Alan Holland: 0:03

I was a former lecturer in AI, but I'll be the first to put my hands up and say, there's some things machines just cannot and will not ever do and shouldn't do but there are some things that are highly quantitative that machines are great at. So when you see a breakthrough technology that, can really reduce your workload, it makes sense to use it lean on it and those who do reap the rewards

Tom Raftery: 0:24

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 Allen. Allen, welcome to the podcast. Would you like to introduce yourself?

Alan Holland: 0:51

Pleased to meet you Tom. My name is Alan Holland I'm the founder and CEO of Keelvar.

Tom Raftery: 0:57

And what is Keelvar?

Alan Holland: 0:59

Keelvar is a software vendor headquartered in Cork in Ireland. And we specialize in advanced sourcing software. So we help big enterprises when they go to market and they're buying goods and services. We specialize in helping the largest enterprises doing that at scale.

Tom Raftery: 1:16

Fantastic. And congratulations on being based in Cork in Ireland, it's my hometown. It's where I grew up. So, uh, no better place I have to say apart from where I am now in Seville.

Alan Holland: 1:26

Yes. People in Cork have a habit of congratulating themselves on their Corkness and celebrating is very frequently. It's what we're famous for.

Tom Raftery: 1:38

And rightly so.

Alan Holland: 1:39

Yes.

Tom Raftery: 1:41

So, let's get back to Keelvar Allen. Why did you set Keelvar up?

Alan Holland: 1:47

I set up Keelvar, because of a problem I encountered when I was a PhD student in fact. When I was studying for a PhD in computer science part-time in the evenings, I was working for my parents, a chemical company just to pay the way. I had to help my dad responding to tenders and RFQs from buyers of chemicals. So there's a big dairy companies, municipal water treatment plants, and so on. And I could see how frustrating it was for suppliers and these pricing mattresses that my dad would always be unsure of year to year whether he'd be winning contracts in the west, so to the east or the north. So he never knew where to place the storage facilities, how much to invest in transportation and manufacturing. I could see that this is an information problem. So that's why I started studying the topic of in academic circles is known as combinatorial auctions. So how suppliers could start communicating packaged beds with dependencies and capacity constraints, much richer information. The great thing for me was that I could see the opportunity for computer scientists and that there was a PhD in this because you needed an algorithms to compute winning outcomes. And so it was all neatly, came together, came across the problem and it was working in a new brand new AI research lab is very well funded at some of the world's top researchers in computer science coming from all over the world Cork and problem fell on my lap and when I finished my PhD, I was publishing case studies, real life examples, and companies started knocking my door saying: actually, this is a big problem for us. Could you help us with what was now been called sourcing optimization?

Tom Raftery: 3:30

So you went from there, you set up a company. Walk me through that whole process.

Alan Holland: 3:37

Before I set up the company I was fortunate to have there's some funding agencies in our enterprise Ireland. It's a funding agency that helps researchers explore the commercialization potential of different research projects. So I was working as a lecturer, but also a part-time researcher. That allowed us to build prototypes and run projects. And one of our first projects was with the local Cork city council. We helped them buy a fleet of 250 vehicles and 16 categories and it was extremely successful. So we were able to prove that not only were buying organizations happy with the outcome that they make savings, but suppliers we're equally happy that they're increasing their profit margins, winning the right combinations of contracts. And that gave me the confidence to say, okay, well there's a business in this. So, I think I was fortunate that coming from a business family, I knew who to go to which solicitors you needed to help you incorporate the business and accountants, what they could do for you and what you're obliged to do when forming a company. So I think coming from a business family helped in those early steps.

Tom Raftery: 4:53

Yeah, no doubt. No doubt. And so how long has Keelvar been in operation?

Alan Holland: 4:58

So I set up the company in very late 2012, I left the university and we took some seed funding from seed investors. We've seen this work for public sector organizations really, really well. Let's do sourcing optimization for government bodies. I think we learned a lesson that just because you have a product that solves a problem doesn't necessarily a meaning a product market fit, and that market needs a certain patiance. And sales process to educate public bodies that they have additionally beyond that or what something they're buying presently that you actually need to be crafting the tenders for them to go to market to buy what you have. That, that was two slower process. So in 2013 and 14, we were building a solution and we kind of led 2014 and we realized that we should be trying to target enterprise companies because they understood what they needed to buy better than government bodies. And they are coming, knocking on our door and saying, I like what you're doing for those public sector organizations. Could you do something similar for us?

