Hubert van Doorne, Nexmosphere

The 16:9 PODCAST IS SPONSORED BY SCREENFEED – DIGITAL SIGNAGE CONTENT

Sensors and triggered content have been part of digital signage for probably 20 years, but they weren't widely used for a lot of that time because putting a solution together involved a lot of planning and custom electronics.

A Dutch company called Nexmosphere has changed all that, offering a wide range of different sensors that trigger content to digital signage screens by sensing the presence of people or reacting to an action, like someone lifting a product up from shelf to get a better look.

Nexmosphere has developed its own set of low-cost custom sensors and controllers that make it fast and easy for digital signage solutions companies, including pure-play CMS software shops, to add triggered content capabilities. Nexmosphere focuses on the hardware and makes an API available to partners.

I chatted with CEO Hubert van Doorne about the company's roots and how his customers are now using sensors to drive engagement in retail.

Subscribe from wherever you pick up new podcasts.

TRANSCRIPT

Thank you for joining me. Can you give listeners a rundown on what Nexmosphere is all about? 

Hubert van Doorne: Oh, yes. I’d love to do that. Nexomsphere is all about sensors for digital signage and to make any digital signage system interactive and by interactive, we do not mean only by pressing a button or touching something on a screen, but it's much more of letting the digital signage installation respond to actually the person in front of it. So it can be a presence sensor. It can be something that you touch or lift the product and that way make an interactive system at Nexmosphere, build a platform of a lot of different sensors that can be hooked up easily to the system and in that way for the system integrator and really help in building this sort of system.

It's something that's been done for decades by a lot of companies who made their own sensors so in that case, it’s not real rocket science, but the beauty of the product is that you can take any of these sensors, bring them together and you don't need any development time to have a proof of concept running and I think that's the real strong point of what we offer. 

Yeah, I've been in this sector for a very long time, and I can remember, when I was working with one company, we had a guy who I think worked out of a motor home somewhere in Arizona or California or something like that, and he would design boards that could be used for very early iterations of sensor-based detection, presence sensor, that sort of thing, but it was a lot of work and there was nothing off the shelf that you could buy to do all that.

I suspect that there's a whole bunch of companies that said, “Oh, thank God, I can just order this and it's already sorted out.” 

Hubert van Doorne: Exactly. It's actually also how the company started in 2015. We were a part of a display company making POS materials for in-store and we developed our own sensor set for Sonos and for Philips to go into the European stores, but what we saw is that the number of stores was heavily declining and the complexity of every system was really growing. So we needed something modular that was not just for one group of displays. We needed something for five thousand, you could make the same part of electronics, but now we need it for 200 stores, something still affordable to build, and that's how the idea started off of making a more standard solution that you can plug and play.

And, first, when we came to the market, everybody was like, “No, we will build something ourselves, that's much easier.” But then, over time you saw these numbers of installations decline. concept stores would only be maybe five stores or ten stores and then slowly evolve over time and that was getting difficult to make something really solid, and robust in small numbers. So that's where we really saw that the companies coming to us. Like indeed, what you say, thank you, you're developing this because it's only a very small part of the total installation cost and compared to a screen or a media player, we are only a fraction of the cost, but it's adding a lot of functionality. So in the end, if it's not working, it's a problem. but it's also not something where you can really save a lot of money because it's only a small part of the budget. 

When you say it's a small part of the budget, are we talking tens of dollars or tens of euros or hundreds of euros on a typical installation? 

Hubert van Doorne: It depends, but if you go for a presence sensor by screen, you have it over, $40-$50 a sensor. So that's only a few percent of the total installation, and of course, we go also to museum experiences where we have 20-40 sensors in one set, and then maybe you go to a few thousand euros, but in the end, it's never a very expensive set of gear, because, in the end, it's just small electronic bits that you can very well pick the right size that you need for your project. So there's very little that you do not use in the end. 

What's the range of types of sensors that you do? 

