Bosch Software Innovations’ M2M/IoT Application Platform. Good? Yes. And perhaps, just perhaps, it’s worth USD1.4bn.

Last week Machina Research spent a couple of interesting days in the company of Bosch Software Innovations, plus selected partners, at Bosch Connected World (BCW) in Berlin.

Bosch Software Innovations has spotted the niche in the platforms market that we refer to as ‘M2M/IoT Application Platforms’, and they hosted BCW to tell everyone about it. A quick re-cap: right now, most platforms (and most participants) in the M2M space are essentially vertically focussed. The telecoms industry, in particular, has spent much of the last few years trying to create vertically integrated behemoths to slug it out for market share. Into this arena steps the IoT-enabling M2M/IoT Application Platform, whose raison d’être is to offer a horizontal application development environment by abstracting across the jungle of vertical solution providers that there is out there. I first wrote on the topic of cross-operator, horizontally focussed, Service Enablement platforms back in 2011. These entities were very much a theoretical construct back then, of course, but now they are becoming a reality in the form of M2M/IoT Application Platforms. If you’re interested to find out more on M2M/IoT Application Platforms, check out this White Paper.

The demonstrations at BCW and the overall theme of BCW reminded somewhat of the approach that ThingWorx takes. The aim is to make the application developer king, of course underpinned by the kong of an open environment and standards (thank you @caro_buck, although I would assert that many of those ‘standards’ will be de-facto, at least in the early years). The application programming environment should be drag-and-drop simple and flexible across a range of verticals and applications. 

Bosch is doing much the same but, betraying Robert Bosch’s Swabian roots, the typical implementation leans more towards a centrally managed on-site deployment rather than a more fashionable cloud deployment. However, I’d assume that the balance will shift as Bosch S.I. seeks to address the smaller scale opportunities that it will need to support in order to be successful in this space.

On the flight back home, and with ThingWorx in mind, I took out my slide rule. I met Russ Fadel (CEO, ThingWorx) in early December 2013 at the ThinkMonk event in Shoreditch, London. There was a power cut, and we met in a room that was really quite dark, and got progressively colder as the meeting progressed. At that time ThingWorx comprised around 50 people, and was looking to expand substantially over the course of 2014. Of course, this envisaged organic growth never happened since ThingWorx was acquired by PTC for USD112m in late December 2013. That kind of money pays a lot of electricity bills. Applying a crude ratio of an implied value of USD2.25m per employee to Bosch Software Innovations “more than 600” employees (let’s call it 625) results in an implied valuation of USD1.4bn. Crazy? Probably. And I’m sure that there’s more to valuation than that, and I’d never recommend buying any stocks on my say-so, but Bosch S.I. does seem to have a significant presence in a fairly interesting niche.

Subnets of Things

I’ve been spending some time lately thinking about the Internet of Things, and began to wonder how we might get from ‘here’ (with an increasing number of connected devices, but typically connected in vertically integrated stovepipes rather than in a horizontally integrated way) to ‘there’ (the fully fledged Internet of Things).

The reality is that, right now, the world of connected devices could best be characterised as multiple ‘Intranets of Things’. For example many homes now have connected smart meters, but the data that these produce is generally used for a single purpose (analysing, pricing and billing power consumption). Likewise, the information generated by fleet tracking systems are generally used to better manage a fleet. And so forth. There are very few situations in which information collected for one purpose is used for a really different purpose, although there are some (for instance mobile operators selling location information to train operators, so that those operators know how many people are on their trains).

The next step must then be around integrated ‘islands’ of connected devices, which we could term ’Subnets of Things’. These would typically be driven by either a single point of control, single point of data aggregation, or potentially a common cause. For example, it is not hard to envisage an emerging subnet of things around a smart city: local authorities would often have access to data relating to congestion charging, public transport, parking space availability, air pollution and potentially a whole range of other data sources. It would not be hard for a local authority to analyse these data sources in such a way as to generate conclusions that are informed by multiple information sources. Similarly, a local health authority (or health insurance company) will clearly have access to information derived from multiple sources, and will clearly be incentivised to mine that information to gain new understanding of illnesses. Qualcomm’s 2Net M2M health platform (designed to support connected healthcare solutions from many different manufacturers) may potentially be another example.

