Insider's Guide to Energy

Episode 1 - ETRM Mini-Series with Previse Systems

• Chris Sass • Season 12 • Episode 1

Share episode

Chris Sass and Martin Hiller (Hosts of the ETRM Mini-Series) are talking to Asbjørn Hansen (Managing Director) of Previse Systems.

"At Previse Systems, we reshape the way Energy Trading & Risk Management software is designed, implemented and operated. We offer a highly performant SaaS solution for European power, natural gas and certificates trading - combining modern software architecture with the latest technology."

To get more insights about all the vendors that participated in the ETRM Mini-Series, check out https://insidersguidetoenergy.com/energy-podcast-education/etrm/

If you want to find out more about Previse Systems, click https://insidersguidetoenergy.com/energy-podcast-education/etrm/previse-systems/


 | Timestamp | Speaker | Transcript

| 03:16.60 | chrissass | Welcome to Insider's Guide to Energy ETRM Educational Mini-Series this week we are excited to bring with us to this episode vice systems and from Previse Systems, Asbjørn Hansen. Asbjørn welcome to the program.
| 03:26.49 | Asbjørn Hansen | Um, thank you I Chris Thank you very much great to be here.
| 03:31.74 | chrissass | Fantastic to have you and of course I have my co-host Martin hiller with me as well I neglected to introduce him but we have both you what we're going to do today is we're going to talk eachr m and before we get into it I think we should start fairly high level I've done this with all our previous guests and everybody going on it was we want to get your industry perspectives. So from my perspective. The eterm space is crowded. That's 1 of the things that made me wake up 1 day and say hey let's do an etm educational miniseries I see a lot of a lot of smoke I see money I see companies trying to reinvent themselves I see new companies coming out and so what I want to ask you to start? The show is. First what's happening in the industry. What's what's going on there. Why why? So much noise.
| 04:12.46 | Asbjørn Hansen | Well I think there's there's multiple reasons for the knowledge right? 1 1 part of it is that you have even though the the model's bit crowded, right? You have certain vendors who pretty much building up large monopolies right? So that's 1 side of things right? where a lot of products are now on the same hat right? So the end of day you have to list vendors out there right? and. But you have then you have the other side of which many small startups like ourselves right? Who's starting up and then saying actually there's a space here and there's lot of value. We can bring to the marketplace right? So when you combine that with a market from from the ah customer perspective right? which is very volatile it has a lot of market volunteers this year a lot of them have. Great surprises by which means that a lot of different areas of the market is certainly changing dramatically right I was just on the phone before this podcast for somebody in the nortics who's basically saying that there's no liquidity left in the market up there and people try to struggle struggle with that and have to find new ways of of getting bias information new ways of trading right? and and similar you see. The the price moves around europe the summer and is alsoum right? where a lot of the otc trading is pretty much dying out a little bit because of that right and people's credit limits are being hit and moving things to to exchanges but then cash and and marching cold becomes a problem for people right? so. A lot of people are kind of struggling and and trying to figure out how do they you know, utilize the cash they have in the best possible way. How do you How do you trade around around the asset you have in ah in a better and less cash- conssuming way because of course this is something that will hit everybody at the moment and systems is kind of 1 of the best way to. Address that problem right.
| 05:46.15 | Martin | So aspina I I heard you saying that on the 1 hand there is more like a monopoliistic structure in the market if you look at the mendorss but you also mentioned that there are more startups coming up with new ideas. So.
| 05:58.19 | Asbjørn Hansen | A.
| 06:02.60 | Martin | Why is there this movement that there are more startups coming up is it because of the monopolistic structure or is it because the the clients really require maybe new technology or new business functionality that monopolistic structures or companies cannot provide.
| 06:16.21 | Asbjørn Hansen | Well I think it's it's more technology actually driving at the moment right? So so lot of people are looking for for new technology right? and technology is is required to to meet the market we are in today right to we've looked a lot of the the legacy systems that were building at a time where you trade the ba diddle you did big thatch trades. Now, you're in the market where elgos are driving things right? people trading very very high pace and and and a lot of small trajes right? Which means the systems you're using needs a different footprint right? It's a different problem to solve right? If you have you know few hundred a few thousand large long-term deals that's kind of 1 problem to solve right? If you have you know. 50 million short-term deals. It's kind of a different problem. You need different technology right? you you can't use use the same the same shovel to dig dig the hole right? It's it's 2 different things. So so you know trying to find good ways of apply technology to do both and ways to separate you know, 2 different bus you type solve it 1 1 packages.
| 07:12.94 | Martin | And okay and and and how does from your point of view Priise fit in that sense I mean obviously you're a more young and startup-driven um company but tell us more about your.
| 07:12.99 | Asbjørn Hansen | Is is a big thing here.
| 07:28.42 | Martin | Solution in that context when it comes to technology and what you can offer to the clients.
| 07:30.95 | Asbjørn Hansen | Yeah, so we started from for a renewable standpoint right in inside say let's build something around ges first with something where nobody had the solution for that. We believe right? But then we took ppas and then we bought build our way back into the normal or classical trading software from there right? We we kind of building our products on on four different. Building box right? So 1 thing is performance and scalability right? So that's really about using technology to scale things up and and looking at you know trade captures and milliseconds not not anything above that right? Ah valuations is in in milliseconds as well. Right? We you know at the moment we have a target to be able to capture the trades on average at 200 milliseconds per trade and we want to be able to. You know have real sub positions in and and and around 500 milliseconds right? and we a good way to get there. Um, so that's 1 thing that that's what we need for the outdoors to work and for for the system to actually add that additional value and give our customers a competitive edge right then usability right is is is another thing right? where a lot of the legs. Your systems were built 20 years ago with uis and things that. Came from that era um, and we saw a lot. We've seen a lot of place where you know people put it spend a lot of money pretty consistent but the system adoption is very low because the users can't find their way around the system too complex to operate right? So building usability into software is is key right. Fast appointment is is really something that we also spend a lot of time on fixing the we be doing it something right now where we beside the contract with the with the new large company the other week and they will go to parallel run next week early um so by be able to put the software in very very quickly and actually scaling to other markets very very quickly without. Having to do much configuration mode. Most deployment is is key right? and a ended as thing is basically automation right? So system automation is is is the last thing everybody needs to be able to handle the high frequency of trades.