Tom Raftery: 6:10

Hmm.

Alan Holland: 6:12

Procurement processes for enterprises are so much different to those for government bodies and governments are highly regulated procurement processes where frequent feedback loops and iterations and bidding not entertained very often whereas enterprises they're much more keen to impose more competitive tension in the process. In late 2014 we took a difficult decision to start from scratch. When funds were low, it's difficult to pivot when you're going to use that most of your seed funding. So that was a big risk but that's, that was the making of the company to. Because we had clarity on what we knew by that stage, what we should have built. But now we just had to go build us as fast as we could. And we were fortunate that we did some projects with very large enterprises who knew what best practice looked like. And builds trust in us because we ran some very successful projects for them that we could build a self-service project and product for their needs. And so some of our early customers, they backed us. And we still work very closely with them. And we have that neutral trust built up over many years that's integral to who we are.

Tom Raftery: 7:20

Okay, where does the name Keelvar come from?

Alan Holland: 7:23

Keelvar is it stems from an Irish word. So the Irish for sensible or intelligent is Ciallmhar. C I A L L M H A R an old Gaelic word. So it's an Anglicised spelling of what is it intelligent system. I used to teach intelligence systems in University College Cork, which is the application of AI and other techniques and intelligence software system. So yes, a nod to my university days.

Tom Raftery: 7:48

Nice. Lovely, lovely. So for people, because this podcast caters to a wide range of people in all aspects of supply chain, everywhere from engineering, planning, through to manufacturing, making, delivering, operations for people who might not be very familiar with procurement what problem is it that you're solving for . Companies?

Alan Holland: 8:08

We saw the problem of how to find win-win outcomes between suppliers and your buying organizations. So, when you're going to market and you're buying at scale. So let's say you're buying transportation at scale. You're going to need to deal with small, medium and various suppliers because you probably have a large network of lanes that you move goods over. You'll have regional vendors, nationwide vendors, and you'll have local suppliers. They all have their strengths. Now, if you're very good at buying, you will allow them play to their strengths and you will also allow them see where they're not as strong nearly. Contracts they would not like to win unless than when something else besides that makes us viable. So what we provide is sourcing our first product was a sourcing optimization. You don't aggregate on behalf of the market. You break down what you need and to small, granular dots. And you say to your suppliers, you guys bundle, design the packages, tell us what we would like to combine things and where your economies of scale are, where your diseconomies of scale are and what kind of capacity constraints you have. So you're gathering much richer information from the marketplace. And so in the case of transportation, that might be, let's say it's a company that's got it's serving. They've already serving customers who are manufacturers in Atlanta. So they have trucks leaving Atlanta at a full, but they have trucks going back into Atlanta that are empty. So they might want to say to you of all the 20 lanes going into Atlanta. If we went in 15 of those 20, we'll give you a 20% discount. Something of that nature. So by facilitating that type of expressiveness and bedding, you're increasing efficiencies, operational efficiencies for your suppliers, and they can bid more competitively. So they share part of that savings. They're giving you a discount, but their profit margins increasing. So that's an example of win-win. No, sorry. There's two sides to the sourcing optimization coin. There's expressiveness for the marketplace. There's also the expressiveness for your internal stakeholders. So when that organization, let's say, that it's transportation, you're buying your given the flexibility to suppliers to be descriptive, but you also should give flexibility to your logistics managers to be descriptive about their preferences. So maybe you're a logistics manager in California uh, doesn't want to award more than 20% of the volume to any one carrier. And wants at least eight carriers, and, it doesn't want more than 50% of the lanes to be switched from incumbents. They may might have all sorts of, creative, um, constraints and preferences around what they would like. And sourcing optimizer allows the buying organization to very quickly model like in seconds, add constraints and see, Okay. I have that constraint and price has just gone up by $500,000. Is that really worth it? So you can assess those tradeoffs easily.

