Simon Schenker: 0:04

Not only it was a lot of inventory, but there was a quarter million saved right there on one material. And that's something that we see quite frequently, especially as we calculate all the materials and have the dashboard,

Tom Raftery: 0:15

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 digitization of supply chain. And I'm your host global vice-president at SAP Tom Raftery. Hi, everyone. Welcome to the digital supply chain podcast. My name is Tom Raftery with SAP and with me on the show today, I have my special guest Simon. Simon, Welcome to the podcast. Would you like to introduce yourself?

Simon Schenker: 0:43

Yeah. Thank you, Tom. So I'm Simon Schenker, I'm co CEO and founder of Genlots, and we're doing optimization and supply planning. So very excited to to talk today to you.

Tom Raftery: 0:56

Okay. And Simon, tell me a little bit about Genlots. You're a startup, how long have you been in business and what was the genesis of Genlots?

Simon Schenker: 1:09

So the Genesis was basically during my master's, at ETH Zurich. I worked in supply chain because I find it interesting. and, I had a professor who was a very good professor and I wanted to do something, applied something with some impact on not only for the library. So the basic problem that a company had, which approached my professor was kind of, they negotiated a quantity discounts, scale prices in procurement, but then kind of supply chain. They ordered all the ways in very small batches because they wanted to keep inventory low. So there was the question of whether there's something, that could kind of resolve this conflict as a neutral arbiter. And, the idea was to look, what is so very quickly I came to the economic order quantity, which probably your listeners already familiar with, but it was not satisfying and I did something on my own.

Tom Raftery: 2:08

For people who might not be familiar with it, you know, maybe you want to explain that.

Simon Schenker: 2:14

Yeah, it's basically, when do you order how much when you have to order supplies for your production in a manufacturing company, and there's this lot sizing, which tells you basically with a formula, what, what is the ideal lot size? What is the ideal order size? And so each time you need to order your order, this size basically.

Tom Raftery: 2:35

Okay. And how did Genlots come about then?

Simon Schenker: 2:39

Yeah. So I did something on my own with reinforcement learning and it worked out really well. This company had around 300 million of revenue, 120 million of, expenses for the supply it was in pharmaceutical contract manufacturing. And, I could show that this really a high potential for savings in the several millions. And so it was cleared for me. I wanted to make a startup. I came out with, with the name Genlots, as a link to two lot sizing and, try to launch myself first time but it was just after university. And I realized very quickly that, you know, there's large, IT organizations to talk to. I was not a developer. So I went for kind of four years into consulting, digital strategy consulting, learned a little bit the trade off, working with developers, designers, and so on and talking to large enterprises. And then I launched myself very smoothly, first, 40%, and then a hundred percent when we saw, that the interest was high and, uh, together with my brother-in-law Arnaud who had previous experience as a startup CEO. And that was 2017.

Tom Raftery: 3:53

Okay. And what's, What's been the progress since where are you now?

Simon Schenker: 3:56

In the beginning we were for a long time. It was just us two. And we toured basically in Switzerland, it's a big center for procurement where we are. There's a lot of procurement organizations here. Basically we didn't have a software. We had our algorithm, we had a slide deck we toured people and looked, like, is there any interest, what is the feedback from the market basically. And then we got some investment in 2019 employ people, started, had our first pilot, had our first implementation in the pharmaceutical area and, then Corona hit and the planners had a lot of other things to do. Just getting the supplies, inventory kind of lowering in inventories, um, uh, you know, achieving cost savings. Became kind of very low on the priority list. So that was a bit of hard time for us, but we continue to develop and started also to getting involved with SAP through the accelerator SAP IO. And that's also how we, Tom, got to know each other, right.

Tom Raftery: 4:57

Exactly. And, for people who might be unaware of the SAP.IO, just maybe say a few quick words about that.

Simon Schenker: 5:08

Yeah. So it's an accelerator. Does the selection and once you're in the program, there's two streams. There's a business and technical stream. The technical stream tries to integrate, your solution better with SAP. Say that it's very smooth for new clients to onboard. And the second part is kind of the business stream where you're looking into models of co-selling the software. So now we are on the SAP store even.