| 09:15.38 | chrissass | So It sounds like provis started meeting a need or or looking at new and emergent technologies coming in. You said you just landed a new large customer. Um, so this is a gray field there right? It's not a green field Play. So So how do you get?? um. Ah, transition What you know if you're a large trader and you have infrastructure in place. How do you convince or how you even if you're quick deployment. How do you convince them to make that move or are they willing to make the move or they only do it for part of their portfolio. How does a company likeise get get into a a large energy trade. It's established.
| 09:46.50 | Asbjørn Hansen| We typically do for for part of the yes we typically do for part of the portfolio and do it in smaller steps right? So you know these you know big bang project steps that was part of energy your impact in the past right and there's loads of horror stories about that that requires a whole bar conversation. But. But but those big projects I think that that's that's that's the past now right? We never but nobody's gonna do that anymore. Everybody knows how expensive that gets and and how big the risks are and and those things right? So we always try to focus on small areas of the portfolio and then try to work with the customers to basically build a landscape where you can plug in and plug out systems right? So you you know it's not enough to have. 1 solution that does everything you're never going to get the right solution for that anyway. But so being able to pick out different solution for different parts of your portfolio and then having an overview built with technology as well. Of course where you have eventdrive trades being pushed across the different systems into a consolidated picture is a much better way. But the customers to drive that business I believe right? So that gives you the flexibility you need to choose different tools for different things and at the same time it consolidate a view across it I think a lot of the customers. We talked to are using a lot of new technology as well in that in that space right? something like snowflake and those sort of. More modern data warehouse ah approaches which allows you to basically block your different systems that way and I think if you start looking at your landscape from that perspective. It's much much easier to scale out and use less different tools say actually maybe tomorrow you start trading metals or you have grains or you have eleg or whatever in your portfolio when you look at that. It might be different solutions that it's the best solution for those products right? Um, you know and that's kind of how we we we get in there. Normally we started in. We started coding in November Two thousand and Nineteen um we had a curse customer go live in July twenty
| 11:24.29 | chrissass | So so how old is your company.
| 11:33.82 | chrissass | Um, and then where's your development done What where what countries you develop in.
| 11:38.18 | Asbjørn Hansen | Well, we have our courts team in Switzerland um, so that's mainly for mainly the the courtship management group that sits here right? Then we have if quite a few different contractors in various places we have ah what so few people in the uk we have ah and. Contractor in Bulgaria and 1 and a set of people in Romania as well. So what we're trying to do is is is space it out and they use different people for different things right? So the core team who does stole all the framework and and all the high- performance things sit very closely together here right? And then we have a lot of crod apis and ui things which you typically you know, outsourced to other people say look you know. Here's how Spec go build this and bring it back to us right? because otherwise the cost effectiveness is not really there, right's it's way of getting using nearoring to to get a better cost effectiveness for our for urban cost fund.
| 12:23.17 | Martin | Yeah, cool.. Thanks! Aspion. So But let's move maybe a bit content wise to a topic that you mentioned already at the beginning renewables I think um, it's worth talking about it a bit more in detail because this is really what the industry is currently to big extent driving. So from your point of view. How important is the energy transition for your clients and what do you specifically offer to manage this transition I think you've mentioned already that you were yeah you grew from that direction. But maybe you can elaborate a bit more on that on your view.
| 12:54.20 | Asbjørn Hansen | Um, yeah, yeah, sure we actually did think about us is we actually only have renewable portfolios at the moment on with our customers right? Even the next 1 reside is also a big renewable producer right? So and we we started out from. From ppas from ges and so but first of all because we believe thats the future right and being able to track your your renewable production is is is crucially important for people right? and we think it goes on both sides both on the producer side but also from the buyer side and there's a whole market there of consumers of energy. Who's currently for my probably a bit untapped right? Who doesn't really realize that they actually need to manage their positions. Same thing goes with this c o 2 and then so on and forth right? There were people as long as the price were very low. Nobody really cared. But now that the price is very high suddenly it becomes much more ah much more important for people to actually look at what they're doing right? and. We. We started out with that. We build our system around that from the get-go right? So while we design things around kind of the a so of aer structures of trades and zpm mail sche and those sort of things we we faced the focused on ges first then ppas and then from there into into into the more classical powered gas. Portfolio record right? So so that that's how we we looking at that of course the the volatility that it brings makes it more important by 2 to Menna this? Well right? Um, both on the on a longer term position with the pbas. But also the short term. But it says it comes out of the the left or was in your portfolio. Why do you need to trade on a very short term basis.
| 14:25.60 | Martin | Okay, understood. So and so that basically you started out on on ges and ppas. But maybe tell us a bit more from from your client's point of view. What does. Makes so or what does ppa and ge os make so special or maybe complex is it like the capturing of the trades with all the attributes is it the pricing the valuation. The risk management I guess it's more than just 1 topic. But maybe you can elaborate a bit on the challenges and what you offer. Um.
| 14:52.81 | Asbjørn Hansen| Um, yeah, sure. Yeah sure start with the ge o's first thing right? because that's kind of the the funniest funest 1 its my view right? because what we see there is that people have.
| 14:58.72 | Martin | As privis.
