Tom Moore: 0:03

Shipping to a warehouse in, in a customer location and shipping them 15 loads in a day, or five loads in a day, that, that doesn't help them in the slightest. Okay. And, and actually one of the, problems that needs to be solved and, and we are thinking about how to solve that right now is, how does a big purchasing organisation buy in such a way that they know that when that truck actually arrives, there'll be a dock door for it?

Tom Raftery: 0:29

Good morning, good afternoon, or good evening, wherever you are in the world. This is the Digital Supply Chain Podcast, the number one podcast focusing on the digitisation of supply chain, and I'm your host, Tom Raftery. Hi, everyone, and welcome to episode 350, 350, wow, amazing. Anyway, episode 350 of the Digital Supply Chain podcast. My name is Tom Raftery, and I'm excited to be here with you today, sharing the latest insights and trends in supply chain. Before we kick off today's show, I want to take a moment to express my gratitude to all of our amazing supporters. We'd never have got to 350 without you. If you're not already a supporter, I'd like to encourage you to consider joining our community of like minded individuals who are passionate about supply chain. Supporting this podcast is easy and affordable with options starting as low as just three euros or dollars a month. That's less than the cost of a cup of coffee and your support will make a huge difference in keeping this show going strong. To become a supporter, simply click on the support link in the show notes of this or any episode or visit tinyurl. com slash dscpod. Now, without further ado, I'd like to introduce my special guest today, Tom. Another Tom. Tom, welcome to the podcast. Would you like to introduce yourself?

Tom Moore: 1:53

Thank you. I'm, I'm Tom Moore. A native of New Zealand, but living in the United States. I'm a founding principle of three software companies that do optimisation for large companies.

Tom Raftery: 2:09

Okay. Can you tell me a little bit more about that Tom? What are the three companies and what kind of optimisation? Because it's a very broad term.

Tom Moore: 2:21

Understood. We started out in 1992 working for Procter and Gamble. And they had a problem that was how do we pick cases and load trucks in a way that is gonna be damage free, legal, and, and most importantly, highly productive. And, and we developed technology on this way and, and also expanded that into load optimisation, which is saying, what's the minimum number of trucks we can get on each load. That it's a highly mathematically difficult problem. And, and so, that was our entree, if you will, into optimisation in, in the supply chain. Subsequent to that, we went in and developed a, a system that actually sits on top of the warehouse management system and tells the WMS what to do. You know, warehouse management systems are, are really good at, at knowing where the inventory is and, and assigning a, a task to a user. And, but, but they're not very good at saying, we're gonna unload this truck over here and we're gonna load this at three o'clock. And, and we're gonna make all those connections between receiving, picking and loading. We've developed the technology that makes those decisions for it, so it's really the brain for the warehouse management system. And lastly, the, the third company that we've developed is the bridge between planning and execution. If you think about the plans that come out of a supply planning system they react to every little change in customer demand, and they do in such a way that they generate I'll call it a saw tooth of really a, a, difficult pattern of shipments. One day you can get 24 shipments, and the next day you can get three shipments doing deployment on a particular lane. And, and that's not a made up example. That's a very real example from, from one of our major clients. How do manage transportation? How do you manage the warehouse? How do you do all those things on that environment? And the Level Load product is designed to smooth that out, looking forward and, and saying, Hey, we've got a a, a big peak coming up. Let's pull that some of that forward. Let's push some of the things that we don't need quite so urgently back and, and let's make sure that we control in such a way that not only are we meeting customer service goals, but we're also reducing the variability in the supply chain.

Tom Raftery: 4:59

Okay. I mean, let's, let's take a step back. You've got these three companies. You still haven't mentioned their names. So,

Tom Moore: 5:07

Well, Well I'm to non-commercial, but the first company which does the case picking and truck loading is Transportation Warehouse Optimisation. The second company, which is the brain for the warehouse is called Auto Scheduler.AI. And the third is Provision AI which is the company that does the level loading. So yeah, we've, we've developed technology and, and again I want to just make this as, as educational as possible and, and I, I really feel badly when somebody is really aah, we're gonna push our own software all the way through, right?

Tom Raftery: 5:43

Good, good, good, so tell me how this all started. What made you wake up one morning and say to my, to say to yourself, I should start three companies optimising, for other companies.