Tom Raftery: 11:18

Okay. Interesting, So do you have outcomes that you can speak to, for, customers that you've managed to make savings for or something like that?

Alan Holland: 11:28

Yeah. So there's a few, a few dimensions to optimization and, and a there's, there's obviously the, the cost savings dimension. Right. And I think that was the primary driver of sourcing optimization in the early days. Most companies realize that, okay. By facilitating that expressiveness. You get savings. And they're typically in the order of five to 15% and can be higher. We've seen 20, 25%, when compared to a process that's just simply just cherry picking. And now the other dimension still, I think that the more mature users of sourcing optimization realized that this is a grade two for improving value outside of cost too. So for example, if you want to start biasing in favour of faster suppliers. Let's say that there are all sorts of non-cost dimensions, but in supply chains, the speed is important. So what you could do is change the objective function to optimize for speed. So you could have a scenario where, okay, I've just completely ignored costs, but this is the fastest outcome we could achieve. And usually, that might be expensive though. Um, so now you're looking at the trade off, so you start to work. When organizations start to quantify, how much more would we pay for faster deliveries? And it's not something you need to come up with a number, because it's difficult to come up with a number for that before you go to market. It's when you, when you go to market and then you see that, okay, you have options, a, B, C, and D. A being the cheapest, but the slowest and D being the fastest, but the most expensive. Now you have a continuum you can select from, and that makes it quick and easy for organizations to move along what we call peripheral frontier and choose, choose an outcome that's the best balance for them. So, so organizations kind of moved, I would say second wave of, usage of source and optimization started to better explore the speed versus cost dimension, but another dimension is appearing and that's around sustainability. So now I would say our top users are including scope three emissions information gathering in their sourcing events so they can start. So instead of the kind of crude approach to trying to gauge rates, individual suppliers, you could actually be looking at a much more granular approach where you're looking at what trucks are being used in what regions, what cities, so that you can start bias in favor of, more environmentally friendly alternatives. So if you have an internal cost of carbon, you can start to be quantitative in and systematic and biasing in favor of more sustainable alternatives. And for enterprises who are trying to operate at scale and need to operate as quickly as much smaller competitors, that's very, very important benefit.

Tom Raftery: 14:40

It's interesting that you bring that up because, we are in the twenty twenties which is supposed of action, and it's, nice to hear that organizations are actually taking this seriously and are looking for ways to reduce their scope three emissions using a tool like this.

Alan Holland: 14:57

Yeah, I think what has been heartening is that that recent procurement events we've attended And every, every, presentation I saw from a chief procurement officer, there was sin singling out sustainability, operationalizing sustainability as their number one objective. So, uh, I think that businesses have realized that in between generally speaking, between 60 and 90% of their emissions are scope three emissions, from their suppliers. So if they're going to make a dent in emissions overall, they have to be gathering much richer information from their suppliers and they have to start becoming systematic in building incentives into how they source. So that suppliers who are investing in this. area are winning

Tom Raftery: 16:20

Tremendous tremendous. Are there any particular industries Alan, that you guys are particularly suited to, or is it across the board?

Alan Holland: 16:30

We serve industries with supply chains. So, say foods, automotive, industrial goods, mining, electronic goods, CPG. In all of those industries, if you look at the top five companies in the world, we will be serving one or two of those top five companies in all across all those major industries.

Tom Raftery: 16:52

Wow. That's impressive and already about the fact that you're based in cork, but your customer base are, they they're obviously not just Cork, where where are they based?

Alan Holland: 17:03

Uh, yeah, we have a global customer base. So from Nestle and Switzerland to Siemens in Germany and Mercedes-Benz, and in the US, we serve the backs of Coca-Cola and Microsoft. So we have a blue chip customer base, predominantly in Europe and us but that's growing in Asia too now.

Tom Raftery: 17:24

That's fantastic. Congratulations. How do you do that? I mean, you are a startup, 10 years on since you started. How do you make that market? Because it is a relatively new market that you're making, right?