Tom Raftery: 5:35

Okay. And has it been useful for you? Was it a waste of time or was it a good use of time or.

Simon Schenker: 5:44

SAP is a very large organization. So having kind of someone on the inside, which helps you navigate it and bring connections to the right people, whether it's technical, whether it's on the business side, we had a very interesting session on pricing with kind of the peoples who do pricing for SAP. So we can learn a lot from mentors inside, but it's also on the technical side. We have access to demo instances and very good documentation, which is sometimes hard to get on you're completely on the outside. So it has been very useful.

Tom Raftery: 6:17

Okay. superb, and as you say, you're now in the SAP store, which is, which is nice.

Simon Schenker: 6:21

Hm.

Tom Raftery: 6:23

Okay. Talk to me a little bit about, I mean, in the, in the prep call, you talked about the triangle between cost, cash and service. Do you want to expand a little about that for people who are listening.

Simon Schenker: 6:36

Yeah, it's a concept that comes more and more that basically you have this triangle between, uh, cash, so inventory working capital and so on, the costs, which might be prices, which might be delivery, transportation, and so on. And the service, the customer service, or in the case of supply planning, the service that you basically give to your production, which should be very high because if production has to stop because you don't have enough material, it's a big problem. But you cannot kind of increase one without making a compromise on the others. So if you want to have a stellar service, maybe you have very high safety stocks to make sure you never run out of material. So that means a lot of capital investment, right? That's one of the examples or you, want to have lower inventories, lower cash, but then maybe you don't benefit as much from scale prices. What we basically did at Genlots is to look at the total cost of ownership. So if you want that the surface of the triangle and trying to minimize that, and then you still can set priorities and say like inventory right now for me is kind of more important than the rest, or maybe cost is more important than the rest but you can then see immediately what is the impact on the other two.

Tom Raftery: 7:56

Okay. And I assume as well, that if you are, ordering large numbers of particular, whatever it is, components that impacts the cost as well, because very often you get discounts. If you order more, is that kind of taken into account as well?

Simon Schenker: 8:14

Yeah, exactly. So scale prices and everything. So it's very interesting because with this total cost approach, you can basically extend it at will. So we have more and more people who are interested in sustainability and CO2 ton emissions. And, we have done a pilot with a client where we calculate the CO2 emissions for the ordering patterns. And then we can say, okay, a one 10 CO2 equivalent is a hundred euros and, integrate this basically in this total cost calculation that you can see the trade-offs, but the advantages that, uh, kind of in supply planning cost reduction means often as well, CO2 reduction, right?

Tom Raftery: 8:59

Hmm. Yeah. Yeah. It's a bit like that. There was a similar concept of that triangle in the printer space, for example. There are three kind of vertices of, of printers, cost, quality, and speed. And you can have any two of those, but never all three.

Simon Schenker: 9:18

Yeah. And it's well known the as well in project management, right? You have the scope, you have the time and you have the number of resources. And if you want to, if you want to have everything at once then it becomes a death March, but, yeah, you have to either increase resources or use scope

Tom Raftery: 9:35

Yeah.

Simon Schenker: 9:36

or increase time if you have unforeseen issues,

Tom Raftery: 9:39

Yeah. And, planning is a, is a big part of supply chain. I had an episode, a couple of weeks back with a colleague David Vallejo talking about it and it's the space that very much you're playing in. So talk to me a little bit, bit about planning and, and, uh, how you work in planning and where you think planning is on kind of the maturity curve.

