In this episode of the CIO Confessions podcast, host Gert Jan van Halem sits down with Thijs Baars, Last Mile Solution's Chief Technology Officer. Thijs is passionate about the transformative power of IT, data, and AI/ML, seeing them as critical drivers for organisations to thrive in this new era. At Last Mile Solutions, he is committed to leveraging technology as a competitive edge, driving strategic growth and innovation.

In this new episode, Thijs discusses the challenges of scaling their platform to handle increasing transaction volumes while balancing stability and innovation. He explains the decision to move towards cloud infrastructure, driven by the need to optimise team efficiency and leverage cloud-based tooling. The conversation also explores emerging challenges like QR code fraud and the importance of thinking "two steps ahead" when managing fast-growing technology companies. Thijs offers insights into balancing immediate needs with long-term strategic planning within the e-mobility industry.

5 lessons you will learn from this episode:

• E-mobility is a rapidly growing and socially relevant industry.
• Scaling technology requires strategic shifts.
• Balancing innovation and stability is crucial.
• The cloud is a game-changer for operational efficiency.
• Fraud and security are emerging challenges in e-mobility.

[00:00:16] Jingle: This is the CIO Confessions, a podcast by Devoteam.

[00:00:49] Thijs is accountable for all products and engineering within Last Mile Solutions. Welcome Thijs.

[00:00:55] Gert Jan van Halem: Great to have you in the podcast. I was wondering, after your experience with FastNet [00:01:00] you are back in e-mobility now with a break with Ball. com. So e-mobility is really your thing or is it by coincidence?

[00:01:21] And especially if you look at 10 years ago when I started my career at FastNet and where the market is now, there's a big difference. So volumes have really changed, the pickup has really changed, but there's still a rollercoaster ride. Growth 100, 200, 300 percent are not unheard of. So it remains very interesting to be in.

[00:01:41] Gert Jan van Halem: Yeah, so thanks for being here because I think that makes it very interesting for the listeners as well, because it's a market with such growth. And change that also the technical perspective from your view, I think it's very interesting. And for me, when I started thinking about it, it started with indeed the relevance of this business as you described, because [00:02:00] it's I'm not sure when we actually started driving electric, but I think that was the big drive to really make this grow and make this adopt.

[00:02:18] Thijs Baars: Yeah. And I think that's really interesting. But Last Mile Solutions is that it started at the very early beginning of EV driving electric vehicle driving passenger cars in the Netherlands.

[00:02:49] And that was from all kinds of different pivots in the history of Last Mile Solutions, the moment where they said, okay, this is it. EV boom, this energy transition is really happening. We want to be there. [00:03:00] So from the very, very early days of the EV industry, Dastman Solutions has been a key player in the market.

[00:03:19] Gert Jan van Halem: So it all started with that board, the connection and the communication. So that also explains all the charging stations that Last Mile Solutions has, but you do also the cards.

[00:03:32] Thijs Baars: everything in between.

[00:03:33] Thijs Baars: so at the moment we stopped doing all hardware developments. So we're a pure software player. And if you look at the EV industry, there's basically two sides of the market. There is the charging point operator that provides energy via charging points towards cars.

[00:04:09] So if you look at so that's also where a big volume comes from, is from ensuring that the CPO gets paid for the electricity. And the mobility service provider gets their money from Boomer, their customers are and ventures that money to the CPO.

[00:04:32] I'm not sure if you're allowed to share stuff or,

[00:04:50] Eso we always talked about triple digit growth, but double digits in numbers on you can see that. So and we're now in this moment where it's the volumes are getting very large [00:05:00] and the percentages are getting slower. But if you look at different markets where we also operate in, so to say in the eastern part of Europe, Romania, that this is still happening.

[00:05:19] Gert Jan van Halem: Yeah.

[00:05:34] And even with doubling the amount of traffic we had, it was still peanuts on what a modern server can do. And now that we see that the percentage wise, the growth is slowing down. The volumes are still tremendously growing. And now we're looking at a phase where it becomes much harder and harder to do it yourself.

[00:05:58] Gert Jan van Halem: And what does that mean more in detail on the [00:06:00] technical perspective? Because I can imagine, not sure if it's right, but I can imagine that you just start with one application, and it works, and it grows, and you can handle the growth as you said.

