Perfect data doesn't exist in any You do create as many pockets of data as you can consolidate to start solving for the next higher order problems and then keep incrementally working on improving your data in your organisation. Eventually you'll get to a shape where you'll feel like, ah, 70, 80 percent of the data is okay. The remaining 20%, however hard you try, you're probably going to still have a struggle. The juice might not be worth the squeeze Good morning, good afternoon, or good evening, wherever you are in the world. Welcome to episode 55 of the Sustainable Supply Chain Podcast, the number one podcast focusing exclusively on the intersection of sustainability and supply chains. I'm your host, Tom Raftery, and I'm thrilled to have you here. A huge thank you goes out to this podcast's amazing supporters. You make this show entirely possible. If you'd like to join the community, support starts at just three euros or dollars a month, which is less than the cost of a cup of coffee. And you'll find the link in the show notes of this episode or any episode or at tinyurl. com slash SSC pod. Today, I'm excited to be talking to Sunder Balakrishnan from LatentView. And in upcoming episodes, I'll be talking to Juan Maicel, who's the CEO of Grip Shipping. Shannon Payne, who's SVP supply chain at MDSI. Mickey Vandeleu, who's the founder of Lakeview Consulting. And Klaus Bretschneider from Lynx AS. So some excellent shows coming up. Don't touch that dial. Stay on here for those episodes as they show up. But back to today's episode. And as I said, today, I'm talking to Sunder. Sunder, welcome to the podcast. Would you like to introduce yourself? Thank you so much, Tom. Yes, absolutely. It'd be my pleasure. Hi, this is Sunder Balakrishnan here. I love thinking and building in supply chain and building towards sustainable supply chains. I had the Supply Chain Analytics practice at this public listed company called Latent View Analytics. We sleep, breathe, eat data, and that's sort of what we do. And in that I specialize in supply chain. I have the the benefit of working across four different industry verticals, consumer packaged goods, retail. industrial automotive and technology. All of them rife with use cases for sustainability. So very excited to be here, Tom. Thank you so much for the welcome you Sunder and so tell me a little bit more about latent view and those particular industry verticals you mentioned what kind of problems are you solving for your clients in those verticals. Latent View as an organisation, came into being in 2006 through the vision of two of our founders who worked in the industry for the while, but then figured that in the early parts of the 2000s, analytics was really starting to bloom in multiple industries. Since they saw their opportunity, they got in and we got off the ground. 2009 landed in the US which now is one of our biggest markets. And some of the big tech clients are some of our biggest clients. As I said, we operate across five different industry verticals, four of

consumer packaged goods, retail, industrial automotive, technology, which is one of our biggest verticals and financial services. That's the fifth industry verticals. If you wanted to sort of get a sense for the cross section of the problems that we look to solve, right from anything to do with customer journeys and understanding customer journeys with data to campaign effectiveness of your marketing effectiveness, if you will, to personalization, all in the customer and marketing sphere, to in the supply chain sphere, a lot around manufacturing excellence, logistics, delivery, efficiency, on shelf availability, a favorite theme of mine. And then of course, underlying theme around that being sustainability to also working on data analytic problems within financial services or FB and a. HR analytics and so on across these industry verticals are just a cross section of services and solutions that we that we built over time. Okay, and your own journey to where you are today. How did you get to where you are today? I started out straight out of engineering in software services, spent time writing code. Then I went on to do my management education in general management with a whole mix of subjects that helped me build up a lot of the baseline, if you will. I then had the opportunity to work on the floor in retail. Sort of big boxes, put it on shelves, sell the product, inward product that is coming in from the suppliers and so on. And then I got a break in consulting with Infosys, spent five years over there, a lot of data driven operations consulting. Then followed it up with PwC PricewaterhouseCoopers. Again, the similar theme of data driven operations. I then had an opportunity to work for a boutique analytics company before I then got the big break at LatentView. And with LatentView I had the opportunity to actually set up the whole supply chain practice, the thinking around what is it that we are even going to offer in the market. One of the themes we talk about is this connected supply chain. The argument being a lot of tech in supply chain means that supply chains are integrated, but the fact that a ship turns sideways in the Sears Canal and it causes a supply issue in