Hubert van Doorne: In terms of applications, we have a couple of real main areas and the first one is absolutely by far present sensing and its presence and motion. So if someone is in front of a display, if he's approaching or moving across or where he is exactly standing that sort of application, but it's very much based on seeing where a person is and depending on where he or she stands, trigger the content or just for statistics, counting how many people are there.

And, the second, large area is what we call lift and learn. So if you pick up a product you can actually display information about this product over video, because we know which product’s being lifted. Typically we have about seven different technologies that can lift and learn. So we have a lot of different sensors for different products, but in the end, it is all for the same application of building a lift and learning.

Another bigger area is LED lighting which you use as guidance. So a third area that we do is, guidance where you really can guide people, for example, in a museum, where to look if you light up an exhibit, but it can also be in a cosmetic store. If you have different products and you make your selection on, for example, a tablet light up the products that match this selection. So LED lighting, it's not so much a sensor, it's more an actuator, but that's what we do a lot as well, and the fourth, biggest pillar is what we call UI elements and that's anything to do with, for example, normal push buttons, but it can also be air buttons so that you just hold your hand somewhere on a position very close to the screen and it already triggers so you don't have to really touch, air gesture that you can do with your hand in the air. That's all elements where you really want the user to do an interaction with the screen that is not on a touch screen and those elements, that's another important group of sensors. And then we have a lot of small, little special sensors. They're fun, but it's not the big chunk.

Did you find as you were growing the company that you started adding these things because people were asking about it or you just saw that people are struggling with understanding where to look next and so on… so let's use LEDs to light guide them? 

Hubert van Doorne: Yeah, I think it’s a bit of both. Of course, we had already very early clients who helped us also. By asking the right things like, “Hey, this is something we could really use. So can you build something for us?” But also looking at ourselves, we have a very strong team, with a lot of ideas where they really look at what is needed now in the market and how we can find a technology that can really be well suited to do this and I think that's also a little bit where we differentiate ourselves from others. 

You do have great industry sensors. They are very expensive. They are very good but, for larger rollouts, and larger deployments, they are far too expensive to make a big rollout or too complex. Yeah, you have the cheap stuff, but in the end, it's also not something that can really scale because you cannot really adjust them exactly how you have to use them in mainly a retail environment, but also we see them now in more and other environments where if you use too many standard sensors then you cannot customize to how it should behave in that area and you can also get a lot of misreads and I think that's the strong point of our sensors that we really build something that you just connect and it works in that retail environment. 

A good example I always give is a presence sensor. We had a competition using exactly the same chipset from ST Microelectronics, but the fact that somebody walked in front of it and there was a minimum timing of a few milliseconds. A person should be there for half a second because otherwise a person cannot appear in half of a fraction of a second and be gone again but if you have a very fast sensor and you think, oh yeah, that was a person you also have to look at the application. It's not in a shopping center that somebody comes in and then a half second is gone again. So that must be a misread somewhere from the sensor, and that's how you very cleverly can filter how sensors should work, and I think that that's why we really make them for the, we call retail environment. That's our main application area, and if you want to use it in other areas there, it will work as well, but retail is just a very harsh environment where it should always work.

Are these devices that you're somewhat buying off the shelf, so to speak, or are you pretty much going down to the wood and designing these from scratch and sourcing the different components and maybe designing your own PCB and so on? 

Hubert van Doorne: Yeah, it's a good question. All PCB design is completely done by ourselves, but the actual sensor is often a technology that we use. It's a standard chipset available from one of the big manufacturers like Sharp or anyone who makes real sensors for the industry where we take just the sensor chip part, put it on our own PCB, and really make it a solution that fits the digital signage market.  

So do you have contract manufacturing that makes this stuff over in Shenzhen or somewhere or are you doing that in the Netherlands? 

