Good afternoon, everyone. We're going to change shift. We're going to talk strategy and finance. I told one of the participants here earlier today that I've finally figured out why it's 52 degrees in this hotel. Do you know why? Anyone? Keeps you awake. Keeps you awake, right. Since this is the Capital Markets Day, we need to be a bit capitalistic. What do you do when you're awake? You spend money. Gambling goes up as the temperature drops. I found that quite interesting. This is the agenda for this session, called the Capital Market session. We're going to discuss our strategy. We talked about it yesterday, and you've seen the keynotes from the divisions. We're going to have Claudio elaborate on technology, then we're going to have a short break, which is needed after a technology elaboration.
Then we're going to talk M&A finance update, then there is a summary and Q&A. Let's dive into it. We've worked hard on our mission and vision, our vision is we aspire to be the leader in creating autonomous connected ecosystems. Yesterday, we introduced to you what autonomous connected ecosystems are. You can say, "Is this corporate bullshit?" No, it's not, but it's the way we classify the future. People talk IoT, what you heard in the keynote today is IoT is just an enabler. You can't be an IoT company. It's just a facilitator of doing the things we want to achieve, which is making stuff autonomous. We believe our mission should be to deliver these digital solutions to improve productivity and quality. Here we go.
IoT in itself is not a target, but to deliver quality and productivity, i.e., cost reduction and efficiency, that is never going to go out of style. Yesterday in my keynote, I showed you the data dilemma, we could call it that. The fact that we produce more and more data we generate data, with IoT, we have machines spitting out data 24/7. What do we do with this data? Are we getting any cleverer? The answer is unfortunately no. We use more data in our everyday activities, but we don't use enough data. We're not clever enough, our systems aren't clever enough to take this new leap into a digital world.
We talked about the accelerator, the way we can keep up or catch up with the data generation possibilities that we have today, that accelerator is a combination of a subset of technologies that we have developed over the past seven years. We haven't really talked about what we're doing behind the scenes. Behind the scenes, Hexagon has built a comprehensive integration and communications module. Which means that we can tap into any legacy software. We could drill into your SAP system, your salesforce.com system, or whatever system it is you want to use. You don't direct a computer towards the problem. Artificial intelligence is useless, that's why you need the verticals to address. When you hear companies saying, "We're an AI company," that's rubbish, too.
You need cloud orchestration, and Claudio is going to speak more to that as well, and edge computing, we saw that in Jörgen's keynote, where more and more of our instruments and hardware are becoming clever, and that's needed, too, to change the flow. Once you have this, and once you master this in your own industry, you can create autonomous connected ecosystems where machines communicate with machines. Yesterday we introduced this. It's a wraparound. This is not a product. It's not even a platform. It's a subset of technologies that enable you to reach autonomous connected ecosystems. The way we operate Hexagon today is we have the innovation hub represented by Claudio, and we have the divisions that have applied capabilities directed towards the problem in a certain industry.
We've identified a set of autonomous connected ecosystems, which is our strategic mission to reach within the next four years. We want to see products in these ACEs. We want to see products in mines, construction sites and plants, factories, cities, and fleets. Another thing that's important to understand is how do we as the company create this? Well, we have six divisions, and they all have their own specialties. If we hone into, let's say, smart factories, Manufacturing Intelligence itself cannot deliver smart factories. We're actually building on the platform that PP&M has developed, and we need Geospatial technologies like Luciad to do it, and we need Geosystems technologies to deliver it as well. We can take another example, like smart construction sites and plants. That is not a Geosystems or PP&M project either.
We need to draw the technologies from Geospatial and Manufacturing Intelligence to deliver this autonomous connected ecosystem. It's all entrenched, it's all ingrained, and it all fits together very nicely. For those of you who have visited Hexagon Live over the years, you see it gradually coming together. This is the power of Hexagon. There is no other company out there that can deliver this. They can deliver subsets, but they cannot deliver the complete solution. If we look at autonomous fleets, what is an autonomous fleet? Well, autonomous fleets are autonomous vehicles that collaborate in a system. For example, in a city where you have traffic and you have buses, you have cars, you have lorries, you have all these vehicles. It's not only to master driving a car autonomously. You need a sort of map that all these cars agree upon.
You need a standard that this is the map we read from. We need to define the traffic itself, and we need car-to-car communication. What do you think an autonomous car does if it hits a red light or sees a red light? Well, it can't see red lights. We need some sort of overhead that directs the traffic and tells the traffic what to do, when to yield, when to stop, when to brake, and so on. That's the excitement that we see in autonomous fleets. There is a subset and a huge potential of services that you can deliver to autonomous systems. If we move on to the construction side of things, here we have two applications. It's plants and it's construction sites. We have, for a very long time, discussed smart plants.
Mattias did a really good presentation today describing what digitalizing the plant means. It means law and order, basically, bringing structure to these plants so that you can improve efficiency and reduce operating cost. We see a scan here of an already existing plant, and the really huge potential for us is what's called brownfield. Old plants that already exist where we don't have enough data around that plant. Auto-tagging that Mattias described would be what you see in the center, where it says Leica TruView. If we could tag each and every component automatically, and if you click on that tag, you can retrieve the supplier, when it was installed, and when it's time to replace it. This is going to bring a lot of efficiency to that industry. If we take the AEC industry, you hear a lot about BIM.
