Most companies developing IoT initiatives are aware that connectivity doesn’t actually add a great deal to their bottom line. Yet, at the same time, many CSPs are struggling to expand beyond connectivity.

One company is bucking the trend. Telia’s Division X aren’t only exceeding industry averages for IoT connectivity revenue, they’ve also opened up non-connectivity revenue streams that account for an incredible 40% of their income. And if that’s not impressive enough, they’re experiencing a 20% year on year revenue growth.

For our new Accelerators podcast series we meet Division X’s Björn Hansen, Head of IoT, and Kristofer Ågren, Head of Data Analytics, to learn more about the secret of their success. They’re on a mission to explore ‘radical new ways’ of growing  IoT revenue, and it’s fair to say that they’ve done more than succeed. Stream the episode now to find out more.


“The ambition is to have a 10x improvement for the customers. Division X should eventually come up to 10% of the corporate revenue of Telia. And being separate is giving us the opportunity to develop new business models, without being restrained by traditional financial KPIs.”

Björn Hansen, Head of Enterprise & Public IoT


Speaker 1: Accelerators from Beyond Now. Hello And welcome to Accelerators by Beyond Now. Join us as we speak with industry leaders, and explore the big opportunities ahead in 5G, IoT, AI and cloud, and the role of the ecosystem. We discuss how to stay ahead, and what technologies innovation and business models are driving the industry to accelerate.

Michal Harris: Hi, I'm Michal Harris, Head of Marketing at Beyond Now. I'm here with our host Jeremy Cowan, co-founder of IoT Now and VanillaPlus, and today we will discuss how IoT is shaping the world outside of connectivity. We are here with our two special guests, both leaders at Telia's Division X. Björn Hanssen, the Head of IoT and Kristofer Ågren, Head of Data Insights. Björn, Kristofer, I'm so glad you're both here. And I'll now pass it over to our host, Jeremy.

Jeremy Cowan: Well, it's great to be able to welcome Björn Hanssen and Kristofer Ågren today. And Björn welcome. First of all, very good to have you here.

Björn Hansen: Thank you very much. Nice being here.

Jeremy Cowan: It's good to have you and Kristofer, a warm welcome to you too.

Kristofer Ågren: Thank you so much. Great to be here.

Jeremy Cowan: What I'd like to do, is understand perhaps a little rewind. Understand exactly what is meant by Division X, and what the goals and intentions are within Telia. Could you tell us, Björn first, a little bit about what exactly Division X aims to do, and how it works?

Björn Hansen: Happy to do so. It was established a couple of years back, to spearhead the creation of a new generation Telco, as they close to our core and explore and build and commercialize, and scale future business opportunities. And where we have a right to win based on our resources and capabilities. And we don't want to put the money into interesting explorations, unless there is no path to scale. And the ambition is always to have a 10 X improvement for the customers, and that that is being catered in Division X, Should then eventually come up to 10% of the corporate value, or corporate revenue of Telia. And having a separate is giving us the opportunity down to develop new business models. Without necessarily being restrained by the current mode of operations, or that in a traditional established operations, financial cape guys on a quarterly basis.

Jeremy Cowan: Kristofer, could you add anything to that from your own perspective?

Kristofer Ågren: Sure. So I head up one of the business areas in Division X, called Data Insights, and we are responsible for Telia's external data monetization agenda. So we offer insights products for our customers to help them become more data-driven, by providing insights, rather than just data streams. So where we have come the furthest is where we create products from our own anonymous and aggregated mobility data. For example; helping municipalities with city planning, retailers optimize store locations. But we are also now working more and more with insights, on top of also other IoT data sources.

Jeremy Cowan: And Björn the typical statistic that gets trotted out at times like this. When journalists are asking questions, is that only 5% of revenue in IoT comes from connectivity? What's the case in Division X? And I guess the very existence of Division X suggests that Telia has found significant other opportunities outside of that?

Björn Hansen: Yeah, true that. Within Division X, 40% of the revenue generated is coming from a non connectivity, traditional net revenue stream. So 60% then is connectivity related.

Jeremy Cowan: That's an extraordinary amount. And obviously delivering considerably to the company, how big is this within Telia, now?

Björn Hansen: We are not at the 10X by no means. We are in the 1 billion Swedish Kraus. So it's still yet to grow. But with the pace that Kristofer and other are having, we are still aiming for the 10 X. But that's a three-to-five year horizon on that.

Jeremy Cowan: Kristofer, could you to tell us a little bit about that growth? Björn's already referred to it. But it sounds fairly rapid. It must be putting considerable pressures on you, not least of all in a pandemic? How are you dealing with that? And what are the experiences as you grow?