| 15:08.41 | Asbjørn Hansen | You know geo is not just a geo right? You have all these different labels that goes with the trades right? So you somebody wants a geo from a newly built ah wind park right? and they'll pay extra expert for that other people want to have a geo that looks at it at a pretty pretty site or has doesn't kill someoneman does kill Birdss whatever is that that the label. Guarantees of the geo means that it becomes a very interest-par market right? There's a there's a big bitsk spread in this market from what I see at least where you you might go and trade generic ges on on more structured platforms right. But a lot of the o most of it is oc trading right? and a lot of that is really about finding buyers who's willing to pay extra from you know, hydropower done locally goes to where their customers are or for wind power done from a specific area with a specifically ecolabel right? So basically being able to capture that then being able to match up right to to. Take advantage of the bit aspi that this brings and this complexity of the contract brain actually both but doesn move to contract more the demand drivenven complexity right? that that that gives us that that's 1 thing that's really something that especially our toro customers spen a lot of time with us building into the software from the getgo so they can capture these things and try to optimize. How do I do that and the same time since the ts expire within a year right you make sure you want to always use the oldest first and and all these sort logic pieces logical pieces that you need to put into your into your software right? So that it's kind of 1 1 area right? and on the ppa side. It's more that of course ppa's typical, long term deals right? Um. With with lots of times and stuff and that's not so different from what it was in the past but every ppa have some sort of of its own logic attached to it. They're typically quite complex agreements right? with all kind of additional things and what we offer for that is we we have a core ppa solution right? but. Since every agreement might be a little bit different. We built our software in a way where everything everything is accessible by our ecosystem and where we app store so that people can build additional functionality just for the specific contract into the software and and have it execute seamless seamlessly right? So you don't have to work. Go outside and do the convention and put it back in. You can actually build your own little little app for this specific contract and and do thosevations something stuff like profit sharing or whatever it is in these pba contract right? is it's always ah hard to model from a generic standpoint and each each agreement would have different differences right.
| 17:30.33 | Martin | So okay, so what I understood from your view is that on the 1 hand more complex steel structures forgos and ppas makes it difficult for your for more difficult than usual for your clients to capture trades. But also the the whole topic. Life cycle for those kind of structured traits is then more? Um, yeah, challenging and maybe in that context because I think that's at least something that I can see that is challenging for clients the whole topic pricing orvalluation of ppas And. Also than risk management of existing ppas in the airport Fieldos I can see that this is not since it's structured trades something that you easily do So my question to you is? Do you provide a pricing and risk management model or do you more follow as of now the approach. Okay I heard now. I Heard you saying app store that you say to your clients here. We have an app store. We We open up our platform for you to develop your own pricing models or maybe you have a partner that that provides you an app for pricing and risk management.
| 18:38.86 | Asbjørn Hansen | Yeah, yeah, it also depends on where you are right? if you you pre-trade a post trade right? and I think there's 2 2 very different things right? So so we we focus main on post trade right? So the whole free tradede part is there's lots of venndors out there who does that's very well like qa and and other pi partner and forth.
| 18:45.92 | Martin | Yeah, yet.
| 18:57.94 | Asbjørn Hansen | Who all does these these pre-trade things right? which is 1 1 problem to solve right? Then we are more on the post- trade part where we're really looking at. Okay, now you've you've done the trade. How do you value it now and how do you manage the risk and and the the different elements of this agreement right? So What we of of course is the captures such a trade. Being able to capture all the different constraints that are buildings of the trade right? and being able to capture pricefulness and attached to the trade If. That's the case right? Um, but at the same time and of course Run normal market to market and those sort of things around around the the ppa's That's this, you know, basic bread and bottom stuff right? But at the same time if you need more.
| 19:19.38 | Martin | M.
| 19:34.50 | Asbjørn Hansen | Complex optimizations across it right? and then allowing people to build that with with while the ecosystem right? allowing people to do at their own functionality or hook in other vendors who have more deep risk management expertise than we do right? So we we're not trying to be the They'll be all right? We actually believe very firmly that we should focus on the main ether part and that people who are more specialist in different areas build their own apps for that right? So um, we don't want to build nomination solutions. For example, we don't we're not going to follow the nomination market very much We're not going to build quantitative analysis solutions. We don't have expertise in that right. Other companies out there who's way better at that and who should provide those things as apps on our platform rather than us spending time on it right? because that would give the customers a better solution and hopefully kubernetesative edge rather than us trying to to you know, spread too wide right? I think at least what I've seen in the past is that the lar vendors tried to build everything right. If you try to build everything you end up doing nothing really well. Um, so try to focus on what's actually really good for the for the main product and for the customers is is important as well, right Um, and then have partners built the list.
| 20:40.40 | chrissass | So so you talk about ecosystem you you mentioned that your customers might want to build something or use a third party but let's maybe walk dive a little bit deeper into how that would work and and describe your ecosystem and you know let's go down the 2 approaches whether they build it themselves or whether there's third parties. Ah, that you said that you'd might to have them involve. Can you explain how that works for your environment.
| 21:02.73 | Asbjørn Hansen| Yeah, so so but the ways it works is basically they we have a kind app framework where people can can sign up for right and and go in and build their stuff right? There's kind of 2 options to do it. They can either build things using our exist stack right? So basically using. Functions and those sort of things you shop code right? and and utilize all the the new gi that we already have it's basically sdk right? But you have all the the framework we've already building all the tools and functions we have built already. You can go utilize so you don't have to you know invent your own stuff around daylight savings or time so handling and all that sort of stuff right? You could utilize all that from. From our software right? Um, so that's 1 1 way and that way you can go build it and apply it in this app for yourself capableably deprive. It. So nobody else can use it that that's your app or you could say actually I want to resell this via the the app store right? in that case, you package it up. It goes through ah some some. View on our side in that case, right? before then gets parts of the apps. So and people people can buy it right? So that goes to both partners and for customers right? So some customers are keen on building their own apps and and they're building them and keep them in-house that they get their own provider information. That's especially around optimization and those sort of things. Others are looking more it this the interfaces right going. Let's get a third party to build this interface to get priced in from Bloomberg writerers or whatever, right? There's no reason for me to build that knowledge in in-housed that me get good. Get it off the shelf and get a pilot to build it and then resell it on platform I just pay a monthly fee for for the interface and that's it.
| 22:31.00 | Martin | So having an open platform where clients can build their own platform sounds definitely a new approach and a new direction to to walk on. Um, so in that context how many apps if you can. Tell us this. Ah how many apps ah do currently exist beside probably the basic platform that you need in any case and how many are maybe provided by by clients Third party.
| 22:53.40 | Asbjørn Hansen| Um, wealthy. Yeah, so so little bit try to do count and my head here right? So and that's a bit. There's a the hard question so there's about 7 with 1 client of which four are built from the outside.
| 23:07.51 | Martin | M.