Simon Schenker: 10:03

So I think just generally it's quite underestimated, the impact that good planning can have on profitability. and I think also, there's a huge range of maturity. Like there's companies, which are very low margin businesses in a sense, like a Nestle, for example, in FMCG, in general, it's very important to have operational excellence, right? So there are quite advanced on planning and trying out a lot of new stuff. Whereas for example, in pharmaceuticals, The supply chain is a bit of a site thought because the margins are very high. So the priorities on R and D and, and compliance and things like that. So, there the maturity might be much, much lower, and it's just some people kind of with some Excel spreadsheets and SAP and trying to do their best uh without, much support from, from the global or central organization, right. And planning is quite large because there is demand planning, right. Looking at consolidating the sales numbers from different markets and looking at your forecast. There's production planning, where you have also kind of a lot sizing process in place. And looking at which lines and where to produce what. And where we are active, specific in supply planning, right? Where then you're kind of between the production planning and the logistics where you, where you say, okay, that's what we know, will come from production, what they will need. And over the next. Three to 12 months, we plan our, our purchases or even 18 months in some cases in pharma, it's a very long, long, long lead time. So also in just different areas, the maturity is quite different. And I think, um, in demand planning and kind of forecasting, there's a lot of things going on right now. And it's interesting to me that a lot of companies, they don't even measure how accurate the forecasts are. I spoke last week to, to a prospect and they started measuring, how accurate the freight costs are and how much they change every two weeks. And they were quite surprised that it was like over 20% of variance between the last two weeks. And so it's very difficult when you have kind of very varying forecasts, to do a supply planning, which optimizes the triangle alright. So for me, it plays well together to kind of optimize the whole chain and not only demand planning, but then also get the benefits from a better forecasting, for example, in supply planning.

Tom Raftery: 12:45

Okay. And while there are these different levels of maturity, are those gaps closing? Are people who are lower down on that maturity curve or are they you know, starting to, to try and fix that?

Simon Schenker: 12:59

I think it's a very slow process because I think there's not a lot of people just working in a, in supply planning for example. And, the people have varying degrees of backgrounds and education. And, what I realized really over the last years is that for example, if you want a CRM, so a customer relationship management tool, if you found a startup. Like one of the first thing that things that you have to do is to approach customers and take notes on your conversations and so on. So there's hundreds or thousands of startups in this space, right. Whereas, supply planning is something which only comes when you produce something physical and once you get bigger. So this just also not a lot of people who are in touch with those very hairy operational problems and have a background, for example, in development or in mathematics to kind of optimize this right. So, it's starting to get more interesting. Also, Corona has shed a bit of a light on the supply chain and it's importance, but I think there's still quite a long way to go. There's only few players active, but there. Then they start to become more and lots of people start to exchange more. but then also people feel like they're, uh, company is very special and different. So also sometimes there's a bit of silos between companies for information and exchange of best practices.

Tom Raftery: 14:31

Okay. And does planning, you know, does it affect profitability?

Simon Schenker: 14:37

Uh, yeah, by quite a lot, because you have this triangle in supply planning, I mean, it really much depends on the industry, right? In some industries like chemicals, maybe 80% of your expenses are raw materials. Whereas in pharmaceuticals, it's maybe 25% of your expenses, but basically as a supply planner, you manage the huge, a huge, huge proportion of the spend of a company. And it often is very big compared to any other thing like rent or, even salaries in some cases. Right. And, uh, so you have very few people who manage a very large proportion of the budget. But they're not very digitized and not helped by a lot, by the organization. So, this aspect and the other aspect, we already talked about this kind of just that in terms of digitalization, there's a long way to go. So for me, it's really an area which has a huge potential to shift. And we have seen this with our clients where we kind of reduce the number of deliveries by 60% and the inventory is by 20% in one go. So the whole triangle got much smaller and, I think there's a huge untapped potential and it's not very visible in a sense to maybe a financial CFO. I mean he sees maybe that inventories go higher, but there's also always kind of a reason that you can cite for it. Like, uh, yeah, we get more sales, right. Then you have some more inventories or things like that. But it's difficult without having kind of a benchmark, to see how well optimized you are, basically.

Tom Raftery: 16:19

Okay. And I mean, you, you, you mentioned some figures there and that's interesting. Are there some, and you're a very young company as well. So are there some wins that you can talk to us with with customers or is it still too early for that?

Simon Schenker: 16:33

So we're not allowed to cite names yet. It's still kind of blocked in their marketing department.

Tom Raftery: 16:41

familiar with that dance.

Simon Schenker: 16:44

But, I can cite maybe some anecdotes to make it more concrete. We had, a pharmaceutical, also a contract manufacturer. Right. And, we did what we do. Sometimes. It's a proof of value where we analyze like on some 50 materials how we would have ordered with our algorithm, how they have ordered a very quick export from SAP manually. Right. And, in the presentation we were like, yeah, there's something really strange happening here because there was one material, that. From our numbers, they, they would have to scrap something like a quarter million of this material, but actually it was true because their, their client canceled the buy

Tom Raftery: 17:30

Wow.