But to move from these ‘Subnets of Things’ to a full ‘Internet of Things’ environment will be a difficult step. It will involve aligning data points from a huge range of data sources, ideally at an individual user, or individual device level. Cue all manner of privacy and standardisation issues. Establishing a fully-fledged ‘Internet of Things’ will be far harder than establishing simple ‘Subnets of Things’. And I wonder if there is much more value that can be derived from an ‘Internet of Things’ when compared to ‘Subnets of Things’. Sure, it helps if an ambulance driver has access to ‘smart city’ traffic flow information, and if the ‘smart city’ can react to ease the path of that ambulance. But that is surely a marginal benefit when compared to the benefits that can be derived from a (much more achievable) integrated health ‘Subnet of Things’, in terms of disease management and treatment?

I wonder if the key to the ‘Internet of Things’ is actually to look for the tipping points of a range of ‘Subnets of Things’?

Opportunities for cellular connectivity in consumer devices

Consumer electronics devices from TVs to games consoles are increasingly being connected to the internet to allow access to remote content. In the near future connectivity options will also be increasingly integrated into an ever wider range of consumer devices, ranging from home control systems to fridges and washing machines. The Consumer Electronics, Intelligent Buildings and Healthcare sectors will together account for approaching 9 billion connected consumer devices by 2020, but these devices will be overwhelmingly connected by short range technologies.

The debate around the potential benefits of adopting wide area cellular (WWAN) technologies for connecting various consumer electronics devices, instead of short-range technologies such as WiFi, has been ongoing for several years, and full consensus has yet to be reached. In this Research Note we set out the arguments for, and against, the use of WWAN technologies for connecting various types of consumer device.

There are a number of arguments in favour and against the use of WWAN for connecting these devices:

• Out-of-the-box connectivity
• Homogeneity of connectivity
• Wide area cellular connections for resilience
• Cost disadvantages of wide area cellular technologies
• Mobile capabilities as a differentiator

I’ve explored these dynamics in some detail in a recent Machina Research Note, but the overall conclusion is that mobile operators should focus on mobile WiFi (aka MiFi) and embedded connectivity opportunities where they can add real value, (including through mobility, resilience and significantly increased utility). Short range communications technologies will suffice in most other situations.

Lots of connected Consumer Electronics. And some fridges.

By 2020 there will be in excess of 4.2 billion M2M connected Consumer Electronic devices in use worldwide, the majority of which will be connected Audio Visual Sources (1.7 billion) and Displays (1.1 billion), primarily driven by consumer demand for web-TV and internet audio-sources.  Machina Research doesn’t count tablets and eReaders in the Consumer Electronics category (we have a separate forecast category for PCs, tablets and handset data), but including tablets and eReaders would add another 900k devices (that’s including media tablets, enterprise tablets and building control tablets – these devices will become platforms, in much the same way PCs are today, and a touchscreen is really just another human interface device), bringing the total to 5.1 billion devices.

The White Goods market will begin to adopt M2M connectivity towards the end of the decade, as smart metering and pro-active energy management become more prevalent. By 2020 the market for M2M connected Consumer Electronics will be worth EUR400 billion, with Europe and Emerging Asia Pacific being the largest regional markets. However, with the vast majority of devices likely to be connected by means of short-range technologies, and with some applications being very data-hungry, connected Consumer Electronics devices are more likely to be a headache for fixed network operators than a revenue opportunity for mobile network operators. Unless the mobile industry can significantly reduce the cost of embedded modules, that is. This is something of a hot topic for me at the moment – I’ve blogged on it before, and I’ll be investigating module costs in some more depth over the coming months.

The main driver behind the adoption of M2M connected Consumer Electronics devices is the potential for connected devices to offer a better user experience than non-connected equivalents. The second big driver will be the potential for M2M connected devices to be more energy-cost efficient by timing periods of heavier power consumption to coincide with periods when electricity is cheaper. But there is no sweet-spot. There is no Consumer Electronics device for which M2M connectivity can both significantly improve user experience and substantially reduce operating costs.