Kristofer Ågren: Super relevant question, on one that we end to crack every single day. It's how do we get to scale? And then we have been able to do so now year on year, for at least a couple of years, when it comes to the insights. We tend to look at this from all the angles. I mean, solving the scale issue is very much about solving the same problem that many customers have, and by doing so in a standardized manner. And buying insights service in a standardized manner, I don't think it's commonplace today. So this is something where we have been quite early, I think. So when we look at, for example, offerings to municipalities. It's a fully standardized offering. It's addressing specific use cases to a municipality, like city planning or events and tourism analysis. And then we surface that in a standard digital offering standard price lists. If we have a 100 municipalities or one, that's pretty much the same effort on our end. So we're building everything from scale, from the beginning. Whether it's the engineering, the product thinking, pricing, delivery, go to market support. That's where we're starting.

Jeremy Cowan: And what'd been the response of your customers. How do they view this?

Kristofer Ågren: That's a good question. And I think our customers, you will get perhaps not a homogeneous answer. I would say all customers when we talk about insights and what we can do, are really interested. It's really typically fun meetings. Where our customers understand that, "Wow! We can do something that we didn't know was available." Then the hard thing gets into, "Well, what are the problems that you're solving now with these insights?"

And then what we're trying to do with them is of course, to help with that. Typically, you will solve X, Y and Z problem. We have found from working with other municipalities, for example. That you can address these issues. And it's easy to get started. It's a subscription service. You don't have to install anything. Try it out. And we're a big fan of actually making it easy for our customers to buy, but also to leave us. That's the name of the game, when it comes to subscriptions. So we need to provide value all the time. And I think that's kind of the key here. And that's what our customers are appreciating as well.

Jeremy Cowan: Björn, there's a lot of talk in the industry these days about smartness; smartness of devices, smartness of services. What exactly is your experience of this? What do you take smartness to mean in your own situation?

Björn Hansen: We tried to turn it around a little bit. Because technology can also be associated quite faster with smartness. But it's really trying to turn the discussion completely 180 degrees and talk about based on their experience, and having interviewed over 750 companies. Whom have embarked on their digitalization journey here in the Nordic and Baltics is really to say. What is it you guys are looking to achieve? And what requirements does that have on your operations? Not just from a technology perspective, but also from a process and operations perspective, in order to capitalize on saying, take something very simple. If I can tell you where your track is, and how far it is to reaching its destination. What should then happen in your operations?

Further down Into the chain, in order to capitalize on these insights that you get to real time. And then we have divided it based on the experience we have done, and also the customers who's embarked on this is  planning to; A, as so either being more efficient with your essence or your processes, or taking new innovations to the market. And the separation between the two is that, taking new stuff to the market requires more from the organization in terms of implementation, and I'm not talking technical but from an operational perspective. And that's a long answer. How I would describe smartness, if you can combine those two, then you achieve a true smartness. And hopefully then can roll out more of the digitalization of choice.

Jeremy Cowan: Kristofer, did you have something to add to that?

Kristofer Ågren: I think Björn explains it well. From my end, which is coming from the insight's angle, smartness is only something that's smart. If it actually solves a real problem, that Björn was also getting into. And what we find really is that it's such a relative term. I mean, in many cases, technology and data is not even used in making decisions. Take heating of a building, for example. This is often managed in a very traditional way. Even though weather data has been around for a long time. But it's not used. So start there and then also add IoT sensor data, for example. And you can add machine learning. But if you're not working smart than just adding some basics, that's also smartness.

Jeremy Cowan: I mean, in the example you use. Can it be allied with attendance data so that, when there is a higher proportion of staff available in a building, it leads to less heating or more heating?

Kristofer Ågren: That's one example. I mean, when you optimize heating in a building, for example. You need to look at things like ventilation, and also what you're optimizing for. So if you go ask a customer, "Well, what should we optimize for?" "Well, we need to have a nice indoor climate." "Okay. Well, what does that mean?" "Well, that's usually 22 degrees centigrade." "Okay. Do you need that also in the hallways?" "No." Okay. Well, we can be smart even without applying the advanced stuff. But of course, to get to the big savings. If he wants to talk 10, 20% energy savings. Then you often need more data sources. The ones that you've mentioned, Jeremy. And then a little bit of machine learning, typically on top of that.

Jeremy Cowan: Björn, I love talking about applications in the real world, the life in practice in IoT. I mean, when you're talking to your customers, I'm guessing that almost every customer situation is different. So how do you have a standard offering? Can you indeed have a standard offering? Or is it something that you have to tailor every single time you start work with a new customer?