| 23:09.58 | Asbjørn Hansen | Um, 3 are built by by us as apps right? that that's 1 1 client another client has 3 at the moment. Ah 2 whom to build from third party vendors 1 built by us and 1 has god knows how many 1015 a lot of them build up build on their own. Um, so so it's very different for different customers right.
| 23:24.36 | Martin | Um, okay.
| 23:28.58 | Asbjørn Hansen | And what what they're looking for right? What we tried to do of course is to standardize this a lot more. So as you get more and more customers. It becomes easy and easier to to onboard and just go and pick and choose the apps you want from from the ecosystem right rather than you have to go and readend the wheel by. So.
| 23:37.82 | Martin | That makes sense and who who is who is doing the maintenance of of of apps Once they are built. Okay.
| 23:45.82 | Asbjørn Hansen | The vendor who produced it right? So if customer builds his own app then if he's of course responsible for doing that right? If a third party develops the app and then delivers it that party normally does the main support as well, right? So we know to make sure that it always works right? We we version our apis right? so.
| 24:02.00 | Martin | Um, okay.
| 24:04.57 | Asbjørn Hansen | Even if we change our api. The partners app should not break and they typically have 6 to twelve months depending on what it is to move to the new version of the api app that before we then deprecate the old 1 right? so that they can continue to use the app that they build without having breaking changes all the time every time we we change something right? because obviously at the moment we.
| 24:20.83 | Martin | Yep.
| 24:24.90 | Asbjørn Hansen | You know we've got like fifty twenty people putting pushing in code on databases right? And of course the amount of change coming into the platform is ready to be high. Let's say that um and so therefore version bpis is is course very important for this.
| 24:35.47 | Martin | Yeah, yeah, understandable Maybe last question the topic so who is then who is then doing the pricing for for an app is it then with a client who builds it with the vendor or how does that work. Okay.
| 24:46.88 | Asbjørn Hansen | Mr. Ben with with the vendor right to the vendor etc own price. Um, you know I don't I don't Um, we don't get involved with that at all right? So we just take a fee if they set a via up our apps all right? Other than that it's up to them with what how they price it. So there's various different model right? So pricing it.
| 24:57.67 | Martin | M.
| 25:06.18 | Asbjørn Hansen | Monthly fees some yearly fees some some based on per price Index or portrayed whatever it is depends on on the vendor right.
| 25:11.24 | Martin | Um, okay.
| 25:12.98 | chrissass | So you mentioned earlier in your description of when you started the partner ecosystem. Yeah I think you said the word azure. So I think you implied Cloud somewhere there. What tell us what your Cloud strategy is.
| 25:21.84 | Asbjørn Hansen | Yeah, well we're cloud only so so that's basically this strategy right? So we designed everything in the cloud from the get-go right? We decided that we when we looked at this upfront and we we looked at that thought. Okay, let's just just build a full cloud native solution. Um. And that's what we've done right? So of course we can deploy it into a into a private cloud but the main strategy is are a multitening solution where people can utilize the same same as cloud right? So everything is built that way and from the get go also in in collaboration with Microsoft we' be working with it as a partner as well and then it's been quite a. Quite intriguing to see how much you can actually do utilizing the technology that they provide. Um, we we started not so longer using as ah, asure batch for example to to do computations right? and do very large portfolios of trades in very very little time. Simply because of the computation of power you can scale up and scale down as you need right? and that that's definitely something that the classical and-premise solutions just couldn't couldn't do right? not without having a huge data centerer and a huge cost to run it. This is way more cost. Effective.
| 26:25.40 | chrissass | So do you consider cloud normal today is that is that mainstream today or is it still on the front end of the curve for some traders.
| 26:32.90 | Asbjørn Hansen| It depends on which part segment you're talking about right? If you're talking to some of the major traders right? It's very much cloud is the is the way forward right? If you talk to some of the classical stat where is that's very little cloud in there and then you know we have a few actually where a few cases where. Russian companies basically say well actually cloudut is great but they want nowhere near any us focused cloud right? which is kind of interesting as well. But it kind of makes sense as well, right? So ah, but other than that I think everybody is is moving towards cloud and and ta will be where we go right? And and the 1 thing it does first us as well which is very different than. Classical solutions right? is that we do continuous release right? We release software. It's production this week four times so far. Um, right? And of course that is a big differentiator right? because it means you have no upgrades right? and so so being able to. Continue to release new instruments new product into production for everybody is is a big gain for for everybody and gives us this as thought about earlier the the faster time to market right? It gives us exact that edge right? So if there's a new instrument on e x we can have it in software next day right? and and that's of course ah a big big game.
| 27:38.98 | chrissass | So so I'm familiar with agile which usually sprints So I'm not so familiar with continuous release. Maybe you could educate me a little bit.
| 27:50.34 | Asbjørn Hansen | Sure, well basically what we do is we we take? Um, we take whatever so feature function or defect or whatever it is right? We you know, build it pass it through q a and and the beta run beta testing before we then push into production right? So each featured function gets released. Ah. Independently pretty much right? Of course there are some cases where we had to release multiple things together because it goes cross-comp componentent right? but most of the time you can release a lot of fixes independently and we also do it under something called feature flax which basically means that we we release code into production but not turn it on so production will continue to run with the same codebase for a few days before we then switch on the flag which starts executing the new code so that if something goes wrong, we can turn the flag off again pretty quickly right? and and have the old solution running right? and and those those are some of things we've we've been doing. We also do that for customers doing actually left release new feature. We think it's going to add advantage to everybody. We'll build it. We release it in. Turn on for 1 customer only. They'll go test it and use it when they're happy with it. We'll then go and turn on for everybody else, right? So but being able to release code that way gives us shoot a bad to dis discover to the more classical approaches right um.
| 28:57.26 | Martin | I stood. Okay, um, but how does it work. Then how do you manage the dependencies to vendors that code own apps and might also have a new version of an app and want to release that um I guess that. Have quite some quality assurance mechanism in place in order to prevent that and release of an app might crash down the whole system. So How does that work.