Is BIM the Holy Grail? Is BIM going to solve everything? No, BIM is just bringing CAD functionality to an industry that didn't even have CAD before that. We can make pretty 3D drawings using BIM, and then we can issue instructions to the work site, but we never hear back from the work site. That's the dilemma with AEC. There is no reporting-back system in the entire industry. People go about doing what they're doing. If you're expected to get a ton of cement today and you've called in 10 workers to work with that cement and it doesn't arrive, you're not going to get a signal that it didn't arrive. You're going to see that in a delay report a couple of weeks later. That's the problem with this industry.
It's an industry built on entrepreneurs that are independent contractors, and they do not adhere to a system. What we're doing with SMART Build and other initiatives is to create an ecosystem for the construction industry where you not only can present a pretty 3D drawing of the building you want to build, you can follow it through the construction process and get reports back in real time. If we move to smart factories, finally, we discussed this. As I said in my keynote and as Norbert showed in his keynote, this is a model of an installation we've already done and that we're running in China. This is actually a live autonomous connected ecosystem. You've got a coordinate measurement machine to the right, you've got a CNC machine to the left.
The coordinate measurement machine is driving the CNC machine and tells it when you're not producing quality components, you need to recalibrate, you need to change the cutting die or whatever it is. This is autonomous because there is no human interaction, and it's obviously connected. If you roll it out, it would be a smart, connected factory. Finally, the last target area for our Smart X solutions is smart cities. Smart cities could be anything. Everyone is doing smart cities. You can see tons of PowerPoints on smart cities. I think we said it very well in SI's keynote that what does it matter if you have a smart city if it isn't safe? We need to start building on our public safety systems and our infrastructural capabilities to document the city in three dimensions and then overlay with activity.
We're going to work with public safety organizations to create safe cities. We're going to work with utility companies to create smart utilities with as little outages as possible. Finally, we have traffic. That's how we describe a smart city from Hexagon. With that, my introduction is done, and I would like to welcome Claudio Simão, our CTO.
Hi, everyone. Thank you, Ola. In the context of being a leader in autonomous connected systems for the industries we serve, I would like to focus this presentation not in all the spectrum, all the comprehensive portfolio of technology that Hexagon has, but in the underpinning elements that are to support the concept of ACE. This is the idea. When we have the break or when we have the networking, if you have specific questions on other technology, we can then address that. Basically, I would like to show how transformational, and could be even disruptive, can be the concept of ACE. Which are the influence, or which are the power of the capabilities that we have in Xalt. In this framework, is a kind of package of a foundational of key digital technologies that we interlink together in an interoperable way to support ACE.
Just to contextualize, I don't want to be theoretical here, but because this is a fact. We can see clearly this happening technology. We know the Amara's Law, I guess Mattias mentioned this today, is about human expectation of predicting technology impacts. This implies basically that we tend to overestimate the technology in the short time, and we underestimate in the long term. We see this clearly with IoT, and also we can see this, for example, with augmented reality. We don't see monetization. IoT today, what I see, I will not refer to specific players, but what I see is automation. Industry 4.5 is industrial automation. It's dynamic dashboard. Connecting A to B is considered IoT. Too much expectation, very low monetization, basically. This is the point that I would like to bring.
Hexagon is focused not in the hypes, but in creating the underlying technologies to prepare ourselves to these new trends and, let's say, benefiting of our domain expertise. As Ola mentioned, our richness of our domain expertise will really create many. We have already identified so many use cases, and which use case we can, let's say, pay more attention because it's creating more value or is more monetizable. What many people are missing? It's all about the explosive growth of data and technology, as Ola mentioned, and this is overwhelming companies. Everybody is trying to find where I am in this game of IoT, artificial intelligence, and the other digital technologies. The gap between the technology, what technology can do and what technology is being done, is ironically growing faster and faster. This is the slide also from Ola.
My responsibility in the innovation hub, in synchrony, in sync with the divisions' R&D teams. Supported by the applied technology, core technology from Hexagon, let's call applied technologies, in the direction of ACE. This slide is just a reference that I use many times because this is clearly happening and also in sync what Ola mentioned. The world is getting more and more complex, right? For smart cities, how can be that dispatching of ambulance, police, and fire is not synchronized, is not interlinked with garbage collection or some event that happens in the city? London is not interconnected. One thing affect directly the other. These elementary things are not interoperating together. The world is getting more complex because now we can see it. We can link A to B in the concept of IoT.
The other element, the technologies are getting more and more complex, this is not so obvious, but there are so many new technologies, standalone, very efficient standalone, but fragmented. The overwhelming effect on the companies is trying to use isolated. Well, basically what we are doing with Xalt in direction of ACE, supporting, underpinning ACE, is a framework of capabilities that are interlinked, and more than interlinked, interoperable. That means if you change here, the other one will realize it and will adapt real-time. This is the concept of interoperable. Xalt then is a framework of digital capabilities, as Ola mentioned, that ties and orchestrates all these elements of all this technology and these elements. This permit us, enable us to close the gap between what technology can do and what we are doing with technology. This is a big advantage that we have.