[00:06:19] Thijs Baars: And that's the interesting situation where we are in now, is that we need to do a double strategy. So we have developed countries with large customers, real enterprise customers that want reliability, they want availability and they're looking to really set down a very professional, solid service to their customers because a lot of businesses are starting to require it.

[00:07:02] So for us to become on the technical part very difficult because a we want to maintain a high level of quality, high levels of availability. We had a big program to really up those numbers. And at the same time, we're dealing with customers that want fast fast moving parts, rapid development.

[00:07:38] So we need to look at different strategies. And that's why we're now slowly starting to move towards the cloud.

[00:07:57] Thijs Baars: Yeah. So I think so for the [00:08:00] last couple of years it's been Very simple for us. We have a very small team to maintain our service because the volumes weren't just that, weren't that large. And a lot of tooling that you have in the cloud is open source. So we apply that as well in our own setup and our own infrastructure.

[00:08:39] So we did an analysis, right? And I think that's one of the things that I've really learned over the last two years that I've been here is that every week is a different week. Like the world has changed. So any preconceptions you had a week ago or a month ago might no longer be valid. And we knew on this cloud decision to stay off the cloud for the last two years, [00:09:00] we knew that was a contentious point.

[00:09:19] And I think key for us is that we noticed that for a long time with a very small team, the Capex investments or the investment in your hardware was key. Now that we're seeing that with org growth, we also need additional people. The capex is no longer leading. OPEX is starting to lead. So the people cost is starting to lead.

[00:09:53] What do they spend their time on? And one way or another, if and even if you use all the tooling that the cloud has [00:10:00] from open source packages, you're always building your own ecosystem. So these people are coming in and everybody needs to understand how do we work? What is our way of deployments?

[00:10:25] You need to maintain that. So that's a lot of time spent on something that you could argue is not justifiable or not directly or core business. And then if you move to the clouds, you gotta get this out of the box, right? People have. Experience with AWS, Azure, GCP, whatever public flavor you have. They don't need to learn all the tooling there because they are aware of it.

[00:11:00] Gert Jan van Halem: And I can imagine because that's what I like about, or what I find very interesting about this is that It's, yes, there's a technical side to it because of the volumes, but it's maybe more an operational side on how do I spend the time of my people in the best way possible?

[00:11:19] Gert Jan van Halem: exactly.

[00:11:30] Gert Jan van Halem: And then I can imagine in a business like you are in, it's also there's to what extent do you do that?

[00:11:49] Thijs Baars: Yeah, and that's and I think that's also safe. People often like what is the difficulty of migrating?

[00:12:12] The way you have your cost in place, so your expenditure, you're like, I need to approve the bill on a new server. And this all of a sudden in the cloud is, something weird happens and the bill might rack up. So you have this different cost features, you have this different this different perspective on how a cloud operates versus your own data center.

[00:12:54] And I think those are really difficult decisions to make. You have a vendor lock in, and you want that vendor lock in, [00:13:00] because it gives you a lot of benefits. And it's, people are easy to start up and innovate, but at the same time, if you ever want to move out. For whatever reason, that vendor lock in is also, might be very dangerous.

[00:13:23] Thijs Baars: Yeah, that's a really good question because it's definitely the place where we are now.

[00:13:45] You can use the built in queuing. It takes you probably a week to build up your own queuing cluster again, and that's acceptable. But if you look at your your identity management, that is so ingrained in our application system, the way you do your authentication authorization. [00:14:00] Yeah. So analyzing do we actually want to do this?

[00:14:12] Gert Jan van Halem: Yeah. So to some extent you think up front. Okay.

[00:14:21] So this is what we accept. The risks or the decisions and the sort of part where we say, Okay, we need to be a little bit slower here because we do want to be in control across the same thing. We don't want to put hard brushes on it. But in the month in the first year, we're gonna monitor these costs very closely and then decide on a path forward.

[00:14:44] Gert Jan van Halem: And so far, that's quite infrastructural, right? So it's cloud infrastructural, but There's of course a lot of software and products running on top of that.