| 29:22.43 | Asbjørn Hansen | Yeah, well it works the way that that each app can't really crash down the whole system does some safety if you just build into the into the framework for that of course, but but each each app vendor has their own release Pipeline. Um and can essentialually release fixxes or change to the pipeline similar to what we do right? and but they can't change is the kind of the. The edge of it right? So you know and then there's also of course you know how many how many times can you hit ah an api and those sort of things right? does safety built into that right? So You can't cop Anhe by mistake right? And that's what we've done to to ensure that but actually the Q a on the apps themselves relies with the vendor right? So if the vendor builds an app.
| 29:52.31 | Martin | Yep.
| 30:00.90 | Asbjørn Hansen| Need to of course ensure that our platform is secure and and fine and but the Q way of their app is with the vendor right? They have to curate their own app and to reachase it out and and and they can feature like whatever they want as well right? Um, so so that that that's when we have doing it right? and we've also did got a lot of work around. Um, um.
| 30:11.99 | Martin | Um, okay.
| 30:20.45 | Asbjørn Hansen| Around other ways of releasing stuff where people are basically putting into docker containers. So basically you write your code in Python or whatever and you place a code docker container in with our with our software right? And then that way you can keep your your stuff completely private and a completely away from us. You can't even see the code right? So if you're releasing stuff just for yourself. You can do that way then hit the api from. From that sort of approach right again same thing apppies the right since we can't see your code. We can't do anything with it. You're responsible with what it does right.
| 30:45.68 | Martin | Okay, it. So maybe last question that and maybe was a very technical question so it sounds to me a bit like your setup in terms of the development framework is Programmer language agnostic. So the client can or the vendor can ah code in any Programmer language. But um, you deployed then on your platform via the deployment pipeline. So that's very flexible to the client and every client has a deployment pipeline that you can use. Okay.
| 31:17.52 | Asbjørn Hansen| Pipeline. Yes, exactly he can english do if he. But if he's an app lenor right? He can e kiss his old department pipeline for his thats right? That's of course the the core stack on our side of course, not it' not It's notgnostic. We use you know as your folks in suchure batch and those sort of things you sharp right? but. If somebody wants to write in Python or whatever they can put that into the container and deploy it that way that's perfectly fine, but that's right, they just can't resell it on the platform because in that model we have no way of verifying what that ad is doing right because we can't see into the code.
| 31:43.20 | Martin | Yeah, understood. Okay, um, it becoming a big back to to the whole topic ecosystems. So Do you also have ah partners that you cooperate with that might provide. A service or an app that they have already and I mean that requires often quite some interfacing So How how do you approach that whole topic partnerships in terms of ecosystem.
| 32:15.70 | Asbjørn Hansen | Yeah, well, it's it in the beginning right? Philosophy has always been that we don't want to build everything right? We want other people to build it, especially if they have more fatigue than we do right? and then they can build a better solution so we work with partners all over the place right? And normally.
| 32:22.77 | Martin | Yep.
| 32:31.37 | Asbjørn Hansen | Jokingly said to people that we we're bet with everybody. Um, so but but based that that's where basically what we we try to do right? If there's if this partner out there company out there who does who knows more about certain things that we do then we should involve those get those guys to build an app rather than us trying to build ourself right? It gives a better solution similar to same thing with a. We have different ways where we have kind of consultant partners right? So doing implementations with us right? We have um, development partners and then we have kind of complementary products right? like like Fiedx is a good example of a complementary product right? where you know it's a solution that adds value to our customers in addition to what we do. So of course, let's work together and let's build interface and things so that we can the customers can get more value of the overall solution. Even though it's not our solution everywhere. The solution for the customer consists of multiple pieces right? Um, and think that debt That's how it has to be.
| 33:21.42 | chrissass | So so you've got this ecosystem there you mentioned partners in consulting. Perhaps So what's your support and implementation model. How how do you support this.
| 33:33.87 | Asbjørn Hansen| Well at the moment since we we're still very small right? We do most of implementations ourselves we have started involving various partners and trying train up their resources to do from cha and for us because our strategy has always been that we don't want to. We don't to build a big education practice right? That's not what we want to do we want to focus on the software. Um, however, as we. This our fourth customer now right? So you know as we as we grow we we tried to train our people for them to go with Jason code forward right? Um, and that that's what we but we do there right on the support side. We've built in a lot of stuff into into our stack right around using application insights and other tools basicallyactively defined problems. And sort of dressed them before customers see them right? Um, that that's ah, that's what we're trying to do there and again we can do that because we have kind of live view of what's going on right? It's all running in the cloud something we couldn't do had a bit on-premise right? We had to wait and then. The customer like a ticket and to go and let and last we might not have the data and yeah, all the sort of stuff that creates slow resolution of problem.
| 34:34.80 | chrissass | And so you're proactively trying to pre pre-identify a problem and address it. So is there uptime or something associated with the the histories. So you've been around for some period of time you've got a number of customers on your system. How does that translate to performance in uptime.
| 34:50.96 | Asbjørn Hansen | Um, well the only time anything's been down is if Microsoft had a problem with their active directory which they had like once earlier this year right where where ah basically the entire sation directory was down and of course in that case, we also our system was actually not down just nobody could get to it which is in a itself a problem right. And but other than that we we actually don't see downside. Really we see downtime still with deployments when we deploy new coke there are cases still where it takes us a bit too long to for new code. We have to that customers know that that have to be be be be down for a little bit but we're working on that as well to get it to with a vision to get it to a continuous release with no downtime whatsoever right. Um, that's an what we we trying to achieve right? so that we can can do that without having a downtime for the customers at all and you need to have a continuous trading right? so.
| 35:37.47 | Martin | Okay, um, maybe back again to the topic integration roadmap that Chris just mentioned what I'm wondering is and what I can also see in in project that I'm on is that usually when you start integrating a new. Solution especially Cloud Native solution. Um, you need to think well about the whole architectural design and make architectural decisions So which systems to connect where you need to build interfaces, etc, etc. So is that something that you leave to the client or is that something you really. Can assist also with your know-how and and work out architectural designs with him.
| 36:20.52 | Asbjørn Hansen | We can only support it right? We we need. We need to doubtut architects and things to to do that right? and it depends on the customer infrastructure and and the appetite and so on so forth right? So for that we work with partners right? We we don't we. We're not the we're not going to go in and and take over customers infrastructure and run that right? we.