This will enable autonomous connected ecosystems and will address problems that we confront today, but mainly, I guess, as important as we prepare Hexagon for the next step. Because the technology is evolving so fast that if you're not preparing yourself, you will be disrupted. A garage company will disrupt you, and then you can buy the garage company with a multiple, a big multiple, for example. This happens nowadays, so we have to prepare ourselves for this, and that is one of the concepts. Xalt encompasses capabilities like cloud, mobility, edge connectivity, edge computing, edge analytics, data compression, data preparation, vectorization, visualization, AI modules. As Ola mentioned, too, AI is not magic. You need a fitness function to tell the direction.
When the system know the direction, you can create many layers of repopulation of the algorithm and the computer, what they are good is doing calculations. They do this very fast, and they go to the direction of what we want to do. AI can be very powerful if you put this on top of domain knowledge. This will then allow us to collect and leverage data from different types, from different morphologies in enormous volumes. Because with edge computing, you can compute at the edge and just extract what is important. You don't need to push all the information to the cloud. Again, these are the main capabilities, not limited to this. For example, we are investing now in data composition.
It's the next front end, it's the next wave of technology when you can compose data of different morphologies, as I said, to make sense of it. This is the next frontier of technology, but these are the main that we are using. I would like to refer to some key aspects of transformation when we talk about technology, because then this is not extensive. I'm not telling everything, but I guess just to put again in context. One of the very important aspects that we differentiate, if you ask me what is different from what Hexagon is doing and the others? These are some of them I can tell. Autonomy at the device level, at the end point. This is basically referred to edge functionalities, not only connectivity but computing and analytics at the edge. This permits that you do, at the edge, data mining.
You can extract, as I just mentioned, you can extract what is relevant, and you can more than this interoperate. That means I don't need to send even what is relevant if this is not timely or if it's not instrumental for the whole architecture of the system. Basically today, the modern solutions should consider architecture from the edge to the premises. Edge network cloud premises to simplify. That means I have 1,000 of cameras in Waterloo, in London, for example, and I don't want to stream all the information to a central AI video analytics to identify somebody's laying on the floor or somebody left a bag in a corner in Waterloo, or if somebody's walking during two hours in the same place or some anomaly, let's say. I don't need to stream everything. I can process at the edge and send what's relevant.
There are thousands of examples like this, but only processing the edge or at the edge is not enough. You have to understand in the whole entire architecture, what is the meaning of that specific edge? This is the concept. Edge intelligence and edge processing opens a lot of new opportunities for real-time. Another element is speed. Edge processing also influence the speed, right? When coming to the example that Ola mentioned yesterday about Netflix, when Netflix start, everybody said it's impossible to push movies, HD movies, high density movies, to your house. Right? How you do this, right? They start compressing. In the beginning, they're compressing the quality, you deteriorate very fast the quality. Then they start sending only what changed from one frame to the other. You don't send the complete frame, you send just what changes, and you compose this at the other edge.
Finally, they develop the vectorization technology. Basically, if you can process, if you have edge functionalities, you can optimize what you push through. If you have predictive algorithms at the edge, you can be faster than real-time. That is new. That's a very sexy word now that everybody's saying. Faster than real-time, that means you know what will happen, and then you act before really happening. We have examples already in our domain for that. The second one is speed, right? Speed is influenced by edge functionality and also artificial intelligence. More and more, everything that we do is getting real-time. It's very important for us that we can really push information, what is relevant back and forth. 5G is coming. 5G will be very important for many of our operations.
Again, a correct architecture of the solution is more important than just push information back and forth. The other element is future-proofed. This is basically what I was telling you. If we expect the use case to pop up to start developing capabilities that will address that use case, you lose the game already. We have to have a framework that is Xalt in our case, that prepare ourselves for the next step. If we embed Xalt in our solutions, we will be prepared for the use case that we don't know yet. We don't know how much we don't know in terms of the so fast and crazy evolution of technology and use case. This is the concept. I'm compacting a little bit because I'm burning my time here. Simplicity. Sorry. Simplicity on top of everything, the customer expect us to be user-friendly.
We have to have easy to use, and we have to abstract complexity. Abstracting complexity is not only the front end, but also maintenance, life cycle. That means our systems, we cannot have patch working. We cannot sell a solution where that if when you have to do something, you have to go there and manually adjust connections. Our system has to be self-healing. That go again in the direction what Ola mentioned of autonomous connected systems. Frankly speaking, autonomous connected system is supposed to do exactly this, right? To add simplicity, to abstract things that the system should do. These are four flavors of the transformational technology that I try to correlate a little bit with what we are doing with this key foundational digital capabilities.