Tom Moore: 5:58

You know, it, it, it, it's not, it obviously doesn't work that way. And, and what happens is a company comes to you and says, I've got a problem. And, and you listen to that problem and you go, Gee, I wonder if there's a solution to that. You know, I, I'm a mathematician by training in, in business mathematics operations research in particular. And, and having worked in that area, you sort of look at problems in a somewhat different and more quantitative way. And, being able to, to take a client's need and translate it into something that solves that need, that's the ultimate way of, of working. But, one of the big challenges is a lot of this is, is extremely difficult from a mathematics perspective. Now this, this whole concept of case picking and truck loading in itself from a mathematical perspective is, is a complex combinatorial problem. And, you know, to solve it truly would take a lot of computers and a, a lot of time to do. So, the question is how you do it quickly? Okay. Similarly uh, when you're looking at the warehouse managements and saying, Hey, how, how I control all these different things that are going on? And, and there are hundreds of things happening 'cause there are lots of different people, lots of different actors, et cetera. And, and so the solution to that requires, you know, making sure that when you squeese the balloon over here, that it doesn't pop out somewhere else. So managing across that, that whole network of, activities in a warehouse is itself a, a very complex problem. But relatively I won't say easy to solve, but, but, but mathematically it's easy to put together. And the same's true in the warehouse network where you're thinking about level loading from a plant to a dc. Operationally you can look at at things like linear programming and and manage across the whole network. The problem is that doesn't work very well when it comes to thinking about what's gonna go on trucks, because most of our clients are large CPG companies. They don't ship between their warehouses with anything other than full truckloads. So while one particular product, we might need 50 cases urgently. You know, what else is gonna go on that truck and how does that dilute the urgency of that, that particular truck? So really complex combinatorial problems there is a solution, but it's because people have come to us and we, we've got this problem and how do we solve it? Okay.

Tom Raftery: 8:39

And, I mean, these kind of problems, why are they important? Talk to me about, you know, what are the implications of solving them versus not for companies?

Tom Moore: 8:51

Okay. When we think about case picking, truck loading, and load optimisation, you know, we are shipping today thousands of shipments from plants to to customer facing distribution centers. Now if we can reduce that, that adds value not only to the customer, not only to the environment, but it also adds value to the carriers, et cetera. So when you think about the, that as a, as an opportunity, and, and we are very proud to say that last year we took 88,000 truck journeys off the road. Well, that's a significant contribution, right? That is a significant contribution. We'd like it to be 188,000, but we're, we're slowly growing that. So, so that's, that's the number one benefit of truckload optimisation. You are saving somewhere between 4 and 8% of the, of the trucks simply by being much smarter about how you put things together on a truck in a way that it truly fills it up. Second when you're dealing in the warehouse you know, the problem that people came to us with you know, my first shift, pilot planner, however you want to think about it, He, he's coming to us and saying, you know, you know, I'm gonna do this, this, this, and this. And then when second shift comes in, they go, oh, that's a silly way to do it. And they come back and change it and do something differently. And then third shift they haven't got a clue. So they don't do anything that, that's particularly intelligent, you know? And, and, and I'm, I'm exaggerating here, but it, it is a, it's a challenge because there's no procedures process that actually helps somebody put all the different pieces together. You know, you think of a large warehouse that has maybe a hundred employees on shift. You, you've gotta manage what all those a hundred employees are doing plus automation, plus you the fact that you've got these random variables with trucks coming in late, and other folks showing up early, and all these, all and hundreds of variables in, in, in the process. And the goal is to get things out the door in a way that the customer, and ultimately the consumer gets 'em on time. So, you know, as we go through the three cycle, through the three companies, this loading concept, you know, what's the benefit? Oh, when you are up at that 24, I mentioned before, 24 trucks one day, three the next. First of all, when you're at 24 trucks, you, you are not using the carriers you want to use. You are scraping the bottom of the barrel. Bad words for, carriers. The next day, your contract carrier. He's sitting there going, well, where's the volume, you know? So that's really what it's, it's really making that smoother makes it good for the, for the transportation cost. And that's where the big savings are. The next savings are in, in labor and the warehouse, both the shipping and the receiving warehouse. Third is, is we don't have trailers sitting waiting to be unloaded at the receiving warehouse.'cause if you ship 24 in one day, 24 arrives at the other end as kind of a fixed period of time. And similarly it's good for the carrier. When we do a carrier bid most companies say, you know, I'm gonna give you 250 loads a year, which they probably think is one a day for, for every. In reality, they give you 20 on one day and zero for the next 19 days, and then they'll you 15 and et cetera. You understand the, the, the issues and they build into their costs the fact that there's this huge variability and, and they're gonna be deadheading trailers in, to meet those, those 15 shipments on one day and, and be looking for work on the next. So, again, one of the benefits, you know, the benefits really are, are, are really reducing the costs of transportation, reducing the costs of labor making life more predictable. And, and of course there's the environmental benefits for, you know, less deadhead miles. Much, much better you load fill, therefore less less trucks on the road. So really a win-win, win all the way around.