| 36:36.98 | Martin | Um.
| 36:38.92 | Asbjørn Hansen | Not experts in that we expert on on building ether M software not a Cloud generate Cloud access across multiple system right? So typically we would try to find a partner to help help do that for the customer rather do that ourselves I think other people can do that way better than we can.
| 36:52.87 | Martin | Okay, chris something to add on the whole topic ecosystems in Cloud from your point of view if not I have ah another interesting topic that we might talk about okay, great.
| 37:06.46 | chrissass | Let's move on I think we've got it covered.
| 37:09.39 | Martin | So aspijoon automation you have mentioned it already at the beginning. That's definitely something that drives the industry josumm mentions some some buwords like Agoe trading I think you heard out of that. So um, especially in terms of algoe trading. Let's start with that. Maybe. You provide really some apps or some functionality that enables the client to maybe build own strategies to run them. Maybe you can elaborate a bit more on that.
| 37:38.39 | Asbjørn Hansen | Yeah, well so again, like Elgos is typical something you don't you don't want to share your elbowgos with anybody else typically right? So that's something that for us becomes a kind of private app right? because that's your your main moneymaker right? for those who are referr right? So what we try to do is is provide the data as quickly as possible and have.
| 37:44.70 | Martin | Yeah.
| 37:58.17 | Asbjørn Hansen | Let the customer subscribe to trades and positions as they change it as they come in so rather than you having to go and run a process every time we we build an event-based solution which basically whenever a new trade hits the Api. The main position part of that gets put onto a service Bo Essentially. Ah, you can listen to these events and get the positions. You know, pulled or pushed into your your real-time solution which then triggers the algo to recalcate and go we put out new orders right? Um, So so we built it that way intentionally. So we could get as much performance as possible to the Al goess right? So while we don't. Write Als herself and and never will right again. I said before that's not something we we great at right? I'll be very much better at that and typically people wont don want to share it anyways. So We want to provide the position so quickly that the al go could get an advantage and go out and create new orders and new things based on new Signal. We got right.
| 38:49.22 | Martin | To what it.
| 38:49.47 | chrissass | So you mentioned if 500 millisecond I think we're like a half a second maybe is that what you're saying your settlement time was so for a human that sounds quick that still sounds incredibly slow to me for a computer and processing what what takes so long.
| 38:54.11 | Asbjørn Hansen | Yeah, yeah, 500 been 6 Yeah, how was it again.
| 39:04.13 | Asbjørn Hansen | Well, if you go through it right? You the first thing is you need to just validate that this trade is actually valid and the data you got is valid right? So you have to validate at least the position relevant part of the trade right? So the way we bere doing. It is basically breaking in bringing it out to 2 different path right? before we set out the event. 1 part is the position relevant field like we go which which timef frameme isn't for what's the quantity price those sort of things right? That's what's important for our position right? We'll go at and at validate that first when that's validated we will persist the trade we'll then send out an event which is the which is the ah position relevant part right? goes into the engine. Then gets valued against the current price curve right? which might also be real-time and if you won look put 12 this together. You did then have a second this is possible. We can get to at home with with with the stack you built so far I actually don't think anybody can do it fast than us. Anyways.
| 39:48.37 | chrissass | So how does that stack up in the market. So what? what is what are other people able to do is a half a second something you're you're very proud of and you're wearing it out front saying that's really fast.
| 39:56.38 | Asbjørn Hansen| We we pretty proud that. Yeah, but yeah, we we we pretty proud of that. Yes, it's pretty good exit right? there for for our point of view. It's it's pretty pretty nice and I said before we're not quite there yet. We are getting very close to it. But that's at least the Target we we set out to be from the beginning right? Um, so we're quite there yet. But we're getting close with them and I don't think anybody's. Nowhere near right? So so we worked with 1 of our customers who who spent a lot of time before they selected us to look at everybody else out there and basically their finding was that they couldn't really find any classic htm systems which would get them below. For five seconds unless they put a lot of weight infrastructure behind it right? So a lot of database servers specially and and and architecture and of course the cost of running that becomes very very high compared to a solution and like ours which is much cheaper to run because it all wants on serverless architecture like um, so so yeah, it is something we we think would be a big advantage for us and. We actually go between to start posting average numbers from for production at some point next year right so we can actually use that as as ah, as information for the market.
| 40:57.21 | Martin | Okay, in terms of automation. What I can see is crucially important, especially in terms of short-term trading you need to have near real time position management features So as a trade I Want to see what is my. Ah, position in any hour. Any fifty minute any market Any product, etc and I heard from you event-driven architecture. So It sounds like you have all the the basic ingredients to to have a realtime position Management feature. But maybe tell us more about that.
| 41:18.63 | Asbjørn Hansen | Yeah, me.
| 41:27.79 | Asbjørn Hansen | And yeah, what we did was we we embedded streaming analytics again from from Microsoft into our software right? So that you know while the algos can go and listen to the same stream afraid right? we can of course. We also build a solution to to do that and show the position inside the application byte which basically streaming analytics. It gets each new event gets pushed into that engine right? Update the position realtime. Um, it basically gives gift. You gives you a position on the screen as quickly as the alco get it pretty much. Um. So so that yeah that we'd we done that to to provide that sort of feature to the customers. Good inside graphs I basically.
| 42:04.85 | Martin | Okay, understood and kept yeah talking about the Al goes again I mean it makes a lot of sense that you wouldn't really share strategies or build strategies for the client that you share then but are you planning maybe on your roadmap to build. Certain Development Frameworks purely for strategy development or back testing framework for back testing strategies at least I can see that the market in some areas is really moving towards building more and more own strategies but in a structured and sustainable way.
| 42:39.90 | Asbjørn Hansen | Yeah, but actually again we would provide Framework to allow people to build their own things right? but we don't want to build it right? We much rather want specialist companies build these sort of features and functions right? um to basically ba the you know ba bar models or baptist the algo or whatever it is right? um.
| 42:54.81 | Martin | Um, so so you would leave the the building the back testing Framework also to the client not do the back testing itself. But the framework also okay understood. Yeah.