That means if we underpin, if we embed our Xalt digital capabilities, our frameworks to our solutions, we will empower tremendously our technologies in many areas. We are discussing with the divisions many use case. We have many use case already ongoing. There are products that we launch with the Xalt capabilities, for example. I'm not sure if Norbert mentioned PC-DMIS Go today, but this is one case, and there are visualization, PP&M visualization, all these how you treat data to stream, how we present data in multi-platform. There are many use cases that we are already doing this, but this will gain momentum because we are getting more and more mature with the capabilities. Let me give you three examples on how Xalt is progressively enabling Smart X and ACE. This example is on Geosystems. We have some reality capture, let's call reality capture sensors.
Could be a scanner, could be a stakeout device or whatever. We have the piece of softwares that are the platforms for managing these workflows of this device. We are embedding edge clients, edge kernel, a subkernel of the logics of the framework. We are embedding in these pieces of software or pieces of hardware. With this, we can push back and forth to the cloud, and then we can feed back. We can use this kind of architecture or, for example, use cases, customer intelligence. We can capture usage, how customer is using it, and we can understand where we can change, what we can do better, which functionalities we should do different, which steps we could automate. That means this is much more customer intelligence. This is the famous virtuous cycles of the CTO of Baidu.
You sell more, you have more customer, you have more data, you use machine learning, you do better products, you sell more, and so on and so forth. It's a virtual cycle that would bring monopoly to the complex. Really the Holy Grail is not that. Where we want to go is not there. Is that you understand how the equipment works. You have a machine learning equipment with domain knowledge. That means we have the fitness function that show us which direction we should go. It's called in data sciences, fitness function. Then we can change the behavior of the equipment on the fly. That means if I'm measuring something and I need more accuracy there, I can reduce the speed of measurement to get more accuracy without the user knowing this. I am making the equipment more accurate on the fly with intelligence, for example.
Another example, this is another product very close to market, is machine health for mining. We have the board computers. We have fleet management. We are one of the leaders in dynamic fleet management. That means the system self-adapts as something wrong happens. We are embedding in the board computer edge clients. We aggregate information at the edge, one of the assets, because this is a network, we push everything to the cloud. We have the predictive maintenance algorithms in the cloud just using what is important, the anomalies. We detect the anomalies at the edge, we live stream into a command center and feed back to the machine. Basically the same thing. We have some studies that shows that we can reduce till 10% the cost of maintenance of these big machines. This is very relevant too.
We are not unique with predictive maintenance for mining, with our capabilities of processing at the edge and orchestrating everything, edge network, cloud premises, we think we have a differentiator for this product too. Another one that is a little bit related to PC-DMIS, Norbert, what you mentioned today. At the machine level, at the edge level, we have edge computing and our edge clients for connectivity. We push information to Xalt core at the cloud, we use information from CRM because this is the fitness function, domain knowledge that we have from the CRM. With these systems today, we can push notifications to iPhone or any mobile device to a manager or supervisor to help him to take correct decisions. This is basic decisions support. Or we can have an algorithm at the cloud level, to change behavior of the machines.
Examples of these are many. One of them, you are measuring a body of a car you detect the right side, the dimension of de merit is trying to deviate, is not so good. You don't want to increase the cycle time of the measurement because it's a production line. You start measuring less in the left side and more in the side that the dimension of de merit is growing. You can push information to the press shop to verify to that specific part why the die is pressing, is stamping this part to cause the dimension of de merit. Could be also in the welding. All this logics is the domain knowledge. Conclusion, wrapping up what I said. Hexagon core capabilities and domain knowledge empowered by Xalt will allow what we call Smart Digital Reality.
That means the convergence of digital and physical. Bottom line, as Ola mentioned, is more productivity, more quality, more security, depends the application that you have. If you want to summarize Hexagon at a very high level. We can position anything anywhere. We can capture reality in real time. We can provide intelligence and domain expertise in the context. That would be situational intelligence. We can design. We have a lot of domain knowledge in vertical applications where we are number 1 and number 2. With our framework of Xalt, we have a major opportunity to capture more value in many areas. That's my presentation. We make a break now, right? 15 minutes? 15 minutes, and then we come back to Ben. Okay. Thank you.
Welcome back. Good afternoon. My name is Ben Maslen. I've been Chief Strategy Officer of Hexagon for just under a year. One of the key components of my role is that I oversee the M&A process. We're going to spend about 15 minutes going through that process, talking a little bit about the kind of acquisitions we're looking for, and then how they tie in with the overall strategy that Ola talked about before the break. M&A, as you know, is a key component of the 2021 financial targets. We aim to add 3%-5% growth per annum from M&A. That is an average. Some years be better, some years be worse, that will facilitate us hitting our 2021 targets. This slide tries to summarize the M&A process at Hexagon overall.
Corporate M&A, our responsibility is to screen all the acquisitions that come into the group. We work with the divisions to build a pipeline, then to take the acquisitions we choose to pursue through to completion. On average, over the last few years, we've done between 10-15 transactions per year. Our capacity to do M&A is dictated by an internal target of 2.5 times net debt/EBITDA. That gives us plenty of firepower to do acquisitions. At the end of last year, we closed 2017 at 1.8 times, so in a very good position coming into 2018. Where do our ideas come from? We get a lot of teasers, external processes coming in from bankers, our role is to filter through those and work out which ones are interesting and to feed back down to the businesses.
