Frank van der Hulst: From Robotics to Water Pipelines - The Evolution of Cutting-Edge Technology

This transcription is AI-generated and may contain errors or inaccuracies.

Leszek

My name is Leszek and I will be talking with Frank van der Hulst about scaling a startup climate, tech trends and challenges in the water industry. Frank, can you take us back to the beginning? What inspired you to start Hulo and how did your career path lead to your focus on water conservation and AI?

Frank van der Hulst

Yeah, really, really nice. Thank you at least first for having me.

Leszek

It's a pleasure.

Frank van der Hulst

Yeah, it's. We've been so my, my road has been. I first started in aeronautical engineering so I've come from like completely different than underground pipes. So where I'm now acting in. But so I'm really an engineering background and then I focused a little bit more on robotics at some point and then I did. I started my career in robotics and self driving cars. So really completely different topic.

And then I moved actually to, to a company acquaintance. They built robotics for water pipelines and that's where really my journey started. So in the water industry it was just really fascinating. What really fascinated me at the beginning was just directly the knowledge of people about water. It's so common that you turn open a tap at least in the Netherlands it's so common that you turn open a tap and water flows out of it. But actually to make that, make that possible is really interesting and how that, how that is all about so how all the systems work. And then through research that we participated in, we found actually that there are still super high leakage rates around the world where those robotics were really focused on the western utilities with low, low leakage rates.

But yeah, worldwide there's around 30% is leaking away. So in the system. So in. So it's, it's an amazing amount. And that actually really made a really good start of where you already started. And then next to that we also saw how the current water industry works. That also is something that we really would like to challenge that status quo.

So with that energy I think we found it Hulo as a company who makes a software for water utilities to really detect those leakages on time and quickly to mitigate all the problems and the water loss considering to it fascinating.

Leszek

I mean now that we've got a sense of your journey, let's dive into the technology itself because it's really fascinating. Can you explain how the technology works especially how it detects leaks in real time without relying on it extensive historical data. It's really interesting.

Frank van der Hulst

Yeah, that's a good question. And I shouldn't dive in too deep of course, but no, in general our technology that's based on the use of that there are now existing a lot of IoT sensors. Right? So I think over the past decade, I think IoT sensors became really popular. A lot of utilities worldwide, they installed a lot of sensors. And now we actually get the question a lot from a lot of utilities that yeah, we get so much sensor data, but we cannot make sense of it anymore. So where of course AI and machine learning in general, really big plays a big part and that's also where our technology really relies on, that there is a large amount of sensors in the network or larger amount of sensors in the network, let's say like that.

So that you really can focus on making sense of all the data as an hydraulic collaboration to each other or hydraulic correlation to each other.

Leszek

So all the water recognize the signal. Basically that's what you're trying to do.

Frank van der Hulst

All the water is in the pipe, it's connected. So one pressure sensor at the left side of the network and the one on the right side, they still are influenced by each other. So that water is still. If something happens. That's a little bit how our technology works. So what we built is we built machine learning models which are on top of sensor data only, so the grouping of anomalies. And then we will move towards where is the leak happening.

Basically that next step. But really that, that part where we rely on our, our machine learning models, that's, that builds like basically knowledge on top of the, or the models on top of the sensor data. And yeah, for that we only use, we use more sensors. So in more in the spatial sense then than in a temporal sense where traditional methods are much more focused around. Okay, you have one flow sensor, for example, that goes into the network with only sort of daily patterns. So people wake up in the morning and then they go to work, so there's a little bit of tip. And then in the evening they go to shower and watch and cook again.

So that's a really the, the standard pattern of, of water consumption. You can also see it in energy consumption as well. It's a same pattern. But that's really an fascinating thing. But what normally that the traditional method is always predicting that daily pattern. That's basically. So a lot of forecasting models are based on predicting the model what you expect to be there today again.

And that always will fail. If that's. For example, yesterday it was really sunny in the Netherlands, but today it's already starting to be becoming a little bit more cloudy. So people will use less water, for example in the garden, let's say like that. So you Already see those changes or other events will happening. So if we also saw with some utilities if a holiday season starts, everybody moves away. Also the complete pattern fails.