| 42:58.86 | Asbjørn Hansen | We don't think we have the right people to do that. So we think other can do that is that.
| 43:04.80 | Asbjørn Hansen | Yes, great and anything we can build it into our framework that helps multiple people will do right? but we don't want to build a specific framework for this. It's little because people use different stacks. Well right? Some people do P or just whatever.
| 43:14.65 | Martin | Yeah, okay.
| 43:23.20 | Asbjørn Hansen | Try to use different other tools to do it right? And it's it's very hard to build something that will be 1 size fits all I believe right? so.
| 43:28.80 | Martin | No yeah, understood maybe then from the whole topic automation from a client's point of view I mean I heard you talking about client engagement in that in that Context. So What is from your point of View. The biggest challenge when it comes to automation Now. From a client point of View. Maybe not from a technology point of view but more from a market point of view.
| 43:51.74 | Asbjørn Hansen | And defense of the client right? Um, well the the biggest challenge in terms of how you get to get it implemented or how you how you do it but because of course you could do different ways if you have a legy system you need to use robotics or something like that to put on top to automate things. But if you go with more 1 solution. You expect that to come out of the box right? So so kind of the solution we we'd be doing at the moment with with our latest customers basically um in today they hit power right? Where where you want to set you create the whole matching settlement matching invoicing stuff against the exchange fully automatically right. So basically trades come in every day they get pushed over to the position engine trades get set matched up automatically and and and create inbus are created and send itscp wire by an automating interface right? So basically if there are no exceptions then are humans involved at all. Um, that's the whole point of ah. Doing that right and getting getting to that sort level. Of course that's a humative bit of effort but people are afraid of producing their jobs right on the other hand those resources can be utilized in way more value value added places right? And if you look at some of the the bigger companies right? There's big differences in terms of how many people you need in back office. Ah, depending on who you are with systems you have in place right? That's extreme gaps right? or extreme differences right? right? from even if you trade some some number of tradeits just because you you don't have modern systems in place to to do things.
| 45:18.41 | chrissass | So we've we've gone a pretty full journey about bits your technology What you said early in the conversation is you started at the edge and you've kind of backfilled into more legacy features and functionality. So where are you in that journey. How how complete is your system.
| 45:34.31 | Asbjørn Hansen| Well, it's getting getting pretty good right? so um so we've take the geo side I think there we've we've we've pretty much feature complete for what we want there's still a few things we want to do around geo matching to improve that further. Um on the power side. We. We pretty much have everything we need. There. We just went live with ftrs and and strip spread options actually last week um so that was that was good. There was a new new thing we wanted to put in right? Um, other than that we would go to new market things. There'd be more things we un do there but other that it's pretty much proxielity is there for what needs to be right on the gas side. Um. Then we still have a bit of stuff to do around valuation around unbund capacity right? So we can you have ununded or 3 hree-way orders where you have multiple ways you can go with the capacity. There's still some stuff to do there. There's still stuff to do out gas storage. Um, but beyond that it would be to a to relatively to be good place. So I think we'll be. Kind of where we want to be towards next summer then we should be pretty much feed feed feature complete for for most of the up and market for power and gas and and the certificates. Ah as they look currently right Um, and then from there we will start looking at us functionality and those things um next year
| 46:46.70 | chrissass | Awesome! So Martin before we go into our next section. The speed round. Do you have any final questions.
| 46:51.96 | Martin | Yes I just heard you saying 1 comment us market. So so far I would have assumed that your footprint is mainly in in europe which might be true or not. But maybe you can tell us more about your location. Um, yeah targets.
| 47:01.44 | Asbjørn Hansen| It is well do do you do a side of things is is basically yeah, we be getting pulled there by customers right? So so basically yeah base. Yeah, you know that's how it is right? So ah so some of our existing europeanp customers.
| 47:13.88 | Martin | That's nice.
| 47:20.99 | Asbjørn Hansen | But already have business in the us or will'll start trading in us and of course when that happens then the need folks enter in the us as well, right? So we kind of been being driven that way we want to do it anyways right? This just happens a little bit bit sooner than then be planned like like many things happens start that way. so um so ah so yeah that that's really the the driver right is basically being able to supply Fox answer to existing customers and of course later on then start looking at the building building most of the us. Well right? And but yeah, that's what that's what forgetdding us there.
| 47:48.41 | Martin | Um, okay thank you.
| 47:51.20 | chrissass | Awesome! Well I think it based on the time we're going to switch gears and go into our speed round for the audience who've been listening to this series so far this is where we ask ebby bender the same 10 questions. We're going to alternate back and forth between martin and I and they're just short answers just to. Answer what they are so just to give you a feel I'll go ahead and kick off and then we'll just dive in so will the number of vendors for each year m solutions shrink or expand in the future in your opinion. Maybe you can expound on why you think that.
| 48:17.84 | Asbjørn Hansen| Expand well because think the the the product coverage will also expand right? I think as you get into more and more things you you you call eacher eacherre right? There'll be more and moremoities there'll beucod' like hydrogen right? You have device trading. Go up, you'll have way more more needs for these sort of products right down to to smaller customer level right? so so I think there's a need across the Board. You see that from c o 2 it. What? what? right? All industrial suddenly have to have some sort of solution to maintain their suor quoers and the ges since so of forthright and of course for that reason the demand will be bigger and because the demand is bigger. But also be vendors.
| 48:57.19 | chrissass | Okay, thank.
| 48:57.34 | Martin | Okay, second question how many ds per minute must the modern etm system today. Be capable to import.
| 49:09.88 | Asbjørn Hansen | Ah, 10000 nancy
| 49:12.24 | Martin | Okay.
| 49:14.29 | chrissass | Okay, what commodity types and energy types are you offering and which ones would you say is your biggest strength.
| 49:21.62 | Asbjørn Hansen | Yeah, what we offer ges eas ah and power and and gas right? I Think the biggest race is the Ge Os Um, because that's where we started right? Um, yeah, ah, that's the short heads are least. Yeah.
| 49:35.10 | Martin | Yeah, okay, do you offer a real time position Management Module and in case, yes and does that work real time So event driven.
| 49:35.24 | chrissass | Okay.