Hubert van Doorne: Some of our PCB manufacturing or most of our PCB manufacturing itself is done in China but we also have some clients who prefer the European stamp. So for some of the products, the PCBs are also made in Europe, and then end assembly and end-of-line testing are all done in the Netherlands. So that is where we really make the product and put all the software on and package it as a final product. 

If you look at the qualification, most of the products are really made in the Netherlands. It's not that we have a contract manufacturer somewhere in China who builds our complete box. We want to have that control at the end still in the Netherlands. 

Yeah. So you're definitely not just rebadging something that's made for the China market and client your own? 

Hubert van Doorne: No, absolutely not. For this, it would also be a little bit too complex. I have to say, of course, some accessories like cables or that sort of thing is of course, the standard stuff we sell, we do not make anything nonspecial ourselves, but all the sensor part controllers, everything that's, that's true, including if you look to the products, all the casings, everything is our own design.

You mentioned presence detection. Can you work with computer vision platforms and drive analytics, or is it very kind of basics, sensing the saying, “In an hour, these many people appear to be in front of the display?” 

Hubert van Doorne: Yeah, we do a little bit different than computer vision. I think honestly the computer vision guys are making great products. The only problem often is that you cannot put a camera or it's something that needs to be calibrated to the space. Typically our sensors are something that you can just drop in the space and it will start working and also on a different price point, but if you look, what we can provide is that we can see if there's someone in front of the sensor and we can measure the distance on a, like a centimeter-accurate. So we know it's a half inch where someone exactly stands and if she, for example, now moving towards display or moving away we get all this trigger data delivered to the digital signage player.

And with third parties, partners of ours, for example, you can have beautiful dashboards that can give you insights like, “This was the opportunity to see” and “So many people engaged with the product.” There you can see a complete dashboard exactly at that time of the day. It was busier because, in this area, more people were standing at that sort of statistics we could easily provide them. 

Yeah, for many years I have wondered about computer vision, the applicability for a lot of digital signage applications, this idea that you're going to serve content based on the demographic profile of who's in front of the screen.

It just didn't hit me as something that would ever be done very much, and I'm curious if your customers and partners generally find that, you know what, just having a sense of how many people the opportunity to see and so on is more than enough. We don't need to get that granular. We don't need to micro-target individual viewers.

Hubert van Doorne: I would say if you talk to the advertising companies, it's sort of the holy grail and they're all looking for the solution, and it's not that it's not coming, but it's also the question of how would you target your audience and I think that's something that’s being asked a lot, but if you see what is the technology needed for it, and if I see how far digital signage is now off still often, not even, if you see just in the store, like looping content, yeah, you could already be much more responsive than just looping content.

And that's also not being done. So before we leap into light years ahead of what we could do, and yes, technically we can do, but if, for example, already the advertisers have problems delivering the right content at the right time, this becomes something very, time-sensitive to really have it in an auction platform and very quickly delivered. So for sure, it's something where, slowly the whole advertising market will go, but I personally do not have the feeling that it will very quickly come into large deployments. 

Yeah. People are still freaked out about the idea of a camera being on them, even though there are cameras all over a store watching for security reasons.

Hubert van Doorne: Yeah. But it's also about how would you like to use it. I think showing content at the relevant position. There was once a client and this was many years ago, but it was about in-store materials and he was a manufacturer of cat food. So he says, “If somebody is in front of the shelf with all the cat food, that's the perfect moment to advertise because I know my customer is a cat owner, or at least someone who wants to feed the cat, and if he's male, female, and their age” and I think that's a very relevant thing. 

If you look into a store layout or for all the media advertising in the store if you know where the shoppers are already in the store, you can also tell a lot about what would be relevant content to show and that it's not only based on age demographics or demographics like male, female. I think that's something typically the advertisers do to slice the market and spice up the market. But I think there are different ways to do that in a very elegant way as well. 

Are you competing with touch or is it just another approach?