Then if you want predict that exactly that day then you need to give feed it a lot of external knowledge. And we much more rely on the internal knowledge of the the network itself. So with all the sensors they have available to predict other sensors basically. So that's what it's come down to. So if you want to really flatten it, we make it. We build multivariate time series models based on the existing sensor data in the model.

Leszek

Okay. And how efficient can you get compared to traditional water management methods?

Frank van der Hulst

So basically.

Leszek

Yeah, or leakage detection methods.

Frank van der Hulst

Yeah, I think the. The one I think it's in. There are three metrics basically what are really important for utilities and, and water loss. That's basically time. So how long when how fast can you detect a leak? Second is which size leak can you detect? And then third is the the location of the leak.

So how well can you localize it? I think in time we are exceeding everything what is currently in the market. So we detect in seconds basically when data comes in we directly detect that there is something going on in the system. So you can give utilities really fast response times where the current status quo is called night flow monitoring. So that basically is checking the night flow of today. So at 2, 3 o'clock because that's where you expect the least amount of consumption. So stable consumption of water.

Compare that consumption to the day before and if that is so is that it's a little bit increased then you could think of that there would be a leak in the system. And that is the status quo. What we were we one of the things that we battle against basically. And then so the fastest that they can detect is in one day. And if you then want to have a little bit more security around your yes, it's really a leak instead of that just there is more water consumption in the evening then you need at least for example two days or three days. So leaks are generally quite long in the system already. So that's where we challenge on time then on the other one is on the size of the leaks.

Yeah, we in that sense we are limited by the sensor. So how. Well how good are the sensors? Which kind of noise do sensors have or the density of the sensors in the network? And I think in general we are there on. On the same level as that night flow monitoring. So nice flow monitoring can really see small increases and we can do the same.

But I think if you look over a week, for example, you see steadily that, that there's definitely an increase. I think that gives a little bit more certainty. So we are there on the same level as it, but not better. So I'm not going to here convince you that we have the holy grail on leak detection in the water industry. That's that I won't be. And then the other one is on localization. And I think on localization we can detect leaks on like say 250 meter radius area.

So that's not pinpointing it, it's here in the street, but it's really in this area of the network that you have to look. And for your reference, we are monitoring in general around 300km of pipe per area. So that's quite a large areas that we are looking at. So if you can reduce that to 200 meter, maybe 500 meters, so that in that range, that search area you will reduce a lot of time for water utilities find leaks in the field actually. So that's really good. But we are not. There are methods in the industry that can really pinpoint where a leak is, but they are also really cost intensive and resource intensive.

So it will take you a month to find those leaks, but then you have a tightly spinpointed. So we always picture that to the utility as well. So we are a monitoring system and as soon as we detect a leak, you still need somebody physical in the field searching and pinpointing the leak, the traditional method. So what, they can find it in one day. Instead of that they have to search for a month, for example. So that's the difference that we challenge in that sense. So yeah, I think that's a little bit how our system compares to the current status quo.

Leszek

Just a follow up question. I can only imagine that the pinpointing part is totally different when you're looking for leakage in an urban sort of location or kind of terrain compared to a countryside, I guess it's totally different kind of venture, right?

Frank van der Hulst

Yeah, for sure. Especially if you have no insights at all. So normally in the city areas they have for example five sensors per 100km, let's say it like that. And in the rural areas it's maybe two or three or something. So also the search area is much reduced. But then with a lot of like expert knowledge you can quite reduce. That's also partly what we do as well, like to reduce the search area based on the leakage patterns that you see, for example.

But that's really more a traditional method. It's leakage pattern that you see that's really is for example of a pipe with this diameter and with this type of material. So then you can zoom in a little bit if you know where those materials are in the ground that you can see. Okay, this, this leakage pattern is basically connected to PVC pipes, for example. Then you know that you only have to search for PPC pipes. But it's something that we really at the moment still do as almost on a ad hoc basis instead of really having it in our platform at the moment.

Leszek

All right, can you share some success stories? How Hulu helped reduce water in specific regions.

I don't know. And maybe before that you could explain the specific locations that you operate in.