Simon Schenker: 17:32

and they forgot to cancel their material, their materials at their suppliers, and it has the six months sell shelf life. So you can imagine they were like during the presentation that we did, they were calling their planners and they were like cancel, cancel, cancel right. Not only it was a lot of inventory, but there was a quarter million saved right there on one material. And that's something that we see quite frequently, especially as we calculate all the materials and have the dashboard, you know, prices here seem strange and you see like, okay, you know, sometimes somewhere it has to be divided by a thousand because it comes in the box of thousands, but the prices in the systems are wrong. Like this one factor is missing. So it's a kind of a master data. Cleanliness gives you sometimes some surprises, but yeah, that has been, I think the most extreme example of it, just the more material we could to save quarter million, they can catch.

Tom Raftery: 18:31

Nice. Nice. Nice, nice. You, you're doing a lot of those kinds of presentations. I imagine you're a startup and you're very early on. What's the supply chain space like for startups? Is it hard to break into.

Simon Schenker: 18:47

I think it's, it has its positive aspects and its drawbacks. I think one of the most, positive aspect is there are not a lot of startups, as there is not that much innovation going on. It's easier to get into contact with people who are in the field. And, uh, they're interested in talking to you. Right because they don't get bombarded by hundreds, thousands of propositions. Like if you're a chief marketing officer, I imagine you have to have a very, very good spam filter because there's a lot of companies wanting to sell something to you. So for us, it has been interesting. Also learning people take the time to explain, what are the issues they're dealing with after the drawback? A little bit as a startup is kind of. Getting budget and supply chain is much, much harder than for example, in the, in the sales department. It's just not something that, I talked, To company last month where, you know, say like you have SAP, you should, work better with SAP, right. And use, use it. You have it already, but there's not kind of a budget for innovation or something like that or, or having startup collaboration it's getting better. But, uh, when we started out five years ago, basically, uh, supply chain doesn't have an innovation budget.

Tom Raftery: 20:43

Yeah.

Simon Schenker: 20:44

Whereas other departments, R and D especially, right. They can experiment and so on. But supply chain is also, the people are sometimes not very happy to experiment because it's kind of, don't touch a running system. You know, things work. If it breaks, it has a huge impact. So, that's a bit the drawback as market for, for a startup.

Tom Raftery: 21:05

Okay. And you mentioned again in the prep call that you're using not just machine learning, but reinforcement learning. And just for people who are listening, you know, everyone is familiar with the term AI and a particular part of AI is machine learning. And a particular part of machine learning is reinforcement learning. So could you talk us through the differences between those three and why you're going with reinforcement learning and what advantage that gives you?

Simon Schenker: 21:39

Yeah. okay. So AI is a very general term and it includes much more than computers, right?

Tom Raftery: 21:45

No, Siri on my phone.

Simon Schenker: 21:49

Yeah, exactly. I mean, it's, it's also kind of a, when do you consider something as intelligent? Uh, it can be biological. I had a very, uh, at my university EPFL there was a very interesting project of recreating neurons digitally on a chip, right. And simulate what happens to them and they react similar to a real brain and things like that. Very interesting. The machine learning is for me I like to say getting value out of data, it's kind of number crunching. And, uh, within machine learning, you have different areas and different approaches to machine learning. And one of those subcategories, as you said, this reinforcement.

Tom Raftery: 22:33

Some of some of the big, well-known examples of machine learning would be things like Google's AlphaGo, right?

