework that we could use and see how well that worked to come up with ideas that we could actually put into practice because as everyone said, there's been lots of external factors over the years. They've all been different recently. You look ahead, how can you plan to be more resilient to those impacts and effectively bounce back quicker when you get those influences? We know that we don't know what they're going to be, but we know the type of things that can impact us and therefore what are the type of things that we need to do to be more resilient in the distribution operations. 

Alan

How easy do you find it within the business to map out possible scenarios, to create scenarios, populate them and start extrapolating into potential futures? How easy do you find that? 

F

I don't think it's too difficult for people to come up with things that they know have happened or could happen based on events in other parts of the world or in other organisations. I think the challenge is more what do you do and how do you rank those and decide whether you should invest in any capital or resources or perhaps training to make your workforce more flexible or whatever. I think that was more the challenge. You can come up with loads of different things that might happen, but it's more how do you rank and prioritise the activities to make you more resilient going forwards. 

JP

Thanks. F. G, you’re next? 

G

Yeah, I work within electronics to durable consumer goods. I actually like the word ‘shocks’. We talk a lot about disruptions, but even some technical trends can create disruptions and I think those are easier to identify than these shocks like we had on the pandemic or the geographical ones or these vessels which get stuck. So can we actually predict them? A big question for us if we talk about resilience. Yes, there are certain scenarios which we can probably think of, identify them and get from the crowd knowledge, hey, what could go on? The number one which comes to our mind is scenarios. Some can come from people's past events. There are certain patterns which you can take data in and create, say, hey, there are patterns around those events. 

That's number one. What you want to create are scenarios. That's number one which we are, for example, trying to establish. The other one, which I found very interesting with one of our consultancies working together, one is really okay to a certain extent you can't think of events, but there are probably a lot of events you can't even think of. So what can you do? Really what they say is working with digital twins, so you have your network but digitally represented and they say you actually should be going for the weakest links you're having. Look at your network, it doesn't really matter if it's a volcano or if it's an airline on strike or if it's whatever, try to just cut off one of the links and see how it works in the network. I found this a very interesting way of going and finding the weak points in your network. 

See with the digital twin how this then works out and see what the impact is and then address those. I just wanted to bring this in here in this group as one of suggestions, which actually was not so obvious to me, but now getting into this is actually maybe a good way to do it. Digital twins on the other hand, it's difficult to get set them up, especially in a global company. I think one of you guys as well said, hey, we need to present it then to whatever management or senior management. Yes. It's a question about how can I present it. Our way forward is a bit more going into these aspects of scenarios and digital twins to present the implications. 

JP

Thanks, G. B. 

B

One of the real consequences of the disruption we have had over the last years is there's been a senior leadership focus on revenue protection against the type of issues that we've had. So, whereas before, I think one of the other contributors talked about business continuity plans being on paper and now needing to be in effect, that certainly applies for our A class products. At the top 30, 40% of our revenue, we've certainly seen that those plans need to not only be on paper, but be demonstrated. If you've got dual sourcing of a product between two manufacturing sites, the expectation is that not only can they do it, but they have done it. They've got the raw material sitting there, they've got the expertise sitting there. They're probably making a selection of products out of the portfolio of products throughout the year. So those skills stay there. 

Whereas previously the fact that you had it on paper, you had the skills, you had an S&OP somewhere would have been good enough. Now it's got to be proven that capacity is there, that expertise is there. Certainly to protect the revenue of those major products through dual sourcing. It's a big thing. 

I think the other thing we've really seen again for the top products is a relaxing of inventory expectations. Previously, we were really managing inventory levels down hard and for your largest sellers, your most stable products, usually you can push your safety stop pretty low because you're making them all the time. We've seen a reversal of that, where we're increasing the inventory coverage of those real high revenue SKUs and as a consequence, taking greater risk with some of the C class products, where we'll accept poorer customer service. We'll accept a missed order in order to create the space and the inventory capacity to look after those top SKUs. 

JP

Thank you. H? 

H

I just wanted to check an idea out. Prior to my stint in Sushi for two years, I worked in food logistics for 18 years. Were supplying people like XXX and XXX, their consolidated supply chain and what we saw through a year, notwithstanding the odd spike, but you had a fairly predictable seasonal pattern. What I did in my latter term there, so the last five years we took a decision that we would say no to anything above 75% of the maximum expected in the year of demand. My point is perhaps there is room for saying no? I'm assuming by wanting to get 100% of what's out there, it's about maximising how much we as suppliers supply. Whereas actually if you say okay, we're going to only resource up to hit 75%, where let's say that your modulation through the year is such that it moves between 100% being the absolute spike and 75% the lowest point of the trough. 

If you shoot for a lower point or the midpoint between that modulation, you're going to get better utilisation of your assets and all that stuff. The peaks and the troughs you give to agencies, you let other people worry about that. There are people, there are businesses certainly in the logistics world who thrive on that. I don't know how appropriate that is for all you here, but just rather than going okay, I want everything and therefore I need to have strategies that allow me to cope with it, just say well no, I'm going to manage it for this amount and ignore the rest, ignore the shocks. 

Kane

I actually just comment, I've heard of that several times, H, being used in the past by different people. Like we'll take up to 75, 80% as our norm. Some people go a lot lower and then hand it out to 3PLs after that or another supplier base to deal with it and spread the risk almost. So you're not alone in that. Demand planning around that, logistics planning around that. There are tools developed to do that for you as well. Now the AI can do it with you. That's not unheard of. 