Most of the acquisitions Hexagon does are originated in the divisions. They will build long-term relationships with potential targets, incubate those ideas, when the time is right, they will send them back up to us for review. That review process focuses on two things. Firstly, does it fit with the Hexagon strategy? Secondly, does it meet the financial criteria that we use to work out whether these are good deals to do? In terms of the strategy, we're not just trying to fill a revenue quota with our 3%-5%. We are trying to buy businesses that will accelerate the strategy that each division has. We spend two weeks every six months, reviewing the strategy of the businesses, go into it in great detail. We work out what technology they need to deliver that strategy.
Should they develop it, invest more in R&D, or can we buy something to accelerate that process? We focus very hard on the synergies that an acquisition will bring, and we give greater attention to those acquisitions that support the Smart X strategy that Ola talked about before the break. In terms of the financial criteria, we're looking for companies with a strong market position. We want profitable businesses, so we want the margin to contribute to the upward trend in margins that we've seen historically and we expect to continue going forward. We're looking for companies with a significant software element, high levels of recurring revenue. It's important for us to look for targets that allow us to maintain our valuation discipline and create value for shareholders. I want to spend a second just talking about the ongoing shift to software.
If you add up all the acquisitions we've done since 2012, 65% of the revenues we've acquired have been pure software companies. You can see some of those in the box at the bottom, MSC, Vero, Luciad, and AgTech that we closed earlier in the year. The 65% excludes software that we might get bundled with a hardware company or sold together. An example could be NEXTSENSE that Norbert talked about in his presentation earlier. Handheld laser scanners, 30%-35% of the revenues of that company are software. That's not in the 65%. This shift is having a big impact on the financial metrics of Hexagon as a whole. Software and services now represents around 55% of group revenues, and that is contributing to rising profitability. It's reducing the inherent cyclicality of the group because these business models are more stable.
Since the economic downturn in 2009, we've obviously added Intergraph, MSC, Vero, and so forth. They set the company up very well as and when we see the next slowdown. Finally, the shift towards software is accelerating the reduction in working capital that we see, and the improved cash flow that we've seen over the last couple of years. A very positive overall impact on the group. Maintaining financial discipline in current M&A markets has become more difficult. As many of you will know, we have a very hot M&A backdrop at the moment, fueled by low interest rates A volatile growth environment that is forcing companies to acquire to keep their top lines growing. Private equity funds that have a lot of resources at their disposal, and obviously financial markets themselves, spot multiples are very high. That just raises overall expectations.
How has Hexagon acted in this market? We've kept disciplined. We've walked away from acquisitions or lost deals where we don't think the financial metrics make sense. We've still managed to close 10-15 acquisitions per year at what I think are very attractive valuations. How do we do that? Firstly, we screen a large number of targets. For the 10-15 acquisitions that we close every year, we'll probably screen a couple of hundred of opportunities. At any one time, we have a pipeline of about 100 projects at various stages for us to draw from. Clearly, by looking at more opportunities, you can be more selective and make sure that the valuations work for you. Secondly, we focus on the highly synergistic projects.
By identifying synergies, we can make sure that we can accelerate the growth and margin progression of companies that come into the group, that helps bring the paid multiples down. Thirdly, where we can, we try and structure acquisitions with earn-outs. We think this makes very good economic sense. It allows us to pay less upfront, align the incentives of ourselves and a company coming into the group to make sure it performs very strongly. That all together manages to bring the multiples down. We have a couple of inherent advantages. As I said, the businesses will incubate these opportunities, build long-term relationships. We are often the preferred acquirer for some of these smaller companies that want to be taken over. We have a fairly light touch in terms of how we integrate companies.
We like them to flourish, support entrepreneurialism, which means that, again, for many companies, how they want to come into the group. We look at a lot of different areas, niche technologies, where some of the bigger players they don't play. This is tying back to Ola's strategy, enabling the Smart X, and autonomous connected ecosystems that are pulling together the different technologies we have across the group into solutions for smart factories, smart cities, smart autonomous fleets, smart farms and mines, and smarter construction sites and plants. The M&A process feeds into this as well. What this shows you is just some of the acquisitions we've done recently, and how they feed into increasing the kind of technological reach that these solutions that we're working towards have. For smart factories, MSC, NEXTSENSE, Vero.
Smart autonomous fleets, we have Vires, a simulation software company for autonomous vehicles that we bought last year. AutonomouStuff that we signed that transaction and announced it last week. That will play significantly into our longer term view of smart autonomous fleets. EcoSys, our project control software business. AgTech, takeoff software feeds into smart construction and sites and plants. Lots of new technologies coming into the group. I have a few examples of those now, and obviously you can spend more time with them in the zone. The first technology is Luciad which we acquired in the fourth quarter of 2017. This provides a visualization platform to fuse, visualize, and analyze geospatial data in real time. Real time is the key. It's on the fly.