Hubert van Doorne: I think we work well together with touch even. I think we are hyper-focused on the sensor market, and that's the same, as what I say about computer vision. I think they make lovely products and also people who do mobile phone tracking, and it's a huge market of all kinds of inputs and sensors you can do and a lot of manufacturers have great touch screens, and I think what we do is just add another layer of possibilities of sensing something that can work well together. 

There's a bit of an orchestration to some of these experiences, right? if you have presence sensing, you can be 10 feet away and you might get one message if you're closer. If you move closer to a screen, it can trigger a different message and maybe guide them using the LED light bars and things like that. It seems to me I've seen demos of this when I've walked through your stand at shows like ISE. 

Hubert van Doorne: Yeah, that's definitely the presence sensor can just be to make something different than a looping video. I always say for the presence sensor, it’s the same as reading a book. I would like to start on page one, and if you want to tell a story to your customer, why are you starting somewhere in the middle? Because it's a looping video. You will never be synchronized. If somebody approaches the screen, it'll always be in the wrong spot.

And I think with a presence sensor, you can have a nice triggering content that really attracts the people to come close to the screen, and when they’re close and when they are in range, I would say you can start firing your message and have your story being told. A nice thing about the sensor is that you can actually time how long is the person standing in front. and is he watching the whole advert or is he halfway moving away? And that comes to another thing that you can very nicely do is, of course, the A/B testing. If you have different content, you can say, hey, with this content, just where the end is a little bit longer and a little bit more flashy, people stay six seconds longer, so this is better content to show.

So in that way, you can really optimize your content and your message by measuring how long people are looking and how can we improve. 

Now, if I'm a CMS software company, or a retail solutions company of some description, if I want to do that, am I having to write the software to create that experience, or is that somewhat templated or available within your end of the solution? 

Hubert van Doorne: That's a very good question. Technically we are a hardware provider. So the only thing we do is deliver the hardware that's capable of delivering triggers to the media player, and that can be over USB or over network or RS22 and we deliver that trigger message, and then you need a CMS that can actually really sense what to do with the trigger and there's absolutely no software from our side. It's just then the software of the CMS where you can, really in that UI setup like, for example, in a BrightSign, you have Bright Author, and then you can select from a list like, If sensor three is triggered, then I want to start video number five and when this is done, I want to have that started. So that's something that can be done via the CMS. 

I noticed on your website that you have three kinds of distinctions or tiers or whatever or partners. You've got full partners. You've got fully compatible and then partially compatible. What's the distinction between them? 

Hubert van Doorne: It's a very good one. We are still missing the official explanation on the website. So we'll add that over time, but no, maybe good to start with our partners that are really strong partners in the industry where we often exhibit together, we do trade shows together and we do projects together where we really team up and make something work for the integrator as a complete set. So that's our real industry partners. 

There are also a lot of CMSs that deliver a nice software solution, but in the end, we are not very closely working with them together, but the application works well and that's where we say it's a compatible CMS. So that means that all functionality will work, but we don't know that many details about the CMS. So if you want to know how our solution will work with that CMS, best to contact the CMS and they can tell you more about how the integration works because we cannot keep track of 50 different CMSs, how they all work, and how they all work in detail.

Lately, we added the partially compatible and that's the reason that we have a few that do a lovely integration on two or three of our sensors, and it's a very nice UI where you can set everything, but it's not working with every sensor. It's just, for example, our few popular sensors where it does work with, and they say yeah, but we can easily make a widget for this client and then it can work with any sensor, yeah, but it's not something that you can just connect and it just will work. So we say, with these companies, it does work, but for some specific centers. So please watch out and contact the company if you want to use it for this specific specific sensor, and ask them if it's working. So that's why we made these three different tiers.

These sensors are generally very small, and in a lot of cases, you would never even know they're there, right? 

Hubert van Doorne: Exactly. 

Is that by design, a necessity even? Or is it just that's the state of miniaturization where you can do that? 