JP

Thanks for that, H. K, please. 

K

I just wanted to come back on digital twins. We've been using digital twins now for a few years. I think we started about three years ago and the data was a challenge at first, but then we've learned how. We've developed a methodology for pipelining data from our transactional systems to help us build those quite quickly and keep them up to date. Our whole risk and resiliency strategy for the last couple of years has been around building digital twins to identify where the biggest resiliency issues are in the supply chain, where our vulnerabilities are. We've started to build options modeling on top. One of the things we found is that we are spending all over the supply chain to reduce risk, but sometimes kind of over investing in other areas and dramatically under investing in other areas as well, where our biggest vulnerabilities are. Now we've developed this options modeling tool. 

For example, you'd identify that maybe products thrown into a particular country or a particular supplier might be the biggest vulnerability and dual sourcing, even though there's a cost associated with it, is your best way to improve or reduce your risk for that particular supply chain. We're using digital twins now for business continuity planning, for risk assessments, desktop risk assessments, but we've also started to use it now for our production system. We're looking at value stream mapping and looking where the efficiencies can be gained and improvements can be made in the supply chain. What we haven't done is connect it to an operational business process yet. They're all strategic processes and that's really the next step to figure out how to bring this simulation capability into more day to day processes. But we haven't started that yet. 

JP

Thank you K. 

Alan

Sorry K, if I can just jump in quickly there. How did you go around setting up these digital twins? Did you bring in a consultancy or was this purely in house talent that you called on? 

K

It was in-house talent. My previous role, I was leading digital transformation for the Pharma Division. When I went into the role they demonstrated, they built digital twins of our bioreactors, which could predict yields, cycle times, et cetera. I asked, well, could we do this for an end to end supply chain? One of the data scientists set off and within like a couple of weeks came back with an initial proposal. When we just started doing test and learns and trying it out and then like the end to end supply chain community got sight of it and took hold of it and it's really flourished since then. It's built on a platform called XXX, which is a kind of off the shelf modular simulation tool. It's really easy to drop out a node, or add in a node, add in an extra lane and just see what the effect is. 

Alan

Cool. 

JP

Thanks, K. Just conscious of the time, everybody, we're into the last five or six minutes, so please jump in if you have any final questions. We got time for a couple more. 

C

Sorry I wasn't able to follow the complete discussion. I think it was quite interesting and I kind of got sparked by K's comments on the digital twin. I had some questions there, how that was managed successfully. Because I do see some challenges on that side. Within our business, one part is data, but on the other side it's also, I think, people availability and people's acceptance of working with digital twins. I'm a bit curious on the second because working on data is one thing, but working on the second part is something completely different. That's the challenge that I'm foreseeing at the moment, mainly because we have also been working on data and data accuracy, but getting people to work with a digital twin, accepting that digital twin, that its output is, let's say, at an appropriate quality, how did you get that? 

What I understood is that you started with something small and then basically because you already had some momentum, it was accepted or yeah. How did that happen? Because I just see some challenges with us where even without a digital twin, with almost real time data, our operators are challenging what they see with automated outputs. 

K

Yeah, so I think the game changer for us was handing it over to the business. I was in like a digital team leading the digital transformation program, and we'd proven it so many times, but you don't get that adoption. Handing it over to our value chain management organisation and letting them prioritise be part of the development process so they kind of see where the kind of accuracies or maybe weaknesses might be in the model. It became like a business led program and that drove adoption and understanding of where it can be applied. Where it can't be applied because it's not one size fits all. It doesn't do everything you need. 

JP

Just say at this point that I will connect everybody in by email afterwards. If there are kind of individual strands of conversation you'd like to pick up offline, then you'll be able to do so. We're covering a lot of ground. We've got one end, people who are still largely using Excel to manipulate forecasts and to try and get a better handle on plans, and the other end all the way up to digital twins. I guess just to give the final word to Alan and Kane. It's not an easy one to answer, I suppose, but what do you see as the best practice in terms of sequencing that journey? What are the next steps? If you're at the earlier stages of that and taking those building blocks towards getting a better handle on volatility and getting more predictability into your planning processes? 

Alan

You like to ask a good question and give three minutes, don't you?!

For me, it's always just thinking about what you're trying to do. I think K's made a superb contribution at the end. Start small. If you try to boil the ocean, you're just going to go, it's too much. As with all things, I'm fortunate. We are a technology provider. We all know that the change management that C has touched on here, that inertia and resistance that naturally exists within people and businesses is something to bear in mind. I think if you can verbalise that strategy clearly to the troops, start small, show wins. For me, that feels like how you can start getting people on board. It's a bit stating the obvious, but from experience, that tends to be the best route. 

Kane

I'd echo that for sure. Also I ask you all to consider that I'm a consultant by trade. I work for Tagetik to advise businesses on data integration and how to move tools. The key thing is don’t ignore… I know we haven't touched it, but it's the ESG journey. You're collecting this data and moving it forward, you're not going to do this within Excel, and the actual single point of truth, when you're coming along to move all of your data, that's got to supply some form of carbon output. You are going to need tools to do that. Start to look at what you want to achieve and how you're going to achieve it, and then go try and maybe find a vendor that could take you on that journey and expand with you. I don't believe Excel has got everything it needs. 

JP

Great. Thank you, Kane, and thank you, everybody. We have to leave it there, I'm afraid, but like I said, we will put a write up together for this, send it to you and connect you via email so that hopefully you can carry on any discussions where you'd like to. So, finally, thanks once again and have a good rest of the day. Bye.