Mexico, this can only and only be built through the association data points give you and that is the essence of a connected supply chain. That's what we are building and thinking every day and, consulting our customers as well. Okay, and what has surprised you most on your journey from leaving university to where you were today? In university, when you study concepts, you talk a lot about strategy, talk a lot about direction. I think when you get your feet on the ground and start executing, you realize that there is so much of a delta that. And all that learning can only and only come when you're working with teams, when you're trying to execute different concepts. We were in a client presentation on Friday, we were building some demand forecasts and showcasing to customers as against what we designed six weeks back to what we were presenting yesterday, there was a lot of delta because reality start hitting you. Data start hitting you. People realities also start hitting you. Process reality start hitting you. And how do you tune yourself to adjust to those realities? Because your tool is only as good as the people who can use it and make good decisions out of it. So those are some of the learnings from university to, execution that I've learned over time. And we're on the Sustainable Supply Chain podcast under sustainability is often looked at as a cost or a compliance issue. How do you see data and AI reframing it as a value driver, or do you? Absolutely do. Yes. Interestingly, in the last six years or so that I've been leading supply chain practices and having conversations with various customers, you're absolutely right that when you think of design principles of supply chain, the way I like to express it to most customers also is supply chains traditionally have been designed for a cost min way of working, as in you always look to minimize cost with tolerable service levels. So whatever service you can achieve within that min cost is what you try to get to. And then the afterthought of a filter is, okay, let me see what's the sustainability score I got out of doing this. I've been trying to talk to customers about this almost triangle way of thinking about the supply chain, the three vertices being service, cost and sustainability. As you keep tuning your supply chain, as you keep taking different actions, there's almost like an operating zone, right? You could pick a spot in this triangle and every vertex implies a certain metric measure. Now, I've started getting into metric measures and the application of algorithms and AI and ML to actually help you orchestrate your supply chain in a certain way. What I mean by that is, let's say a particular transport from point A to point B costs you $100 and it's going to give you an a service of 90%, let's say. Another gives you $120, but it's going to give you a service of. 96% percent your cost min will probably say don't take the $120, take the $100 because It's $100. Like It's a minimum cost. Whereas the second one, the alternate way of thinking of it would be that yeah, I'm going to get six 7 percent additional service by doing$120 if the 7 percent justifies the additional $20 and why not do it? But I think the third parameter here is in trying to cover that distance, are you also having an incremental lift in your carbon footprint, say, or in, in wastage measures, et cetera, Adding that as a parameter in this equation is where the influence of algorithms and AI can come into designing sustainable supply chains is at least my belief. Okay. I mean, that makes sense, but can you walk us through any real world examples of how AI has supported sustainability goals and supply chains? Absolutely. can think of about three different examples and I'm mentioning these three because, I would encourage anyone who's thinking of AI analytics and supply chain to again, think of it in terms of a maturity curve. At the absolute base of it at the stage one, if you will, is the foundational elements of data in sustainability. So you have to build your baseline data foundation. Examples is a time manufacturer has come to us and said, Hey, I need to do a whole lot of compliance reporting around sustainability. I have all this data, largely supply chain data, but a whole lot of other financial data also spread across multiple systems. Can you help me connect with all the data, ingest it, engineer it, shape it, so that it can facilitate my reporting. It can also facilitate some of the internal dashboards that are used. So this is sort of your foundational level of analytics. Silence. Silence. Silence. Silence. You start taking it a step further. You can come into what we call causal and predictive analytics. So there's an organisation who manufactures chemicals and functional farms. Their problem statement was, said, Hey, run a factory, the consequence, natural consequence of running this factory is that there is a lot of waste water that gets generated. It has to be released into water bodies. There are permissible limits. Anything about this, I have a fine attached to it. right. Now, the, initial problem statement was, can you tell me when I'm about to breach the permissible limit? We started having further conversations with them. What