Tom Raftery: 13:20

Okay. And, You've been in the business, as you said, for many years now. What kinds of changes have you seen, particularly, let's say, in the last five years in those sectors?

Tom Moore: 13:34

You know, I, I think that, there's always been leaders in the business. You uh, we work with a few of them, Procter and Gamble, Nestle, Unilever, you know, folks who, who are prepared to make significant investments and understand the values of, of eliminating waste in, in their supply chains. I think what's really new is smaller companies are starting to look and say, well, hey, we can do this too. So we are dealing with now companies that are, are historically that would never have come to us because they wouldn't have even have thought of some of the technology or the need for such technology. Now that I think they're seeing it. So that's, that's one major improvement I think in, in total supply chain. You know, as, as you think about the, the other factor, and, and I'm, I'm old enough to remember that, you know, many years ago, a lot of companies had these little operations research groups, and I thought I was gonna be part of one of them many years ago. And, and they all kind of said, oh, you know, these, these systems don't work. They get lost or they don't work. You know, they, they come in and they have great promise and they never quite get off the ground and they don't get accepted, et cetera. And therefore that was sort of out of vogue to use systems. If you, if you talked about a solution that was okay if you talked about a system, eh, you know, systems, systems are expensive. They don't work. And I mean, that's whole thing. And, and now you, you see people much more willing to accept, Hey, we can't get this through with, with manpower alone. We have to use some form of intelligence. That intelligence really revolves around systems and, and which flows very nicely into this whole thing with AI. You notice that two of our companies have AI in their name. Okay? Why is that? Well, AI continues to develop, and I'll, I'll just talk about this level loading product. One of the challenges is of course, that you have to, to ship things, as I said before in truckloads between your plants and your distribution centers. You can't afford to ship little bits and pieces because that's not how large CPG companies work. And one of the challenges is that there's no really good way of determining what goes on a truck without spending a lot of time doing extensive combinatorial optimisation, which is what our, our, our product from transportation warehouse optimisation does. Except it takes a lot of time. It takes time. So with AI, we're able to come up with some much better truck loading estimation, not, it's not precise, but it's estimation of what's actually gonna go on each truck. And we were able to, to do this estimation really rapidly. So when we are dealing with a network, we can use a linear program, throw it over to, to AI and, and heuristics they solve, this is what's gonna go on this truck, send it back. And we, we can flip back and forth between the two. You know, five years ago we couldn't have done that. Okay.

Tom Raftery: 16:52

Wow.

Tom Moore: 16:52

To today, we can, and, and we're very fortunate, we received a patent for some of the technology, not wrong words, we can't patent technology for, for the processes that we've developed. So, really you know, again, what's happened in the last five years explosion of AI and, and everybody's talking about it. Some, some actually know what they're talking about and some don't.

Tom Raftery: 17:16

As one of the people who doesn't know about it. Talk to me a bit about it. I mean, obviously, there's huge interest now in the likes of large language models as a result of the launch of chat GPT last November and the various other open source large language models that have sprung up in the meantime. But. You're not talking about large language models, are you? You're talking about another kind of AI, and this is, this is, this is the thing about AI. There are so many different types of AI, everything from Siri on my phone to chat GPT also on my phone, but, and, and, and everything in between. Talk to me a little bit about the differences in, in the AIs and the ones that are particularly useful for what you're doing.

Tom Moore: 18:02

I, I'm not gonna pretend to be an expert in AI, so just stop there. I, I can tell you that we use something called reinforcement learning. Okay? And, and, and reinforcement learning is conceptually very easy operationally and, and you know, whoever came up with this is brilliant. And so I'm not gonna gonna pretend I, I fully understand it, but the, the concept is if you can give some form of reward function and give a, a computer enough time to, to think about how to solve a problem and check itself off of a reward, it learns how the, the best way of doing things.