As Jörgen talked about in his presentation, you can pull in 3D maps to cities, you can overlay that with sensor feeds to create what we call a 5D digital reality, which allows you to show moving objects on a map and see how they pan out over time. This obviously has a potentially very big number of applications in our Smart X opportunities. As we can see here in San Francisco, city applications, traffic, people flow, and so forth. Longer term, autonomous fleets. Anything you visualize something moving on a map has use in aviation, mining, and agriculture. MSC, we acquired this just over a year ago. It's a leading supplier of simulation software. Very strong position in automotive and aerospace simulation, as you can see on screen. This has been within the group now for just over a year.
Margins were already very strong when we acquired it, the integration's gone well. Growth is now accelerating nicely. What MSC does, as Ola described earlier and in his keynote last night, is allow us to take the metrology data coming off a production process, which is a source of truth of what you've made, and feed that back into your CAD/CAM software, the manufacturing workflow, and ultimately with MSC back into your design suite so you can use that data in the next iteration of products to make it better and improve quality. Finally, a little video on our AutonomouStuff. Again, we signed this last week. We expect it to close over the next few months. Their core product is providing platform vehicles that integrate hardware, software, all the key controls you need to drive and support autonomous operation.
They sell to OEMs, tier 1s, tech companies, universities, startups, anyone who's hoping to build a product or a technology that goes into the autonomous vehicle project, have their own software for speed, lane control, and so forth, lidar object processing, as you can see somewhere in this clip, and solutions for capturing and processing the vast amounts of data, terabytes daily, that these cars kick off. I've sat in one of these cars. Quite scary. Very exciting. I think it'll take me personally quite a while to actually have an autonomous vehicle. For the group, we think it's a fantastic acquisition. Very clear synergies with technologies that we already have. Positioning Intelligence, the GPS and the correction services that Michael's business has, simulation software in Manufacturing Intelligence, obviously map production and processing, where Hexagon is a market leader.
Longer term, clearly the technologies that are developed to produce autonomous cars, that can be translated into solutions that we will be working on for Smart farms and mines. Just in summary, this is the acquisition history for Hexagon over recent years. As I said, the last couple of years, the run rate has been about 10-15 acquisitions per year. So far in 2018, we've closed four acquisitions and signed AutonomouStuff on top of that. In terms of the historical contribution to group revenue growth, 2011, the 33% is when Hexagon bought Intergraph. There were a couple of years after that whilst that deal was digested and leverage brought down where the contribution was a little bit less. Since then, it's stepped up nicely, and we're running at the 3%-5% run rate that is our commitment to hit the financial targets.
For this year, if you include acquisitions we did in 2017 and the spillover from that, plus what we've already closed, we've got 3.5% revenue growth in the bag for 2018. In conclusion, M&A will continue to be a core component of Hexagon's strategy. We'll focus on acquisitions that have a high software element, high synergies with the rest of the group, and contribute towards the development of our Smart X initiatives. We feel confident that we can, even in fairly elevated M&A markets, keep adding 3.5% revenue growth per year, and with that, increasing shareholder value. With that, I will hand over to Robert.
Thank you, Ben. Good afternoon, everyone. My name is Robert Belkic, and I am the CFO. I am going to take you through some finance slides the next couple of minutes. Starting with the historical overview of Hexagon and the path to progress. Hexagon's path to becoming a leader in digital solutions really started in the year 2000. At that time, we had sales of EUR 500 million and an EBIT margin of 5%. Since our transformational journey began, we have delivered strong sales and margin growth. In 2017, we had sales of EUR 3.5 billion and an EBIT margin of 24%. Taking you back to December 2016, when we launched our financial targets, our new financial plan, our 2021 plan. In 2016, we closed our books on EUR 3.15 billion, and we had an EBIT margin of 23%.
The plan that was launched had the base case scenario or has the base case scenario, EUR 4.6 billion and 27% EBIT margin. We also have an opportunity scenario where we are to reach EUR 5.1 billion and 28% EBIT margin. Once again, looking at 2017, where were we? We were at EUR 3.5 billion with a 24% EBIT margin. Clearly on track toward these targets. Ben zoomed in on this one. This is the basic assumptions of our financial plan and the components of it. Ben focused on an M&A box. I will focus a little bit on the organic growth box. Once again, what we have said, as Ben reiterated or said, was 8%-10% is the total growth per year on average. 5% of that growth should come from organic growth.
It is important to stress this is an average over this period of time. Any given year, we can be above or below the 5%. It is important to stress 5% is an average over this five-year period. 3%-5% will come from M&A, as Ben pointed out. When this plan was put in place, FX was assumed to have a zero impact on us going forward. Looking at 2017, how did we end up? Total growth 10%, organic growth 5%. The contribution coming from acquisitions was 6%, and we had a negative impact from FX of -1. Once again, well in line with our targets. Shifting gears a little bit, I would like to talk about a few key points that explains how our strategy has contributed to strengthening our financial metrics. Starting with the shift in our business model.