we realized we could solve for is By the time you hit the permissible limit, it's already too late. It's already, you are in damage control by then, but we use the power of analytics AI to actually start detecting patterns, anomalies, and tell you when you're breaching your control limit, meaning you go beyond this, you have not yet hit the fine level, but you are still in trouble. And if I can detect it enough, I can give you a signal back enough for you to then start adding control measures so that you don't exceed your pollution ever, but also you operate well within operating conditions. This is more of running predictions, predictive algorithms, a similar thing with a food waste for a food and beverage company where we did a prediction of how much food is likely to get wasted or stale because of practices in the supply chain? What can you do to arrest some of that? And the final level of maturity is sort of getting more cognitive with a lot of external data as well as an organisation who came to us, an automotive company, and this is where it sort of starts getting into scope three, because what they said is I have a lot of shipments that are flying around different parts of the world, and they are being flown around by a third party, a 3PL logistics provider. Is there a way for me to understand the traceability of every package of mine, so that I know that what's the total carbon footprint, accumulation that is happening because of the transit my packages are going through? And what can I do to work with them to consult with them to say, Why don't you think of an alternate route or why don't we think together about an alternate route? All of these are practical problem statements that have come our way and we've found different ways to try and solve them or find data solutions to And what about the likes of digital twins, you know, can they be leveraged for sustainability and do you see their potential? hasn't been realized yet anywhere? I think digital twins in itself are still fairly nascent in the industry. there is still a lot of proliferation to be done, but I think the power and maybe for the audiences who may be hearing this term for the first time, a digital twin is meant to do two things and two things very well. First, create a digital replica of any physical system. So it can be a manufacturing line, a manufacturing plant, a warehouse, a distribution center, or it can also be a replica of your entire network. That is all the connected parts of your network from manufacturing to the the truck that is transporting to warehouse, the truck transporting to DC and so on. When you, the first part is to create this twin so that you have visibility to everything. I'll also add one element to this is a lot of digital twins in the market tend to construct what I like to think of as a conceptual twin in that you have a concept in mind of how your supply chain is supposedly structured. And then you almost create a simulation or a replica based on a conceptual diagram. The better way to do it is take actual data from your systems, learn the actual relationships between different sites, and then construct it. But the second part of a twin which is important is simulation. What if scenario planning, and that's where you really start getting to say for the inventory waste problem that I spoke about damages and stairs. If I stacked a pallet with 56 boxes versus 66 boxes in a pallet, how much is it going to make a dent on or a reduction in my waste is the kind of simulation problem you could solve for if you started constructing digital twins. They are more visual as well. So therefore they are more intuitive. The forklift operator on the ground could actually maybe do a quick simulation. It's like the flight simulation type of a problem that they could see that if I'm going into lift, pallets in a certain way, it was a certain other way, Um, Or if my forklift is traveling around my warehouse in different routes, is it consuming lesser fuel and therefore lesser carbon footprint? These are the kinds of things you could construct and simulate with a digital tool. That's the core application. And you talk about data driven sustainability as well, and you know, that sounds great, but where do most companies fall short when they're turning data into actionable insights? This is as true for sustainability as it is for other places in supply chain efficiency visibility as it were. But generally when I've done mapping of that maturity that I spoke about earlier, for organisations that are roughly about 2, 3 billion dollars and above in size foundational data is usually a check. Some level of dashboards for descriptive and diagnostic insights, usually a check. Almost think of these as greens. You come to the next layer of maturity, causal insights. Why did certain phenomena happen? If there was a delay in delivery, why did that delay happen? Or if for a particular lane, my carbon footprint spiked, why did that happen? Other kinds of questions you want to answer. I'd say it's an amber. Kind of. And then you start going into prediction and prescription, the next two levels. For a lot of organisations, it's kind of a cross, a red cross in both of these. That's sort of how I imagine