Tom Raftery: 18:41

It's gamifying AI

Tom Moore: 18:43

is, it is, it is indeed. And, and you know, it takes, you know, hundreds of thousands of iterations and, and hours of computer time to come up with. I, I now know how to do this. Okay. That's brilliant. Okay. Unfortunately, people are using AI as kind of the, I'll call it the panacea. And one of the things that AI, as in its current form, is not particularly good at is chaining events. And, and go back to my warehouse example, you've gotta unload something before you can put it away, before you can pick it, before you can stage it, before you can load it. Those things have to be tightly controlled and, and put together. AI is not particularly good at that. Okay. But go to the other extreme where you're looking and saying, How do I, how do I look at demand patterns and, and say I can forecast what it's going to be over the next weeks based on, on different knowledge. Like, you know, the, the concept of I, if I can see what's going across the scanner in a, you know, large retailer, I can possibly estimate when they're gonna replenish that. Okay. And those are the kinds of things that AI have been, has been well used for and, and that's got a good long history of doing that.

Tom Raftery: 20:04

Cool. The other big trends I think we're seeing in supply chain are things like the rise of automation and IoT and things like that. Have you come across any of those at all or is that more or less outside your remit?

Tom Moore: 21:31

You know, we work with those we, we are not doing the automation. But if you think about again, go back to the warehouse, automation has been, been slowly filtering in, you know, in our world, automatic layer packing, everybody's working automatic layer packing. It makes a lot of sense. Okay. When it comes to, you know, robotic forklifts, again, been around a long time. Moving product around, whether you call 'em AGVs or, or, or self-guided forklifts, you know, they, they work and they work very effectively.

Tom Raftery: 22:07

Sure.

Tom Moore: 22:08

Automated case picking eeh, still the jury is still out on automated case picking because of the, the complexity associated with it. There are some examples of it and it seems to work in some instances. I'm not convinced that it doesn't actually cost more in terms of support and labor today, but automation's there. But you have to think about, you know, activities in, in, in a warehouse in particular. We, we call them, you know, actors, right? An actor could be a person, it could be a automated layer picker, it could be anything. And, and actors have some capability to do particular tasks at a particular rate. And, and from our perspective, we just need to plan around that. One of the challenges with automation is, you know, you can't make those AGV forklifts move any faster. Okay.

Tom Raftery: 22:59

Sure.

Tom Moore: 23:01

They still have all the safety features, which is good in, in place, but, you know, you watch one breakdown in an aisle and then you watch the next one comes along and it sits there and waits for the other one, and then the next one comes along and it waits for the other two and et cetera. and, and, you know, it, it's, it's sort of humorous, keystone cops kind of thing. You know, there, there are still things that need to happen to make automation work more efficiently a and improve warehouse operations. But as we say, the, our auto scheduler product has to look at all those things, understand what the capacities are, and, and, and manage to their capacities. Know you, you, you can't put a, a, a, a eight foot forklift through a four foot door. Right. So, you we, we are still dealing with those, those kinds of, of, of issues.

Tom Raftery: 23:53

Sure, sure. And we looked back over the last five years and saw what changes, or we just talked about what changes have typically happened in the last four or five years. What about the coming four or five years or 10 years if you want to go that far out? You know, where do you see things going?

Tom Moore: 24:13

Well, you five years ago, Gartner predicted there'd be what they call convergence. Okay. And the convergences of the planning systems to the operational systems because we are not linking the whole supply chain yet. Okay. And, and that convergence, you know, the, the level load product is a product that, that drives convergence. We, we are converging from, you know, essentially the manufacturing and, and planning system all the way through to the warehouse. But you know, We are not reaching back yet. And, and this was went back to vendors and we're not pushing back out to, to you know, the many grocery and, and other warehouses that we're shipping to. So we, we've, we've looked at this much of the supply chain. We're gonna eventually get to this, this much of the supply chain. You can't see my hands, so, but you know, we're getting bigger and bigger and, and, and the whole goal here really is, is, is converging all the disparate parts. The big challenge of and this is, this is really hard because you've gotta now start trusting your vendors and you've now gotta start working and, and having your, your customers trust you. And, and that's probably where the biggest gap is from a, from a, from an pure systems and, and processing. Yeah, I think all this is doable. The, the, it has to be a leap of faith and, and I don't see that leap of faith happening today, tomorrow, but, you know, five, 10 years down the road, maybe there'll be a lot more connections. A a among nodes in the supply chain.