Björn Hansen: Coming back to what we talk about, customer needs and what's really drives th digitalization stopping there. I think the answer is no. In the sense that, we build this architecturall horizontal. But we need to sell and understand in the vertical. Without getting trapped than in building a community solutions. Because nor we, nor the customer want that. But that will be something that has to have a lifecycle management down the road. And you want to be building it on standard components as much as possible. So build horizontally, sell vertically is a short answer, Jeremy to that question. And it's easier said than done. But that's how we tried to run it.

Jeremy Cowan: I going to try and remember that. That's a very neat answer. When you're breaking this problem down. I mean, presumably you're encountering pain points all along the way. And they may not be the same pain points that you felt when working with another customer. Kristofer, is it easy to avoid talking about technologies? Customers after all they're not terribly interested in the technologies they want the solutions?

Kristofer Ågren: Well, it depends a little bit on who on the customer side that you're actually talking to. I mean, if you discussing with, for example, an IT Department that already, what they do is buy tech. So then you do have a technical discussion about. But when it comes to insights, usually what we find is that the buyer of insights. So it's like energy optimization or city planning. Are not the people in CT and IT, but rather the ones operating that part of the organization. So, one of the key questions when we work with this is really, as I mentioned earlier, solving a problem that many have. And then you naturally get into the use case discussion.

If you talk to a customer and you can explain, well, what are others in your vertical? What are competitors doing? What kind of use cases are they solving with the help of data and insights? For example. Then normally you get into discussion that's very far from tech. And there's really no shortcuts to get there. For the verticals we play in, we will also work with partners who know those verticals deeply. I don't pretend that Telia is a deep vertical player in everything we do. So here we rely also on insights from our partners. And we use user experience methods to gain the insights needed. We listened to our customers to let them steer the product development, as much as possible.

Jeremy Cowan: Björn, somebody once said to me, and it stuck with me that CEOs, that your customers, never leap out of bed in the morning saying, "I need an IoT solution." What they need is a solution to their problem. When you're bringing them an IoT solution, are you conscious that there may be other things in their mind, and indeed maybe other solutions that you can offer them down the road?

Björn Hansen: They're definitely, and that comes a little bit back Jeremy, to your question that you just raise before. It has a tendency way too fast, become a technical question. Because those on the customer side, there technical people that want to show that they understand a certain technology. And there are engineers on our side that is ready to build it for you. But then you ask the question, what is really your pain that you're trying to solve? And if we can agree on that, coming back to Kristofer's comparison about the indoor climate in the corridor, versus in the actual building. Or the room where people then socialize for the larger parts of the day.

You have to come and pause the technical discussion, to really get back into that one. Because there's a trap easy for us to fall into, as an engineering company. When we meet the Digitalization Officer of Choice. Because every company today have a Digitalization Officer of Choice. And he or she has a small pocket of money, just to spearhead innovation in their organization. And say, "Hey we're looking at this cool digitalization projects." Okay, but is there real pain attached to it? Is there a potential business value at the end of the rainbow? And if it's not, don't go there. Because then you will get stuck in pork swamps.

Jeremy Cowan: Kristofer, there seems to be a certain amount of mystique around the subject of artificial intelligence these days. And people use it almost as though it's going to solve all problems. It is after all, as I know you said before, it's a tool. How are you using it today? And how do you anticipate using it in the near future?

Kristofer Ågren: Well to us, as you mentioned, AI is just quite simply a set of known tools and capabilities stuff that we can use. And then when we talk about AI, I mean, AI generally is machine learning. Which is a discipline that has existed for a while. And it's simply the ability to make an algorithm that learns from looking at data. Rather than you have a program of putting in different rules in advance in your system. So we use it in a few different ways. We use it both internally for us to get scale and ultimation. So for example; we're in the business of insights. We process billions of data points every day. Our customers rely on our insights to take decisions every day, on those data points.

So that means we need to assure that even though we have these billions coming in, as avalanche of data coming in every day. We still need to check it. We still need to make sure that it's meeting our data quality standards. Well with that amount, I can't simply have a team looking at it. And it's not possible to devise rules around it. Because data quality, as anyone who has worked with data quality would say is, that you find new things every day. So here we use AI methods to detect when something is wrong. And it helps us to scale internally. So I don't have to have a team that's manually checking these... don't have a fully rules-based approach. It's just too much data on too many variables to look for.