Frank van der Hulst

Yeah, yeah, for sure. So we are currently operating in the Netherlands. So our home base in the Netherlands, the question is a little bit different than in other countries. Next to that we are in making waves in. In Brazil. Next to that we are also having some progress in Oman and in Africa. So an African continent.

So that's also nice. And some other European utilities that we collaborate a little bit on. And yeah, in the Netherlands. So why is the Netherlands different than other countries is because the Netherlands has a really low leakage rate. So they are in all the charts, basically all the rankings. They are at number one. If, if it comes down to.

To leakage and the low amount of.

Leszek

Leakage that they have, what's the reason behind it? Is the infrastructure new, high quality or are there any other reasons for that?

Frank van der Hulst

Nah. So there are multiple reasons. One, I think the most important and takeaway for water utilities abroad is long term strategy planning. So I think 30 years ago they started with a lot of data gathering already a lot of. So a lot of really cool good insights. And also in like really okay, which pipes should we replace and make them really that on just not on ad hoc basis, but really making studies of that. So they had really good strategic planning next to that.

Yeah, we are of course the soil that we have in the Netherlands. There are almost no rocks in our soil. For example mh it's flat. So there's a lot of. So not too much ground movement. So we are in that sense we are also lucky. But if I compare it to other utilities who abroad it's definitely.

I think the success factor of everything is that they did a for the last 30, maybe 40 years. It is really good strategic planning within utilities and work together and so they make it really good steps there.

Leszek

Where are you planning to expand next?

Frank van der Hulst

So we are planning to extend Next in Brazil, that's really where we want to scale for us. It's a really interesting market because they are quite far with their digital maturity. So they have a lot of systems already operating. They still experience high leakage rates. So that's somewhere we can really help directly onboard those clients. Instead of that, we have to help them first by actually placing sensors.

Leszek

Interesting. I'd like to talk about challenges. Every innovation faces challenges. So what were some of the biggest changes you guys faced when first developing Hulu, whether technical, financial or related to market adaptation adoption? Sorry.

Frank van der Hulst

Yeah, so we have a couple of challenges and I think one is really on the technical side. It's like also the discussion of cloud and fees on premise and all those kind of discussions because water utilities are quite, quite conservative and water is also seen about most in every country as the level one security, I think it's called or I'm not sure in English, but at least that it's of state important that you protect the water sources and protect also water delivery to people at homes because so they always have sufficient drinking water. And that's due to that, for example, terrorists can attack the drinking water sources and have a large impact on everything. So security, they have a lot of high standards. That was one of the challenges that we had to overcome. So to be trustworthy to let them know what we talk about, we develop a cloud solution. So that also gave some extra challenges to it because we get often the request for on premise solutions, but that's for us.

Yeah, solution will also be much more expensive. So that's not what you are really looking for. And so that was one of the challenges that we had to first really decide and we tested everything. So like okay, yeah, maybe we should go on premise and maybe we should go in the cloud already or maybe we should support both. And at the end we made a decision to support both also due to mentors who had more experience than me in the technical sense. They basically said, yeah, now you should try to stay away from on premise solutions, especially in the complexity that we work with with multiple systems from our clients. That's something that we had really some issues with.

And I think in another challenge that we in general have is also due to the water utilities are conservative so the general adoption of new technology. They have quite a lot of long roadmaps and we are of course a really young company. So that's also what you see in the most in the most in water utility industry in Sweden. Most of the companies, they are actually consultancy companies. So they get like small assignments and they try to develop a software tool to decide. Whereas we said no, we, we don't want to do consultancy at all. We want to do build a product.

Once you know that the sales trajectories with water utilities are more than a year basically. So it's a long, long sales lines. So that's also why we got external investors on board to help us in that sense with, with exciting that Runway and also give focus to the product that we can deliver something that is really valuable for water utilities. But we'll have his payback in a couple of years. Basically when you went through all the process with water utilities and you can go for to really work together with them and show the value of that sense as well. So that was really a challenge. Yeah. To find the financing.

We are luckily we were in a really lucky position I think that we didn't have too much challenge for that. So it was a pretty. Went quite well for us on that side. So looking forward to the next funding round. So that's. See how that goes.