As is apparent from this slide, 55% of Hexagon sales today is software and services related, and 40% of the sales is related to recurring revenue. Going back in time, 2010, the recurring revenue part of sales was 20%. Since a couple of years later, we are now at 40%. There are really three things that have been driving this transition. Firstly, it is a shift in strategy from selling products to selling more solution-focused products. Where we bring more value to the customer and getting a response from the customer and them having an ability and willingness to pay. Secondly, our software-centric business has typically outgrown our hardware-centric business lines. Thirdly, as Ben pointed out, the M&A strategy of Hexagon has been acquiring software companies who have a recurring revenue model.
From Ben's slides, once again, 65% of the companies that we have acquired the last couple of years have had these features, software and recurring revenue. Okay. Where will these key metrics go going forward? What kind of targets can we have going forward? Once again, a very hard question to respond on and to give a scientific answer on, because there's many variables at play for us improving these key metrics. It is, once again, based on this development will continue to improve. You will see a continuation in this development based on the three arguments I gave to you in the previous slide. Okay. My next point is then our strong focus on R&D. As you know, that's the core and DNA of Hexagon. On the R&D side, we spend typically 10%-12% of annual sales on R&D.
This is really an essential part of our strategy to be able to continue to deliver profitable growth. From time to time, we also get questions on capitalization. Why do we capitalize? How much do we capitalize? The effect we have on our P&L statement, is it increasing or decreasing going forward? The answer is that we capitalize 50% of R&D spend every year, and we amortize it over a two to six-year period. This has been the case for the last 10 years. A very consistent approach when it comes to capitalization. The reason for why we capitalize is really because it's mandatory according to IFRS. You should capitalize if the development expenses pertain to new products, if the product is expected to generate considerable earnings in the future, and the cost is significant.
Looking at the benefit then from capitalization, as is evident from this chart, the gap between capitalized R&D and the amortized R&D is roughly two percentage points. This gap has shrunk and is continuing to shrink and will also continue to shrink going forward. Final comment on this slide is really, the R&D focus and core of Hexagon, you've seen multiple examples of that throughout the day with all the divisional keynotes, and all the interesting launches embedded in those. You will see furthermore of that portfolio when you do the tour tomorrow in the zone. Okay. Another slide and another strong development. This is our cash flow, starting in 2011 until 2017. Cash flow from operations has increased with 112% over this period of time.
Looking at the cash conversion, our average cash conversion during this period of time is 85%. That should then be compared to our target of 80%-90%. The improvement is very much driven by the mix shift I mentioned on previous slides, this also the recurring revenue-driven model that we work with. Final remark or comment on cash flow when it comes to seasonality. Typically, our cash flow is a little bit weaker in the first six months of the year, and typically stronger in the latter six months of the year. Next slide, looking at the working capital to sales. Once again, a very positive development for us throughout the years. Back in 2010, prior to the acquisition of Intergraph, we were at 30-plus, and now in Q1 2018, we are at 13%. A very positive development for us.
Actually internally, we have had an internal target of 15%, and that has now been overachieved. It is a little bit of a soft target. We will never come down to zero, as we are not a 100% software company with that kind of model. Us being on these levels is very impressive. We have bits and pieces of the Hexagon business and the Hexagon divisions that will always tie up working capital as we grow. Still, 13% is an impressive number. Final comment and our final slide, which is related to cash flow. Taking you back then to Q1 2018 and what we presented there. We will have two larger one-offs in the cash flow statement during 2018, where we are investing in two facilities that will impact the cash flow negatively this year.
It is Calgary, which is being inaugurated in July 2018, where we are moving into an R&D facility, and that facility will house employees from Positioning Intelligence, PP&M, SI, Geospatial. The second larger investment we do here is in Huangdao in China, where we are investing in a state-of-the-art technology park in Huangdao, where we will showcase all different technologies and all different divisions within Hexagon. The impact that these two investments will have on the cash flow statement is CapEx of EUR 90 million-EUR 110 million for 2018, which means that the amount in investment in tangible assets will amount to approximately EUR 145 million-EUR 165 million for 2018. Looking at our return on capital employed. Here we have an internal target of 15%. We are currently at 12.6.
Considering the high profitability of Hexagon, you might think that the 15% target is a little bit on the low side, but you need to consider and you need to be aware of that with the very ambitious M&A strategy that Hexagon has had historically and is going to have also going forward, it means that we have done several larger M&A deals, which means that the goodwill amount in our balance sheet is large. Also, we are a little bit of a young company, I would say. It takes time for a company to grow into its balance sheet, and with 18 years of track record, it will take time until we grow into our balance sheet. Eventually, we will grow into it, and this ratio will come up.
If you, from a theoretical perspective, not saying that we should, but if you still do the alternative calculation, if we would take away the goodwill from our balance sheet, the return on capital employed would actually be 37%. For us and for me, it is important that we are moving in the direction of reaching the 15% target, and that in itself is a testament of improving margins within Hexagon, a very disciplined approach when it comes to M&A valuations, and also, once again, the underlying shift that we have in the business model. Talking about margins, starting with gross margin. 61% is our gross margin on a rolling 12-month basis. Once again, what is the underlying effect or the reasons for this one improving? Well, we have a lot of small incremental steps in the right direction within the Hexagon divisions.