And then externally, of course we provide insights. So sometimes that is the result of an algorithm that, that is the insight. You need to have an algorithm that looks at these different data points. Makes a prediction or some sort of inference. But what I think often is missed in the public narrative is that you also need to tread carefully. Because when you provide an insight, to say that you should make decision A instead of B. Customers are going to know why. Why is the algorithm recommending A or B? Or even in order to feel safe to make that decision, how can I then know why it's recommending that?

So you need something called explainability. Which is a big area in artificial intelligence. And that's really understanding what variables are influencing, and why is it saying A or B or X or Y. And that's really challenging when you work with cutting edge AI, like deep learning, for example. That's a big research area, trying to figure out how to have explainable AI models. So more often than not the simpler algorithms or even simple descriptive analytics does the job. And is that AI? Well, it depends on how wide your definition is. And then of course we have partners applying algorithms also on top of our IoT strips.

Jeremy Cowan: I've heard reference to synthetic data generation. Kristofer, what is that? And how is it shaping the projects you work on?

Kristofer Ågren: It's a super interesting space. And that's really enabled by cutting edge AI, a sets of algorithms called generative models. Of where deep learning can look at massive amounts of data, Learn the patterns in that data. And then this algorithm will instead create a synthetic data set that has the same statistical properties. But it's effectively a new dataset. So in a sense, you could share the patterns of your data, but not your data itself. And this is a early area, a lot of research in this space. But it's a really interesting way. I think this could be a big thing going forward. As a way to expose data or patterns in the data. Because really that's what you're interested in. You're rarely interested in the raw data, you want to get the patterns.

But you can expose that in the way you otherwise would not be able to, for example. For legal privacy reasons or commercial reasons, competitive reasons. You might not want to share your raw data. Because it could give your competitor an edge. But you might want to combine the patterns from multiple sources. So this is a scenario where we're looking a bit as well as it really enables the sharing of data or patterns in data in a very new, exciting way. But this is not where I think I see big revenue streams in the near term. But really a very important capability in the future.

Jeremy Cowan: Björn, looking at the delivery of IoT services, generally. I think most companies tend to view that this cannot be done by them and them alone. That you need to be working with a partner ecosystem. And to make it fly as efficiently as it should, you need to be optimizing that partner ecosystem? How does Telia utilize the ecosystem in the IoT process?

Björn Hansen: Being a Telco entering in through Division X, this arena. We build our technical stack horizontally. And open up then a set of interfaces or APIs than to vertical experts. We don't have to build everything ourselves. But in the respective articles, there are solutions today available. And either the customer themselves have a certain preference, or we know all of those vertical solutions. That can be put on top of our platform. And then coming back to your question, if you take the building as an example. Most part of the building is connected today. But the system A connected to the elevators has nothing to do with the security solution, that puts the alarm in the building.

Whereas, the horizontal platform in the background. They can share similar data. And then you can put the level of priority on the different cases. There's not a problem, if the elevators still or standing still, and there's no people in the building. Then the technicians can come out at a regular place. Whereas, if there are people in the building and it's crowded, all elevators has to work. So we tried just to say that, "Don't be afraid. Nobody will be able to do this on their own." It's natural that we do create the partnership. But do it based on a horizontal common view on how to digitalize a building, or a bus or a city, for that matter. So that you don't build stovepipes.

Jeremy Cowan: Michal, can I bring you in here on this aspect of the ecosystem? What's your view on what Bjorn's just been saying?

Michal Harris: Absolutely agree. I think what Telia is doing is very clever. A lot of our customers are talking about how to engage with an ecosystem. It's very clear that nobody can do it alone these days. An ecosystem is a critical part of it. Think what is also interesting in the case of Telia, that they went forward and also created a lot of the solutions themselves, like the insight bits. Which again, some of our customers are trying to find partners even for this bit. I'd like also to ask maybe Björn, there is a lot of conversation right now around 5G, and how 5G will impact IoT. Can you please share your thoughts?

Björn Hansen: Yeah. It actually has already started. Because 5G is all about machines talking to machines. But 5G opens up for a high device density. Meaning we can connect a lot more things in the same area. It also allows for higher data rates, with a low latency. And those three things in itself is what differs 5G, if you like from 4G. And that will enable us then to have a million connections per square kilometer. Meaning parking sensors, so street lights, road sensors. Where you can have real time data towards the Smart sheet to feed. And that's also a Kristofer's area of analytics, becomes even more crucial and an integrated part thereof.

But the 5G journey, if you look at the business logic. Where we start the discussion, has already commenced with the current solutions that we're rolling out around IoT. And we do that based on low power wide area networks. That is by its nature, very similar in behaving like a 5G network. But they don't have the same device density or high data rates. And the latest is a little bit slower than the 5G networks. So we have stopped it.