Leszek

Keeping my fingers crossed, I can only imagine. But one more follow up question to that which is you guys started and as I understand Netherlands, which is as you said, one of the top countries in terms of low leakage ratio, percentage score, whatever you call it. And I can only imagine that it's really harder to sell in Netherlands so like this kind of service than it is for instance to sell it in Brazil, which is facing all those challenges you mentioned. So my understanding is that you actually started in the market that is really, really hard. I mean it's local, you can sell locally. But it's also very, very hard because the value added is not as big compared to Brazil for example. Is that kind of true or I'm misunderstanding something?

Frank van der Hulst

No, I think it's at least definitely in some extent definitely true. Because in the Netherlands our value proposition is different. Indeed. So in the Netherlands we, we think what we add value to is to the utilities in time. So how fast they are, the response time to a leak and that can have great impact as well on damages to other infrastructure. So around the leak for example. But that's, that's so it's not about the water loss, but it's much more about the damage that the leak is causing.

And yeah, we showed really good. So in every utility that we worked in the Netherlands with we showed really good improvement in response time. So how fast they were at a leak because in the Netherlands a leak is detected by the people who live there, so they call when they have no pressure anymore. That's basically how it goes in the Netherlands they don't have too much monitoring systems on leakage. So that's also an interesting point. And indeed in Brazil the value will be much higher if you can prove your technology there. So that's definitely also what we see.

So it's really a good market to grow whereas the Netherlands is a good market to innovate and to collaborate with the utilities because they're quite open at innovation levels and especially also to other Dutch or European companies. So that's really good and really a nice ecosystem, let's say it like that. So that's. That's something that we have already the benefit in the Netherlands of and that they gave us a shot like a couple of years ago so that they. That we can try and prove our technology. That was already. But Brazil is a much more scaling market and where we can add much more value.

Only then there are other things that are challenging and that's for example the tech system around external companies. So that. That is a challenge. That you live far away of course is a challenge. So there's a lot of challenges. And as well next to. Next to the Netherlands for example the UK is also losing around 30% of their water.

So it's really high. And there you see a lot of challenges around how the market works. So they are quite. They have like five year plans and they invest at the beginning of those five years and you have to wait again five years which is really not as well for a startup of course because you need to prove before and then you need to access it. Next to that it's also quite competitive market due to I think that this all commercialized and people know each other and the other one is for example we also did Italy, we looked in Italy for. Because they also have a lot of high leakage rates but. And they got a lot of funding from actually from the EU for to solve it.

But all the funding went to engineering companies which then have to search for technology partners to integrate. So it's a.

It's also weird system. So that's also again really an atmosphere of where you need to know people and that you need to know the network and to really get into those contracts basically. So it's again a long road. That's I think the biggest challenge in water industry.

Leszek

Absolutely. I mean expanding globally is a great opportunity but it's also challenge and challenging.

Frank van der Hulst

So.

Leszek

So. But I want to shift gears a bit and talk about some other challenges which basically scaling company itself presents its own set of management dilemmas. And let's talk about those. So scaling a startup I guess often means finding the right balance between innovation and managing operations. And I want to ask you, how do you guys balance those two things with Hulu and specifically at what point does management become too much in your overview? In overview.

Frank van der Hulst

Yeah, that's also a challenge that we definitely face I think in Hula. So we started to scale the company, the team a little bit this year. So it's also something that is really recent for us in general. We had I think to get to really to team collaborating. We have an international team as well which is really an. I don't think it's a challenge. It's also an added value to have multiple cultures and multiple diversity within the team.

I think that's really good for the team balance. But to get everybody collaborating and then people were used to their own type of work and then new people come in and it's also sometimes a little bit challenging for everyone in the team. But yeah, on one end. When does management become too much? I think that's a very good question. I think if you are more in meetings than that you're doing actually work in an early stage startup, that's really some point where you think okay yeah, I'm now too much in meetings. But it's also, yeah that's I think where management comes to mind.

Leszek

Yeah, that's got it, got it. Speaking of you know, international team, diverse team, what qualities do you prioritize when hiring and yeah, basically that's the question. What kind of traits are you looking for?