Hubert van Doorne: Yeah, I would say the fact that you can make it very small is because it's retail, nobody wants to see tech. So the more you can hide it, the better it is, and yeah, I think we make quite a big effort to make the sensors as small as possible because they are much easier to integrate. So yeah, I think there are some options where you can have more bulky systems, but we always try to make them as small as possible.

I've spent quite a bit of time lately talking to different display manufacturers and R&D companies about emerging displays, and one of the things I talk about is with stuff like micro LED, there will be a time when you'll be able to include a sensor right within the display. So just because the pixels are so small, you could put a sensor device in there as well. So you're not just looking at a screen, you're also looking at something that can do some degree of sensing. 

Have you started to look at that at all, or is that just out there somewhere? 

Hubert van Doorne: Yeah, we do see that, I think I can take a parallel to a couple of years ago. We were quite growing big in the room booking systems that we did little LED strips that you could add to the system and then you saw the first screens popping up where they had the little LED strips integrated, and we are not selling any LED strips anymore for booking. That's just where the industry goes. But I don't see that as a bad thing, because I think, for example, if Samsung decides to add a presence sensor in every one of their screens, then that's the next step where we say okay, then maybe our sales of single presence sensors will go down, but there are other sensors again, where you would like to measure again. 

So I think it's sort of the same as you see media players integrating into screens. If people are using it a lot, it makes sense to make it in one unit. I'm happy to see that everything gets less complex and is being integrated into one unit, and I think it's more our job to, for all those things that cannot go into super high volume, that there is space in every screen, then maybe you need a present sense or just looking at a different angle then right in front of the screen. That will be something that's never added to the screen. because it's a typical application. 

There are two components, right? There's the sensors and then there's controllers. How does that break down and do you need both? Obviously you need the sensors. 

Hubert van Doorne: Yeah, and you do need the controllers. What we did with our hardware structure is that we have a lot of sensors and they all come back to the same little connector. It's what we call Xtalk and it's like a mini USB, and that's actually every sensor can be hooked into any controller, and then the different controllers we have is just to oversee the whole system setup and to have one, yeah, let's say a sort of USB hub where you have one communication channel to the media player and our controller is then, really doing the power supply for all sensors, seeing what's connected and being the host the communicator to the media player. 

And we have different controls because for example, for something like LED lighting, you can get a controller with an LED driver already embedded on board, so it makes it much easier to control LED strips and the same for example for audio switching and we have a lot of different controllers that have different functionalities built in just to make it more easy and have it integrated in one system. 

How do you know if everything's behaving as it's supposed to? 

Hubert van Doorne: It's a good question. The nice thing is that our system, our controller is completely monitoring all sensors and the system integrator can also always see if all sensors are online even exact serial numbers and everything that's connected. 

It looks like very simple tech, but in the end, you can, within a dashboard, for example, completely monitor your system health and see even if somebody is plugging something in a different port. You can already see that the system's setup has changed. So that's how we can monitor and see if something is not working or needs attention over time. That's something that's really needed if you do large-scale rollouts, and I think, yeah, everything we build is just made for maintaining it or giving it a possibility to roll out in big volume.

Is that your device management dashboard, this is purely an API that you can integrate into a system integrator or whatever. 

Hubert van Doorne: It's just the API where we deliver that, and we have then our partners that can deliver you a complete dashboard already that you don't have to build something yourself, but if you have the as a system integrator, I want to build my own dashboard, you're free to use our API and you don't need any license. Then it's just free. You buy our hardware and there's no license cost. 

Is there an emerging kind of tech that you're now looking at that you think you can add to the stack or could be a big deal like lift and learn and presence sensing has been around for a very long time. People who've been in this industry and doing retail solutions understand it, but is there new stuff coming? 

Hubert van Doorne: Not so much on the sensor part. For example, last year, we added weight sensors also for lift and learning so it's a completely different technology. Radar technologies are improving. So yes, we always continuously monitor what's available and there will be new additions, but I think that will not be complete wow in the industry now there is the Holy Grail that this is possible. It will be just additions over time. 