It will only go faster and more devices and more data will be analyzed on top. But it will take a bit time before we see the real business and revenues generated. I mean, there's a lot of customers already now in confined areas, such as harbors here in the Nordics. Which we have connected to 5G. Where you have self serving trucks and lorries, et cetera. And we are stopping that. But before we see a self service society, it will take a couple of years. But in confined areas, the business model is there already today.

Jeremy Cowan: Björn, Are those private mobile 5G networks?

Björn Hansen: Yes.

Jeremy Cowan: Only private networks. Yeah.

Björn Hansen: Yes. So far.

Jeremy Cowan: Kristofer, the 5G story is obviously bringing fantastic new opportunities and clearly some challenges as well. What are the challenges that you find in your daily life at the moment in Data Insights?

Kristofer Ågren: Well, I'd like to start just to comment on what Michal said, also on the Telia's is doing insights themselves. We don't necessarily always want to do the algorithms ourselves. We're quite happy to orchestrate a ecosystem. And have a partners, typically. These are smaller partners, that are very niche. That have a great algorithm for particular use cases. We are happy to orchestrate that and provide a holistic solution to our customers.

And then we also get into the challenges. So sometimes the hard thing is not building the algorithm. It's actually getting a hold of the data. And it's just going to be even more interesting or more challenging, however you decide to term it, ith 5G, for example. So what we're seeing, having been ourselves in the insights business now, for a few years and also multiple countries. I mean in the Nordics and Estonia. Is that what we spend a lot of time on, is really the data quality. And how do you do insights at scale?

It's easy to run an algorithm in a controlled environment. When you release it into the wild with new data streaming in every second. I mean, we are not doing real-time analytics use cases very often. We would typically have a delay of at least a few hours of the data. And that's hard enough, making sure that you assure that the data stream is coming. What happens if a data point is missing, or multiple data points is missing? What if it's wrong? How do you manage that? How do you tie that into the insights that you're providing either in the form of a report or visualization, or it's fully automated? It feeds into the billing management system, for example. And adjust the heater in the basement. How do you manage these things? And that's about running stuff.

That's about making it work every day. And that's what we're finding. That's a hard problem to solve. Regardless, of the data source. And it's only going to be harder when you get more data points and more data sources.

And that is what I find I think is going to be differentiating in the future, is how you manage that. It's not as perhaps, sexy to talk about as AI algorithms. But it's really what makes it work in the real world. It's figuring those things out. So that's what we're spending a lot of time on. And of course, the value proposition like, what is the problem that we're trying to solve in a particular vertical? So those are the two that I think are the most interesting and challenging. And the ones we spend the most time on.

Jeremy Cowan: The last thing I wanted to ask is, if people out there are considering creating their own Division X, what are the lessons that you've learned from aligning these skills in the one division?

Björn Hansen: I can start by putting a one-liner up. Picking lemons early, I would say. Meaning we don't glorify the fail first philosophy. But we believe that you need to spot those lemons early. And that means use cases that don't have business cases. Because we have had in the past several of those. And perhaps we have been a bit too optimistic and generous to allow those use cases to thrive, without necessarily having the tannics or we started a discussion, at the end of the rainbow. And so sometimes you need to boost some egos and I say, "Hey, this is a brilliant engineering exercise, but there is no business to be had. So please, shut it down!"

Jeremy Cowan: This is a common problem across IoT. And I'm sure you're not alone in this. Kristofer anything else that your peers can learn from the growth of Division X?

Kristofer Ågren: Well, I second very much what Björn is saying. If anything, I would just add to that figure out your go-to-market before you build stuff.

Jeremy Cowan: That makes perfect sense and succinctly put. Björn Hanssen, Kristofer Ågren, thank you both very much for your insights. And with that, I'd like to hand back to Michal.

Michal Harris: Thank you, Jeremy. And special thanks to both our guests, Björn Hanssen and Kristofer Ågren. This has been a great conversation. And I'm so glad you could join us on Accelerators. Accelerators is a podcast by Beyond Now. Hosted by Jeremy Cowan and joined by me, Michal Harris. We hope today's topic is inspire you to accelerate further and faster and beyond. Be sure to subscribe to this podcast on iTunes, Spotify or Apple. This podcast is published biweekly and produced by Fox Agency.

Speaker 1: Accelerators from Beyond Now.

Subscribe to our Accelerators podcast

Get notified as soon as an episode becomes available

Get in touch