Alan Holland: 17:39

Yeah, that's correct. We focused all our efforts on our product first and foremost. We only hired our first sales person just over two and a half years ago. We were quite, quite late in terms of getting to the sales and marketing game, which allowed us to be quite ma I would say we were match fit in terms of our product. We were for all the customers we were serving. We possibly underestimated ourselves and strengths of our product because when, when we did start selling it and we did start marketing it more intensively, we were very high success rates when we involved in competitive processes. So we should have maybe should have been building a commercial team. And we lacked capital too, right. So we kind of come from, an origin where we had spent our seed money on, targeting the wrong market. And we were bootstrapped for several years and we didn't really look at alternative sources of capital until I suppose we started developing automation. That's our second product where we took a view that sourcing optimization is clearly delivering stellar results when it's applied by companies. But the big companies we serve have a limited number of experts in sourcing that would understand why and how to use sourcing optimization. So we said we won't boil the ocean and try to teach the whole world why and how to use sourcing optimization. Instead we'll develop sourcing bots to automate the eh, sourcing process so that you could have a sourcing bot for let's say packaging, ocean freight and that it would follow best practice, process and execute it to that standard every time. And when we took this to our customers, as our, second product. They loved it. They almost loved too much because they were saying, could we have sourcing bots for this and that? And the other. We said we're still relatively small. We were maybe a headcount of 20 back two and a half years ago. Um, we can't do everything. And so we said it's time to take on additional finance and grow, grow the team much faster. And so we, in the last two years have grown from 20 to over 80 and, we have kind of exciting new product roadmap ahead where we're, we'll be launching a much wider range of sourcing bots so our customers will be getting more of what they're looking for. when We'll be sating their appetite and it's probably more, hopefully.

Tom Raftery: 20:13

So that leads me nicely to my next question, which is what is next? Where to from here?

Alan Holland: 20:19

Yeah. So, we're all about intelligent systems. Whenever we have these product meeting our, company meetings where I'm re restating our vision and focus. What we have a vision for procurement in which software agents to all of the tactical work and all of the grunt work and people in procurement do all of the strategic work We'll be adding more intelligence to our sourcing bots. So there'll be able to know when to go to markets, know who to go to market, to, invite, to compete then know, and be training themselves, on, sourcing strategy that's optimal for, for them, so that if you're in a competitive market, you can use a different strategy. And another market where there's a lack of supply so this, that type of tactical buying where they're sourcing bots can be reactive and can absorb much larger volumes of data that actioned intelligence. Now you were saying 2020 has been a decade of action. That really speaks to the imperative for intelligent automation, but it's not sufficient to have actionable intelligence, you want actioned intelligence. So, where we're going next is, is making sure there's more actioned intelligence.

Tom Raftery: 21:37

Nice. Nice. Okay. Superb Allen, we're coming towards the end of the podcast now, is there any question I have not asked that you wish I had, or aspect of this, we haven't touched on that you think it's important for people to be aware of?

Alan Holland: 21:53

Yeah, I think that, you know, when we're talking about automation and intelligent automation, procurement, you know, sometimes we get asked about its impact on the future of work in procurement. And I think it can have a very healthy and positive impact on the role procurement plays in an organization. It would be a mistake to think that this is replacing jobs or moving jobs. Instead, what I would say is that it's making the jobs in procurement much more interesting because the focus of people's time will be on strategic activities, negotiations with suppliers, risk management, innovation discovery, and so on. Where you know there's, there's so much work machines can't, you know, I I was a former lecturer in AI, but I'll be the first to put my hands up and say, there's some things machines just cannot and will not ever do and shouldn't do but there are some things that are highly quantitative that machines are great at. So you apply machines when you see a breakthrough technology that that can really reduce your workload. It makes sense to use it lean on it and those who do reap the rewards.

Tom Raftery: 22:59

Very good. Superb. Okay. Alan, if people want to know more yourself Allan Holland, about Keelvar or any of the topics we discussed on the podcast today, where would you have me direct them?

Alan Holland: 23:13

Yes. And there's our website Keelvar.Com there's also our LinkedIn page is a good source. If you want to follow us on LinkedIn, where we're regularly, producing new content with a great content team, um, who almost on a weekly basis producing new and excellent insightful content.

Tom Raftery: 23:29

Fantastic. Alan that's when great. Thanks a million for coming on the podcast today.

Alan Holland: 23:34

Thank you.

Tom Raftery: 23:36

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 like the 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.