What we do see is that there is a very strong appetite for green signage products and what we will do is next year we will launch a new product in that area so that we can also deliver to more verticals than only the retail and experiences as we become more and more interested in other verticals. 

I was curious about digital signage. If you work in the industry, you think it's quite a big industry, but in relative terms, it's quite small and narrow. Is there enough of a business for you just within digital signage, or is it important to branch out? 

Hubert van Doorne: No, I think it's definitely big enough for digital signage now because there's almost hardly any competition. I always say our biggest competitor is no sensor. So in a project somewhere that the sensor is being skipped, but, it is fast growing and I think digital signage itself is growing, but I think also you see much more areas where it becomes, yeah, it's a discussion, is it digital signage or, it's a screen somewhere in an area, if you call it digital signage or not, we see more and more screens coming up and more and more data becoming available in the digital world. 

So I think, yes, the number of screens will be quite growing in the next few years and I think, there's a very big market we can service. 

Yeah, I've written a number of times and talked to, or when I've done talks with companies, about what I call boring signage and the idea of using sensors that maybe you would think about for parking garages and available stalls and so on, and using that in airports for very busy restrooms and things like that to have displays out front that kind of give you the state of availability of that. 

It sounds dumb, but I think it's incredibly valuable and again, it's just applying sensors in a creative way. 

Hubert van Doorne: Oh, yeah, and we see regularly that our sensors are being used in completely different order verticals. So your toilet example is one where they used our sensors, and even to see if the bin is full in the restroom. So it's not that we are not able to provide those markets. It's only that we communicate and we read the development for the retail space. 

But yeah, you will see more and more areas opening up where it becomes interesting and, that's something we definitely can branch out of, over, over time as well. For the moment, we are really focused on the retail space because I think there's a lot to win still, but yeah, the sensor technology will go not to retail or experience centers only. 

Retail gives you scale too that you don't necessarily get with washroom solutions. 

Hubert van Doorne: No, and I think very important is that, in the retail space, data is hugely important for AI to see. How can I do my shopper assortment? How can I see how people are behaving in the store? And actually by providing the sensors that can really measure. 

We have, for example, now a retailer in the UK that's only using the sensors to see how people are moving in front of the shelf. There's not even digital signage anymore, and I think, that gives the importance of knowing what people are doing, that they want to deploy a system in more than a hundred stores to measure to just see how they can improve and make better operation.

You're in Eindhoven? 

Hubert van Doorne: Yeah, that's right. 

How big is the company at this point? 

Hubert van Doorne: At this point, we are only a very small group. We are around 25 staff and now slowly growing our sales team as we see the market really starting to develop before it was a lot of interest and smaller projects and sometimes a bigger rollout of 500 stores or something, but now you really see, since last year volume coming in and that's really nice that we can now scale and now we are really from the sometime away from the startup and now a scale-up. So that's also where it will happen with the staff in the same way. 

You mentioned being involved with another company back in 2015? Are you completely a private company or do you have ties to a larger firm? 

Hubert van Doorne: No, it's a completely privately owned company. At that time, we were part of a display company, but that one stopped in business and this was completely carved out as an individual company and it's completely independently operating.

If people want to find out more, they get you on nexomsphere.com?

Hubert van Doorne: Yes, that's the perfect place to start, and if there's any question, please reach out to one of our team members or the emails. We are like not all companies always, but we are responsive to our emails, and we like to help people with any question they have because that's the first step in exploring what sensors can do.

You're at ISE, right? 

Hubert van Doorne: Of course, and that will be really exciting with some new products. You should not miss as an integrator, even if you're not interested in the retail space. So I would definitely see common visitors. 

All right. Thank you for spending some time with me. 

Hubert van Doorne: Thank you very much. It was a pleasure. Have a great day.