Simon Schenker: 22:40

Yeah, exactly. So ah Alpha Go is a very interesting example because Go is a Japanese game, like think chess, but a much more complicated, but it has quite simple rules. But the game was for a very long time considered unbeatable by a machine like in chess, you have computers, but you can train against them, which are better than the world's best player. But in go for a long time, they were very proud that, uh, the machine can never beat them. Yeah, it has to be always a human. So there were two different approaches, selected. The first was, AlphaGo by, um, done by, uh, a company now owned by Google and what they did. They looked at a lot of, of past plays of, players, right. World's top players. They had tens of thousands of different games and they trained the machine, in, uh, quite a long process to kind of imitate the players and find the best strategies. Right. And, so you have a huge database of historical data and, you tried the machine to get the machine to imitate this. This approach is used a lot in insurance, for example, to check whether someone is worthy of your insurance or not, or for the credit score and so on. There's a lot of signals that you might overlook as a human, but the machine knows from past experience, did they default or did they not default? And you can kind of infer in a black box model okay this one is good, I will give him a loan, a credit or something. Then they did AlphaGo zero, which was kind of the next iteration, which used reinforcement learning. So the approach was completely different. No historical data needed just the rules of the game. And then basically, the algorithm played against itself itself, many, many, many games, but in a very short timespan, And then it's beat the players, uh, kind of 99 times out of a hundred or something like that. And they beat the, the former AlphaGo, uh, algorithm. So it's a very interesting approach because, for us concretely, it means that we need to know the rules of the game. Like what are my minimum order quantities in our case and so on, but then each time from scratch for each material, we kind of, optimize it by knowing the rules and finding the, the best possible ordering strategy. And we don't need tons of historical data, which also might introduce human biases, right. Because maybe if you imitate how it has been ordered before, but it was wrong, then that's not a good thing. Right. So, um, I think it's a very interesting approach, reinforcement learning. And I think also it will, will come up a little bit more because you don't need to clean historical data, which is kind of a big constraint in supervised learning, for example.

Tom Raftery: 25:44

Okay. Cool. That's a really interesting Simon. Is there we're coming towards the end of the podcast now, is there any question I haven't asked you that you wish I had, or any aspect of this that we haven't addressed, that you think it's important for people to be aware of?

Simon Schenker: 26:00

Yes, I think it's important that supply chain gets a bit higher place in the priority within a company. I think that's something that has been very overlooked and I think also there's, um, there's really a role, in companies for someone who kind of analyzes and optimizes those processes. Because what I see is that a lot of people, they, they ask to doing operations and at the same time, kind of doing a bit more longer-term optimizations and, they know very well the business. But often it's, they're putting out fires all the time and have no time to kind of taking two steps back and I'm looking at the thing and then maybe there's external consultants, but they don't know the details of the business. So, I see us as one part of, kind of a revolution of digitizing. Supply chain. Right. But there's another part, which is clearly people and the time they, they have to think a bit more about the big picture. So I think that's just a very important point.

Tom Raftery: 27:08

Nice. Nice. Okay. Superb. Simon, if people want to know more about yourself, Simon Schenker, or about Genlots or any of the things we discussed on the podcast today, where would you have me direct them?

Simon Schenker: 27:21

Yeah. So for me, you can find me on LinkedIn, Simon Schenker, right? And Genlots is Genlots dot com so G E N L O T s.com. Or you just type in lots sizing in Google and you will find us probably.

Tom Raftery: 27:37

Nice a bit of, a bit of good search engine optimization done there. It sounds like.

Simon Schenker: 27:42

I got a very large company contacting me because we were the only ones they found any information on. And you have one article on different lot sizing methods, which drives 95% of all the traffic to our website,

Tom Raftery: 27:57

Wow.

Simon Schenker: 27:58

because it's just a desert in terms of information. So we didn't optimize by a lot, but we just have one article which you know, in an area that nobody other says no one else serves. So

Tom Raftery: 28:10

Fantastic. Fantastic. That's amazing. That's

Simon Schenker: 28:13

lucky?

Tom Raftery: 28:14

Uh, send me a link to that article and I'll include it in the show notes as well. And, uh, that'll help drive a little bit more traffic. I hope.

Simon Schenker: 28:22

Yeah. Thank you very much, Tom for inviting me.. been a

Tom Raftery: 28:26

No problem. Thanks. Thanks. A million Simon for coming on the podcast. Okay, we've come to the end of the show. Thanks everyone for listening. If you'd like to know more about digital supply chains, head on over to sap.com/digital supply chain, or, or simply drop me an email to Tom dot Raftery @sap.com. If you'd like to show, please, don't forget to subscribe to it in your podcast application at choice to get new episodes, as soon as they are published. Also, please don't forget to rate and review the podcast. It really does help new people to find the show. Thanks, catch you all next time.