Frank van der Hulst

Yeah, we have a quite. Yeah let's say strict interview process where we do first really focus on the technical. If we look at developers so we do a technical test it should be a fairly simple test but sometimes they are quite A lot of people were struggling with it. So there was one part that we, we do in the process and then we also after that this is successful we invite them to do a personality test. So we are looking at what is somebody's personality and how will they fit in a team. So what are good traits and what are good traits for the current team that we have in the current. Yeah, people that we have in a team.

So for example if you have a lot of. In the beginning we were all like yeah, that's more go getters but also much more on innovation. So we were. The whole founding team was not really good in finishing and make tasks and also being doing the repetitive things as well. Like so if you do all the time the same things. That's something that we, we miss. So that's.

That you also, that's something that you are really looking for into people that you hire so that you're complimentary to each other. So we try to test that in that sense that we, we use a tool for that touch tool to map personalities. And in general what we are looking for in our current stage is really people that fit to our values as a company and we set them as we set honesty, adventure, achievement and respect to each other. So that's the values that we are going by and it comes really from. From me and from Robert as founding members and we are searching people as well who fit that. And we recently had our team day and actually we got some compliments from people from outside of the group how well everybody was already fitting to each other like in a such short time that it was.

Everybody was really looking really comfortable. So that was really where you can see that people have some kind of same personality. So it's not like that's. That they are. Yeah. How do you say it? If they collide to each other.

That was not really the case. So that's, that's something that I really was really happy and appreciated to see that I think we have at the moment built a great team of great personalities and people indeed with diverse backgrounds.

Leszek

Cool. Congrats. My final question is related to climate tech. So my understanding is Hulo plays an important role in tackling water management issues. But I want to get your thoughts on how do you see tech, climate tech evolving in the coming years, what's coming next and also how Hulu will adapt itself to those trends and happenings in the market.

Frank van der Hulst

Yeah, that's a really nice question. Can go really along about that as well. But in general what I think is important and what people and companies should really think about is about value instead of like about how much money they make. So that's also something that you see in climate test is that they are always counted on how well they perform. So we are how much water can you say for example, that's something that people will judge us on as well. Next to that if we make money, yes or no, but. And for example, if you do decarbonization, for example, then you will.

How well does can you decarbonize the air. But in general. So that's where I think that is a really important factor to always judge a company of about how they do add value if it's, if they even turn, break even or not. But I think there should also be a value in. In other things and owning money. So that's basically. Basically it.

I think there's some really good programs for that already happening worldwide. So you see larger companies that want to have like zero carbon footprint or zero water use. So you see for example the big cloud providers, they want to compensate for their water use in the data centers to compensate for it by helping the local communities around that. So by providing funds there to help those communities with technology what normally would be out of scope because of price, whether they can actually use it. So that's something that is really interesting and I think that's. That's also an important part in the sense of that people will. That. Yeah, that that ecosystem is much more in.

How do you say it in a. In a balance that you. Not everything turns about profits. I know that the world is not working like that unfortunately. But that's something that I think is really important because at the end also for example in water systems. So a lot of people they make judgments of like pipes. Why do actually pipe leak before, for example, why don't we just over engineer pipes, for example.

So we throw in a lot of different materials over the years. They fill in all different ways and now we are throwing in new types of materials. We don't know how they fail because they will fail in 50 years. But you can also for example throw in a cast iron pipe which is thick enough to that you know that it will last for 200 years for example and will never fail. It's also a possibility. But you make some decisions there and I see that often with death especially. And also I think you low in that sense also part of it we add value so we can show that we lower the water loss of and the water use of utility.

So we value in that sense what it would be even better if we actually were not needed. So if there were not many leaks, for example. So I think that's in a lot of different sense as well because I talked in the beginning about decarbonizing a little bit. That's I think a hot topic in climatestat as well. But yeah, the best thing would be if that was not needed either. So if there was just much less production so always in that system. So that's a little bit compensating on one side and one is yeah.

Making right choices in the beginning.

Leszek

Thank you, Frank. Thank you for your insights. Thank you for sharing the Hulu story, the technology behind it. Well, basically it was a pleasure to have you here.

Frank van der Hulst

Thank you. As well. It's really nice to be here. BetterTech Leadership powered by Brain Hub follow Les Schick on LinkedIn and subscribe to the Better Tech Leadership newsletter.

‍