it visually. I don't mean to say that every part of an organisation is a red cross. In some pockets of your organisation, you would have experimented with different kinds of predictions, different kinds of cognitive, insights. if you had to take like an average of most organisations of that size, it's generally a struggle to have causal, predictive, prescriptive. As you start going to slightly smaller organisations, and when I say smaller, between 500 million to a billion dollars, a billion to a couple of billion dollars, even the foundational level data and basic diagnostic insights for compliance reporting, that also starts becoming a bit of an amber. So, in that data driven sustainability and the maturity curve, this has been my learning of where organisations are. If someone were to ask me, where should I start? Um, Thank I'd always say, first principles is get your data foundations in place first. But having also said that, you can never almost approach this as, let me try and, you know, shape up all the data in my organisation first, and then I'll go to the next stage. Because that, that perfect data doesn't exist in any You do create as many pockets of data as you can consolidate to start solving for the next higher order problems and then keep incrementally working on improving your data in your organisation. Eventually you'll get to a shape where you'll feel like, ah, 70, 80 percent of the data is okay. The remaining 20%, however hard you try, you're probably going to still have a struggle. The juice might not be worth the squeeze, so you'll probably have to take more and you'll continue. But I think that's the way to think of problem solving with data driven sustainability. What about things like on shelf availability that you mentioned earlier and end to end connected planning? They, sound great, but how do they translate into real world wins for companies and consumers? Oh, consumers as well. that's, that's very interesting. I, spoke about this concept of a connected supply chain. I think everything is connected and when you start solving for on shelf availability, you essentially start solving for that consumer and a consumer centric supply chain in that everything in your supply chain is moving and shaking to try and make sure that the right product reaches the right customer at the right time, place, quantity, quality, price. How do you now go about thinking through this on shelf availability is, a lot of on shelf availability is thought of from the point of, Oh, I need a distribution center that can somehow hustle to send product over to the retail store. So that the product is in front of the customer for them to pick up at the time, at the moment of truth. But I feel like the network has such a cascading effect. This is an example of a six months ago. I was talking to a customer. We're in a workshop in their office, and he said, I have this one warehouse that is backed out four days in the yard. So there are four days worth of trucks sitting in the yard. They just don't have the capacity or whatever to unload. And then the plant is keeping on sending more trucks their way. I mean, you think of it any which way, cost, service or sustainability. This kills all three metrics. There was nothing in the intelligence of the system to say that, you know what? I'm actually four days behind at that warehouse. Can we not reroute? Can we not a) go back to the end to end planning itself, which is sending the trucks. Or B), the end to end visibility, which is looking at a truck going in, which can see four days of truck. Can it not redirect to a nearby site if it has the capacity or whichever site has the best capacity available? Fortunately, or unfortunately, a lot of these decisions, even in large corporations are largely heuristics, human led. Ultimately, a human has to make decisions without a doubt, but there has to be a machine that has to enable that human. And the idea of end to end planning plus on shelf availability is just these two problems that I described starting to work in a bit more synchronisation in that if there is an event in the network that is causing a spike in any of these three metrics, one, can there be a fix right now through, you know, alternate decisions and two, can there be a feedback to the end to end planning to say hey, there is this particular issue. It seems like it'll be a long tail to the issue. Why don't we recalibrate our planning to think of it alternately? Both of these need a bit of a dynamism in the supply chain. We've been a little too static for a long time in the way our processes, systems, and people have operated. And that I think is a key need of the hour. The next time, I hope the next time of COVID doesn't happen in our lifetime or any of our future lifetimes, but any such red herring event happens, we talk about resilience. The resilience of a supply chain is not this big chest beating initiative or transformation. The resilience of supply chains are built in these small events and the ability to tackle these small events so that your deviations are never too far out. You're always staying stable and in control. That's the, that's my view of end to end planning plus on shelf availability operating towards