| 49:47.37 | Asbjørn Hansen | A zip vote. Yes.
| 49:50.75 | Martin | Okay.
| 49:53.50 | chrissass | Okay, do you have an automated workflow for straight through processing.
| 49:58.40 | Asbjørn Hansen | Yeah, we do and we do. We have a whole workflow model that will build and again eventriven right? So Ah so depends on when you but you know there's a trade status workflow where you can customize your own stuff or you can utilize the eventriven framework to build build morefoxelity that you want to build right? and of course all to be able to integrate different workflow right. You might, but but still have a you know nomination workflow a regular to of reporting workflow a confirmation workflow and so on to forth that you want to integrate and and build build across the different workflow fact. So that's box after the box. Yeah.
| 50:28.78 | Martin | Okay, fantastic. Can your etm solution or you um, 1 of your modules prize asian options.
| 50:37.00 | Asbjørn Hansen | A our solution cannot but you can buy an app. Um, so there is a place in the code where you can hook in your app to value that typically I would probably use Tur waconn something like that to value that right I trade so these thingss in the past but but but yeah, there is ah the ability to choose.
| 50:52.82 | Martin | Wow We are.
| 50:56.52 | Asbjørn Hansen | Utilize ah utilize an option model to to do that vi an app.
| 50:58.28 | Martin | So even mentioned the ah the model Termbu wake men. Wow Okay, cool. Thanks.
| 51:05.47 | Asbjørn Hansen | There's an error in the original model. But yeah.
| 51:07.75 | chrissass | All right? So are you offering full integration and implementation services for your each youram solution.
| 51:14.77 | Asbjørn Hansen | We are. We are um but we try to get partner to do it as well right to to expand the footprint and and make sure that customers are fully dependent on us right? I think 1 of our mission missions always been to create vendor dependence right? We don't want people to just be redep dependent on us. And so we also want partners to provide the sort of service. We have some partners already who's who's doing that to a certain extent.
| 51:36.20 | Martin | Okay, now come into question where some of our guests have to think a little bit longer. Um, so what is the biggest threat from your point of view for third party etm providers if there is any threat.
| 51:52.94 | Asbjørn Hansen | The biggest threat I think actually the biggest threat at the moment I think the worries be the most we take it that way is is really I security things right? Ah, at least we've seen that future recently in this industry where people have been hit by ah by. Attacks right? there as so best as the other week was so last week was down and then my right ah reli or somewhere that kissed us was having a lot of problems with with cyber tax right? Which of course is is terrible. So so at least to me at the moment. That's the biggestest rate. That's the 1 that scares me the most right? Of course we do everything we can. We have fix certain security companies to help us with that sort of stuff right? but. Still that's something that that's terrible for anybody I really feel feel sorry for the the companies you get gets hurt by.
| 52:31.39 | Martin | Yeah, ah, but maybe to elaborate a bit more on that because that really triggers my interest and I would actually assume this is a threat then for for the your uses. So The the company's using your system. But why do you think is ah is it a threat for. You and the party Etam providers.
| 52:51.10 | Asbjørn Hansen | Well, it's fit for us right? because if our system suddenly doesn't work that kills our business right? You know the the reputational damage of something that that can have is is is very high right? Um, so um, so yeah.
| 52:57.44 | Martin | Okay, okay, no thanks.
| 53:04.36 | chrissass | Cool. So how does your licensing model look today and is it likely to evolve or change in the future.
| 53:12.90 | Asbjørn Hansen | Yeah, Well we we try to do do things new all the time right? But but yeah at the moment it's all ah you know per user license basically on a on a monthly or quarterly whatever subscription basis right? Um, That's how we do at the moment if we might well change right depending on what customers want and how we we add more Value. We. Have models where we look at how many users locked in for example, rather than just ah, the use a model like it really depends on and the size of the customer and and which sort of business they're in but um, will that change. Maybe um, but let let's see what what comes out and what the needs are from the customer side right? because I think. Looking forward, You'd have less users essentially right? because you have more automation going on so you have the issues with the software. So either. You need a higher per user price or you need to basically have a enterprise type licenses insteadpoint.
| 54:03.11 | Martin | Okay, great. So last question are you offering an api and what is the key technology behind it.
| 54:12.90 | Asbjørn Hansen| Ah, be Mus we we had everything so apis right? So everything is built by but in rest apis right? Um, so it's it's basically with Jason payload um, across the board right? That's technology behind this is that your function a certainency sharp right? So ah, um.
| 54:18.70 | Martin | M.
| 54:29.00 | Asbjørn Hansen | But standard stuff is essential to be all built on and the max perform.
| 54:30.85 | Martin | Um, okay, fantastic.
| 54:32.93 | chrissass | Well cool. So we've we've been through the segments of the show. Um, perhaps you want to take just a moment to bring it all together. You know, give you a minute or 2 to kind of summarize up what you would would love our audience to know about provide systems.
| 54:47.23 | Asbjørn Hansen | Well I think and kind of as it as a summer right? We we be kind of a young young company right? with but with the lot experience right? We've been around this market for a long time and and our mission here is to build better and cheaper software that adds more value to customers right? Ah I always say to people that you know the The only thing that makes people pay us is we if we ensure that they make money right? So so our job is to ensure that the customers get a commitment to edge and that they can make money themselves because of course if they do that they are more likely to pay us our fees right? So so that's that's our view of the world is what we tried to do right? And and that's mentioned before they there's a few building blocks in terms of. Form usability and and and the whole framework around ecosystem might that that's basically why we we structure our software from from the get go right? So yeah.
| 55:35.52 | chrissass | Well thank you so much for participating. Thank you for being on the show today and thank you for for our audience this concludes this episode of insider's guide to energy etmm educational minieries if you haven't been tracking the series there.
| 55:39.63 | Asbjørn Hansen | Um, thank you very much that's be it.
| 55:42.10 | Martin | Thank you aspian.
| 55:53.61 | chrissass | Episodes with many vendors to listen to go ahead and dive in look at us at insider's guide to energy dot com you'll find a page which has all the vendors listed you can compare and contrast and hear what every vendor has to say we look forward to seeing you on another episode and thank you for your time today.