The software and services part increases, the recurring revenue part increases, the solution-oriented sales increases. All that is enabling and fueling this development in the gross margin. Us improving the gross margin is clearly an indicator that our new products that we constantly launch are able to command a higher price than previous generations. Looking then at the EBIT margin, once again, a very positive development for us. Rolling 12 months, 24%, which should then be compared to the 27% and the 28% target that we have. Looking at our debt level then and our deleveraging capacity. What we have in our financial bank documentation today is one financial covenant, net debt/EBITDA of 3.5 is the max. We were hoovering around that level back in 2011 following the Intergraph acquisition. We then took a deliberate decision to deleverage our balance sheet.
We set up an internal target to reach 2.5, and as is evidenced on this slide, we reached that target within two years. Since then, this ratio has continued to come down, with the exception of two spikes. The first spike was in the fall of 2014, when we, in the same quarter, acquired two software companies, Mintec and Vero. The other spike is related to the acquisition of MSC, Hexagon's third largest acquisition ever. That created a spike in this curve as well. What I think is a key takeaway from this slide is really we will have temporary spikes in this curve, but with our strong cash flows, we have a tremendous capability of deleveraging very fast. Although the curve is spiking, we're coming down very fast immediately after that.
Once again, if we just do a theoretical mathematical exercise, 31st of March, we were net debt to EBITDA-wise at 1.77. All things equal, that give us a headroom of almost EUR 2 billion when it comes to net debt and EUR 558 million when it comes to the EBITDA. Looking at our financing down here, we also have a very transformational story, I would say. Historically, we were very dependent on the bank loan market, and that picture is very much different today. As is evident from this slide, today, when our functional currency is EUR, we have 98% of our debt being denominated in EUR, 93% of our debt sits within the capital markets, i.e., commercial paper, MTN notes, and bonds, and our average interest rate on our funding today is 1%.
42% of the funding of our debt is long-term, 49% of the interest rates that we pay are fixed, and the average duration of our debt is 17 months. Also, a final remark on this slide, all the short-term debt of Hexagon is at all times fully backed up with the revolving credit facility we have in place. A EUR 2 billion facility sitting with 14 banks, which has a maturity in September 2021. To conclude then, and to summarize what I've been talking about today, well, I've been looking at the shift in the business model towards more software and solutions, which will further strengthening our recurring revenue and margins. I talked about the working capital, the continuous improvement of working capital, and the very strong cash flow of Hexagon.
On the return on capital employed and the target we have there, it's an improvement, a continuous improvement, despite the very ambitious M&A strategy that we have within Hexagon. A final remark that we are clearly on track towards the 2021 targets. With that, I hand over to you, Ola, and welcome you on stage.
Right. It's time to wrap up, and a few summary slides. Hexagon's expansion is going to happen in four areas, expanding the core. I think you've seen today, especially with Jörgen's keynote, that we're modernizing and we're simplifying our products, and it's much easier to use the new products than the old ones, which is super important for Hexagon because we can access a much greater market. Capitalize on digital transformation. We've heard that word over and over again, but I think it's not a mantra that you can use generically. You need to know what you want to transform. Smart X is our wraparound for the future applications that we see, and they are in factories, construction sites, autonomous vehicles and fleets, smart cities, and mines, and agriculture.
We have new opportunities, and we talked about autonomous X, where we definitely have a strategy for acquiring that. We have the AEC market, which is the new market for us, especially in terms of software platforms. Content monetization is something that we've discussed, and it's a way of transforming old hardware revenue into recurring service revenue. I think the best example we have of that is Geosystems that used to sell airborne sensors. Now we're selling data to users of data, and it's a recurring revenue business. Synergies and innovation, we're continuing to leverage synergies across the divisions. We have innovative sensing technologies and Xalt as a framework for all the divisions to draw from. It's important to remember that it's not separate businesses. We're orchestrating this, so to say, to reach the Smart X.
To create a smart construction site, we need the base platform from PP&M. Without the acquisition of Intergraph, Hexagon would never have been able to do this. The Geosystems technologies and sensing technologies and knowledge about that market is absolutely crucial. Geospatial provides 5D capability, and Manufacturing Intelligence is helping with accuracy as well. You can take each one of these, and we can tell you a story how we draw resources from all areas of Hexagon to make this happen. I think when you look at other companies and you compare our companies to those companies, you will never find this depth and breadth of applications and technologies, and this is what's needed to deliver the future. How do we set about reaching our financial targets? Well, it's a combination, as we've heard today, from divisional development, where we have divisional five-year plans.
On top of that, we have group projects where we pull resources from the divisions and from the hub to enable the Smart X. On top of that, we have M&A, which for us is very much a make or buy decision. We have a quote here from a famous professor, and that is, "Autonomy can only be achieved by converging the digital world with the real world," and that is absolutely true. With that, we're going to end the presentation today, and I think you're going to see more of this when you go to our technology center. Is that tomorrow? With that, we open up for Q&A if there are any questions in the auditorium on what you've seen or anything you wonder about. Yeah. Yeah. I want to start this. There's two parts to the