sustainable supply chains. To kind of achieve that level of resilience and agility, you obviously need to have AI built into your supply chain, but very feorganisationsns seem to have achieved that. Why do you think that is? You know, what kind of barriers are they facing trying to integrate AI into their supply chains and how can they overcome them? For a lot of organisations, when I sort of start assessing where they are at in their maturity curve, I think some of the foundational elements of the data are things that they need to be painstakingly built. That's the part that's never going to be easy. Building AI algorithms is not as complicated if you have good data, but building good data is just painstaking. I think that's stage one. Stage two is again, you have a bunch of systems. You, you, sort of buy different systems and then figure out how to integrate them, how to get them in, that generally becomes one of the second stage gates, if you will. But I think the, the bigger mountains to climb, are people, change management. It's one thing to build a box and say here is an AI. It's going to spit some results out. Go follow what it does, but it's another to the human who's going to receive these numbers Naturally, and it is something we speak about internally have spoken in client presentations also is Why should a human want to trust a box that is putting some numbers versus something that they have seen over time, believed over time. They believe they know better than a box does educating them around that. That the idea is the box is not there to replace them. The box is there to augment their capability. Secondly, that box should also be able to take feedback from the human, if the human in the loop says, I think I have a better alternative to what you're suggesting. Let me try what I am doing. The feedback to you is no, I'm not going to do what you're going to do. Let the machine observe what the human is doing and learn from the fact that if the human took an action and it was a better outcome than what the box suggested, take that input back into your feedback loop. These are elements that are still maturing over time. I think, I mean, today, nowadays, no conversation is complete without mentioning generative AI. There is a certain power to generative AI in that it can take feedback from humans. It can generate elements that make the journey, the decision making journey for the human much simpler. I think a mix of these solutions coming together is what will aid, this maturity journey. I think we are some distance still away from that. I keep joking this, joking about this with a lot of my team members that I think there are so many good problems still to solve in supply chain. And I could very well retire and there'll still be a lot many more problems left to be solved. And that's sort of the learning journey that we are all on, right? That, there's a lot of fun problems to solve to keep their door on. Okay. Well, I mean, seeing as you brought it up, you were looking into the future. How do you see supply chains and analytics evolving in the next five, 10 years, particularly in the context of AI and sustainability? There are, of course, a lot of organisations have set themselves goals for 2030, 2040. I think that in the journey to get there, data is going to be at the core of taking smarter decisions towards those goals. Almost want to say that organisations who want to achieve those goals, you can almost split them into data driven sustainability organisations and, data agnostic or data absent sustainability organisations. I think the first category will probably get there faster, will probably get there much more efficiently, much more, intelligently than the second lot who do things based on gut feel. I think that that's one, I'd see that, that the journey there will split into two lines that you might probably start seeing that more and more clearly. Again, within this, the qualifiers that most companies collect data. It now comes down to which of them use it well and which of them don't. There'll be that split. The proliferation of AI use using smarter algorithms, using better decision engines will, I think, play a big role. I expect to see digital twins playing a role. Especially in how you orchestrate responses from your supply chain, how you design interventions in your supply chain towards either better service or better cost or better service or a better sustainability or maybe better still a balance of all three in order to get those realities. I think stronger data, stronger AI applications and digital twins with simulation will come into play. Finally, I think the last element to this is taking your people along in this journey, helping them mature, helping them learn, grow, educating them more about making them data powered sustainability champions, if you will, will again go a long way in helping organisations accelerate their journey. That's my thought. what metrics or KPIs do you think companies should focus on to measure the success of their sustainability initiatives and ensure continuous improvement? I think, naturally there are sustainability goals that have been set up, and generally these sustainability goals have, fairly clear operational