[00:14:57] Thijs Baars: That's a really good question. So I think yeah, so I, I [00:15:00] think this is a little bit the classical core, non-core do we believe that this is competitive edge for us or not? We're coming from a a history where we did everything ourselves because we could.

[00:15:32] But it's always in a very that's very tightly controlled. There's a very, yeah, we have a very close control of where do we decide this is no longer our core or we don't believe we can compete in this. But I do think that's all infrastructure site. And the ability that cloud gives you is especially on the product side on the software development is to really speed up your innovation.

[00:16:10] Jingle: IT Confessions Box Every IT career has its challenges. In the Confessions Box, our guests share their lessons learned and resilience gained.

[00:16:36] Thijs Baars: I think my life does get better a little bit with AI, but it's a bit of a mixed bag. I had a discussion with a partner of ours and they were telling us that A. I. Is a bit of a weird beast. Things that we consider very simple, counting the amount of arts and strawberry as a great example.

[00:17:08] It's a little bit hazy where we can see the benefit because of this adjustment of things. I do think, and that's so the ecosystem that the cloud provides is also an ecosystem or the road path, the road map towards AI. And one of the, in the analysis we did do we need to stay in DCs or do we need to move towards cloud?

[00:17:37] Gert Jan van Halem: And then developing the basis yourself is too much of an investment, especially for you guys. It's not your core, of course.

[00:17:58] Gert Jan van Halem: harder.

[00:17:59] Gert Jan van Halem: So at the [00:18:00] moment it's monitoring what's happening and how could it help us more than having your own advanced idea on this is how it will change.

[00:18:07] Thijs Baars: And basically aligning us that once we see the opportunity, we can directly use it. We can step into it and we

[00:18:18] Thijs Baars: Yeah. And I was thinking on this last week and you see all these small steps for people saying it's just a big hype.

[00:18:40] Let's see. But so you need to be on path on these smaller steps. And if you're a small company with small volumes and you're focused on, running your own business, you have this tunnel vision where you don't really realize this. And now that you're looking, okay, we're really going to pivot.

[00:19:02] Gert Jan van Halem: And you mentioned edge. And I think for with charging stations and the man in the middle sounds a bit scared, but you have to do some stuff on the edge as well, right?

[00:19:16] Thijs Baars: So it's a really interesting setup. So in a sense, every charge point is an edge computer or an edge server. It accepts input from a car or a charge card or any other device. It does some processing and then it sends it to our central server. So you could say that we're running one of the Europe's largest IoT networks.

[00:19:58] So even though you might be thinking [00:20:00] it's a car, you put it to your charging point. It's a local connection that needs to be authentication of that car or a charge card. We need to understand what is your car made off. You want to get maintenance metrics from either the car or the charge point.

[00:20:18] Gert Jan van Halem: So you have to have a few on the whole chain that's involved to make a decision locally.

[00:20:27] Gert Jan van Halem: It almost feels a little bit like in the old days when we had dumb clients, smart clients.

[00:20:39] Thijs Baars: Yeah, exactly. Yeah. And I think so what they try to do is let's try to move away from a lot of authentication authorization stuff. Put it on the edge, right? So let's try to remove around it to the central server, especially if you're in the city, the connection will be fine.

[00:21:10] I know, I've been there. Yeah, so you wanna, so how can we speed this connection up? So let's do the authentication authorization in the charging point. But then the charging point needs to have a continuous up to date database of charging cards in the world. Is this card still allowed? Is it enabled?

[00:21:44] Is coming back to us with people understanding that if their card is cached, but they disabled it and the cache hasn't been updated, you're able to charge. Or let's say we don't want a localized payment terminal, but we give people a QR code. People are gonna use, [00:22:00] scan the QR code for payment.

[00:22:16] Gert Jan van Halem: And it's actually been a nitpick of me for a long time.

[00:22:35] Thijs Baars: Yeah. And the idea behind it. So the idea of the QR code is very good. We want people to be able to charge without having a charge card. So make it very accessible wherever you are in the world. Everybody has a phone, so everybody can scan the QR code and do their payments. But then not realizing that anyone in the world can also put their own QR code sticker on top of that.