Tom Raftery: 25:57

Okay. What are your plans for your companies for the upcoming 3, 4, 5, 10 years?

Tom Moore: 26:03

You know, we are growing in our niches. Okay. We, we, we are definitely niche oriented, niche focused and, and defending our, our turf obviously against others. That, that, that's, that's the, the, the simple thing. Where we see, you know, load optimisation happening is, is not just for the large, you know, Unilever's of the world, but all the way down to, you know, companies that are shipping, you know, 10, 15 trucks a week. You know, down, down at that level. We see it also for not just deployment, but we also see it for, for customer shipments where customers are then have, are now having faith. I was working with a company the other day that ships to Tesco in the UK and they're shipping 46 pallets on a truck that takes 52 might be 48 on 52. go why are you doing that? Well, because we can't fully, accurately predict based on, the various pieces and that are in the load, how we're gonna fit that jigsaw puzzle that's a load together. Well, you know, I think we, there are things that we can do to help that. Okay. So that's one, one area that I think's gonna add distinct value. As we talk about the warehouse it's not just the warehouse, you know, we just announced, this week, Auto Scheduler is now partnered with, with one of the visibility companies. So the visibility company, Fourkites in this case is saying, Hey, this truck's gonna be late, or this truck's gonna be early. Now that adds value to the warehouse operations. And, and, and we can in fact now plan with, with more certainty rather than, we think this truck's gonna be there at, well, 10 o'clock, you know. And, that expansion and, and again, comes back to, to expanding that convergence concept and the whole level loading process, we see that level loading process moving out again upstream and downstream. Customers shipping to a warehouse in, in a customer location and shipping them 15 loads in a day, or five loads in a day, that, that doesn't help them in the slightest. Okay. And, and actually one of the, problems that needs to be solved and, and we are thinking about how to solve that right now is, how does a big, purchasing organisation buy in such a way that they know that when that truck actually arrives, there'll be a dock door for it? Okay,

Tom Raftery: 28:35

Yeah. Yeah.

Tom Moore: 28:37

so we are scratching the surface here. We've got a great opportunity to expand.

Tom Raftery: 28:43

Cool. Cool. We're coming towards the end of the podcast now, Tom. Is there any question I haven't asked that you wish I had or any aspect of this we haven't touched on that you think it is important for people to think about?

Tom Moore: 28:57

You know, that's always the toughest question in the world. Okay. And I, I, I think that that one of the areas that we continue to, to work on is, is, is, hey, what's the customer doing? What's the, plant doing, what's the warehouse doing? We don't spend enough time talking about carriers. Okay. Trucking companies across the world are suffering and, and they're suffering from, driver retention issues, from managing, an extremely volatile business, pricing, certainly in North America goes like this. It is really a challenge to to manage and be in the trucking business. It's a boom bust cycle and, and somehow we need to, to, to work on, on managing trucking in a way that we think about it just the same way as we think about any other resource. There's a, a, a person from Forrester, George Lowry, who really, thinks that, that we should be including carriers basically in, in our supply chain planning and helping them with it. The problem we have certainly is, is in North America, you know, majority of trucks are in fleets that are way less than 50 trucks, so that there's a real challenge in this whole management because we are dealing with such small entities who don't invest in technology, et cetera. So this is a whole area that I think, you know, worthy of, of a lot more discussion and, probably with people who know a lot more about it than I do. But, but something that you haven't asked about because frankly, no one does. Okay. That's just, we just assume we've got infinite number of trucks and infinite capacity and have at it.

Tom Raftery: 30:49

Cool, cool, cool. Tom, that's been fascinating. 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?

Tom Moore: 30:59

Uh, They can go to any of our websites, transportation where optimisation.com or auto scheduler.ai or provision AI dot com, but they can also just send me an email, tom.Moore@t-wo.com.

Tom Raftery: 31:20

Fantastic. Fantastic. Tom, that's been great. Thanks a million for coming on the podcast today.

Tom Moore: 31:25

Always good and it's great to see a man with a great hat.

Tom Raftery: 31:30

Thank you. Okay, thank you all for tuning in to this episode of the Digital Supply Chain Podcast with me, Tom Raftery. Each week, over 3, 000 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 our guests or a personalized 30 second mid roll ad. 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 the digital supply chain. Thanks. Catch you all next time.