metrics around, say, example, food waste, cutting down the volume of food waste, measuring your carbon footprint and looking at optimising carbon footprint, electricity usage, and so on. There are a bunch of operational metrics that you anyway, have a measure on. You can very naturally start measuring how much of an improvement are you making quarter on quarter, year on year and so on for these operational metrics. I think what will also be interesting for organisations, and I'm hoping organisations are already doing this fairly well, is you have a, almost a project management plan from now up to milestones to your 2030 or 2040 journey, and those have specific governance metrics. Those have milestone achievement metrics versus did not achieve metrics. And it's not just the chief sustainability officer's job to track that. It's now going to be the job of the chief supply chain officer, plus the chief sustainability officer, plus the CFO as well. To track. Are you even hitting your milestones? Somewhere along the lines, I would like to see organisations start tracking one additional metric around, people skill development, your learning and development skills in your people. Are they getting certified? Are they doing practical projects? Let's say the number of practical projects that they're doing six sigma projects or anything that is targeted towards sustainability. To sustainable development of your organisation. These can be interesting governance metrics over and above your operational metrics, to track how you're evolving as an organisation more than anything else. Very good. Very good. Yep. And left field question. If you could have any person or character alive or dead, real or fictional as a spokesperson for our supply chain sustainability, who would it be and why? Oh, wow. I'm also a football slash soccer fan and just for the reach that two of these players have. I could almost think of a Ronaldo or a Messi and given that a lot of these players already talk about, you know, initiatives with UNICEF or a David Beckham for that matter, UNICEF, et cetera. So somebody like that joining the bandwagon to be able to message it out, the message will reach so many millions. But if you wanted some character from history. I am very fascinated by Abraham Lincoln and the power of his oration, the pull he had. Somebody like that can also be very interesting as a spokesperson for driving sustainability across organisations, it can be quite interesting. Nice. Nice. Great. We're coming towards the end of the podcast. Now, Sunder, is there any question I didn't ask that you wish I had, or any aspect of this we haven't covered that you think it's important for people to think about? I think the last element of this, and I have a five year old daughter, and when I think of very simple things like wake up in the morning, brush your teeth, don't keep the tabp on while you're brushing. it. You could keep the tap off. Some simple things like that. We see if we can try our best to practice some of those things. I can't say that we are super successful. We try our best, but education at grassroots levels for I mean, why should sustainability education start after you finished your school, your college, your, post graduation, and then you actually start, and at least that's been the reality for someone of my generation, let's say that the realisation of sustainability has been the last 10 years, maybe, that's when it started to become mainstream as a conversation. But my daughter, who is now going to go into her first grade next year. Shouldn't this conversation start in her classroom in the first grade itself? Why not? I think that is one thought that keeps coming to mind, that should we start creating more responsible citizens by bringing some of these themes right at that grassroots level? Maybe that's the one immediate thought that comes to mind Nice. Nice. Great idea. Yeah. Superb. Sunder, if people would like to know more about yourself or any of the things we discussed on the podcast today, where would you have me direct them? You could, of course, reach out on LinkedIn. You'd find me by the name Sunder, S U N D E R. Balakrishnan. You can also visit the website of my organisation, www. latentview. com. And you'll find a lot of information about what we do across different spaces. Don't feel shy to get in touch more than happy to have great conversations. Fantastic. Sunder, that's been really interesting. Thanks a million for coming on the podcast today. Absolutely. Thank you so much, Tom. I appreciate you giving me this opportunity. Okay. Thank you all for tuning into this episode of the Sustainable Supply Chain Podcast with me, Tom Raftery. Each week, thousands of supply chain professionals listen to this show. If you or your organization want to connect with this dedicated audience, consider becoming a sponsor. You can opt for exclusive episode branding where you choose the guests or a personalized 30 second ad roll. It's a unique opportunity to reach industry experts and influencers. For more details, hit me up on Twitter or LinkedIn, or drop me an email to tomraftery at outlook. com. Together, let's shape the future of sustainable supply chains. Thanks. Catch you all next time.