[00:23:03] Gert Jan van Halem: thing, don't just scan any QR code. Oh, please don't. But it's also, to be honest, a hard problem, right? Because, okay, that's QR code. Can I trust it? Can't I trust it?

[00:23:27] Thijs Baars: Yeah. And I think it's a an indoor industry is a new problem to solve.

[00:23:47] Okay. The charging point operator doesn't care because if they lose out on eight or 12 euros, that's an accepted risk. And we, you stripe it away. Now, with volumes getting bigger, a charging session might [00:24:00] cost up to 10 times. So 80 euros is not unheard of for a charging session. And of course, there are thousands of cars.

[00:24:26] See how can we combat it while maintaining a good user experience? Because we do want people to be able to pay at a charge point without having their charge cards.

[00:24:38] Thijs Baars: So there's a couple of emerging ideas. One of those is they're going to generate dynamic QR codes, so a two factor authentication code where you see the token refresh.

[00:24:49] Gert Jan van Halem: But still, I can't.

[00:25:05] So I think there's a big push towards that. But definitely more costly than adding a QR code sticker on the charging point. So there's a bit of a balance there.

[00:25:22] Yeah, no worries. And never look at a QR code. But are there many people actually using it with a QR code or is the main way to do it with a card?

[00:25:41] So you have a car, you can start charging via an app. And of course, if you can hack that app in any way, you can start a session for anyone at a charging point. And that's where the volume, so then you can with a forerunner means charge multiple cars at multiple times at the [00:26:00] same time, and then the volumes become really large.

[00:26:04] Thijs Baars: Absolutely.

[00:26:25] Thijs Baars: Yeah. So I think what's if you're in a very fast growth scenario you need to start thinking two steps ahead. And it sounds very logical, but doing it in practice is a lot more difficult. At bold. com There was a serious growth going on where we doubled the size of the the engineering capacity.

[00:27:08] And so you need to figure, okay, what if I add 50? But, it's likely we're going to start with 50, right? So you're going to start, I think, okay, what do we need to organize? If I had 80 or 100 people, how do I organize that? How do I prepare myself that when those people land, they land in a team that makes sense?

[00:27:36] And we want this to keep maintaining the startup vibe. So you don't want all these paper processes. And I think so, and the big lesson is it's not just about adding people. It's about making sure that there's a culture there that allows them. There's a process there, but in the right way that the current people with 30 Still accepted as well.

[00:27:58] Gert Jan van Halem: challenge. And [00:28:00] in the beginning you said, okay, but you have to think two, two steps ahead. Yeah. But you also have to solve all that stuff that's happening today on people, the technology stuff will break. Yeah. Yeah. Still.

[00:28:14] Exactly. So the plane is flying and you're trying to build a plane on top of that on top of that. So that's there's a multiple. And that's and that's

[00:28:28] Thijs Baars: Yeah, so I think the process is key. That's, yeah, it's the balancing act. Yeah,

[00:28:35] Thijs Baars: And I don't want to have coined the term agile, but we need to be agile. So you need to continuously be aware of what is really top priority. What do we need to fix now? You want to stay out of a firefighting mode, but that's really difficult in exponential growth.

[00:29:04] Gert Jan van Halem: Yeah. So it's always

[00:29:13] But as long as it isn't burning because it's a massive program, it's going to take a lot of time of your people. It's a lot of investment in terms of money. You're looking continues okay, is this really top priority now? Or can we do another half year and resolve these other burning issues?

[00:29:30] Thijs Baars: is never a shortage of burning issues. So that's not a good

[00:29:33] Thijs Baars: And all of them, and that's, and I think that's the beauty of the market, is they're all good problems to have. Yeah. They're always issues because you're skating too fast.

[00:29:58] Gert Jan van Halem: And in the end, good problems are [00:30:00] also problems, right? Oh, yeah. Yeah,

[00:30:03] Gert Jan van Halem: Very interesting. Thank you so much, Thijs. I think we learned one thing. Don't scan QR codes. Yeah. But we also I really like your look on how do I look at cloud? How do I put my people on the right place to make the right investments for them as well?

[00:30:31] Jingle: Discover other podcasts on https://www.devoteam.com/podcast/. Devoteam, AI-driven tech consulting.