And, yes, many fridges and freezers will one day be connected. But not to tell you what you had for breakfast that morning, or to send you an email that you are running low on milk. The ‘killer apps’ for connected fridge-freezers are ice making and defrost cycles. Both of these functions consume significant amounts of power, and can be easily shifted to times of day when power is cheaper. Connected fridges and freezers won’t be a big market opportunity any time soon, but they will be one day. And, anyway, a better way to monitor your stock levels of various household essentials would be to place RFID-, or barcode-, scanners near household bins. If you really wanted to do that, that is.

Machina Research Global M2M forecasts

It’s been a busy summer so far at Machina Research.  We’ve been working hard to finalise our first Global M2M forecast, including devices, traffic and connection technologies (all for each of 54 countries, and 5 ‘Rest of’ regions). I thought that it might be useful to post an interim update – the figures below are the total number of M2M connected devices in each sector by 2020, in 000’s.

M2M Devices in Intelligent Buildings                                          4,900,000

M2M Devices in Smart Cities & Transportation                           work in progress

M2M Devices in Automotive                                                      1,400,000

M2M Devices in Consumer Electronics                                      4,200,000

M2M Devices in Healthcare                                                      770,000

M2M Devices in Utilities                                                           1,500,000

M2M Devices in Manufacturing & Supply Chain                          work in progress

M2M Devices in Retail & Leisure                                               140,000

M2M Devices in Construction                                                    33,000

M2M Devices in Agriculture                                                       82,000

M2M Devices in Emergency Services & National Security            51,000

M2M Devices in Smart Enterprise Management                           work in progress

M2M Mobile Broadband Devices                                                 work in progress

The figures listed above already total somewhere North of 13 billion M2M connected devices by 2020.  Mobile Broadband devices such as tablets, phones and laptops don’t strictly qualify as M2M according to our definition (“Connections to remote sensing, monitoring and actuating devices, together with associated aggregation devices”) but it looks like Matt (co-founder of Machina Research) has already identified around 9 billion such devices, and only counting those connected to WWAN networks.  More musings and meanderings on mobile broadband at Matt’s Wireless Noodle.

Machina Research will be publishing our Global M2M forecast in September, and I will be posting the occasional update here as we continue to develop it.  The report will adopt our usual approach of discussing drivers and barriers for growth of the relevant M2M markets and will be supported by an Excel data sheet including granular 10 year market forecasts for 54 countries and 6 regions. The forecasts will cover numbers of connections, traffic and revenue for each of the sectors, together with splits by technology (LAN, MAN, 2G, 3G and 4G, etc).  More detailed information on any of the sectors is/ will be available in our specific sector reports, as we publish them.

Mobile industry closing door on M2M in consumer electronics

From a recent Machina Research press release: “The role that mobile operators, and mobile connectivity standards, might potentially play in the consumer electronics market is severely limited by the cost of WWAN embedded modules. Chipsets for Wi-Fi and other short range communications technologies generally contribute under USD2.50 to Bill of Materials costs, and often less than USD1.00. By comparison, 3G embedded modules cost of the order of USD30-60, with 4G costing more still, and effectively price the technology out of the highly competitive Consumer Electronics market. If those prices don’t fall, then the mobile industry might just have to sit out the opportunity for M2M connected Consumer Electronics and stick to already established mobile broadband devices.”

There are three factors underlying the relatively high costs of WWAN connectivity:

* Any consumer electronics device incorporating WWAN connectivity really has to be multi-mode (2G+3G) and multi-frequency -- there's no real way around this (other than in the case of extremely low bandwidth devices, with sufficiently short expected lifetimes that 2G switch-off will not be an issue) and the increased complexity of the underlying chipsets increases the cost of those chipsets.

* IPR costs (fees for Qualcomm, mainly, and mainly in the case of 3G) are charged essentially on a per-device basis.  That's fine in the case of a smartphone where 3G data transmission rates really improve the user experience: the additional upfront premium seems worth it. But in the case of a digital photo frame, the premium for 3G connectivity doesn't look like quite such good value.  An easy way around this problem (well, easy in theory, if a little harder to implement) would be to persuade Qualcomm to accept royalties on the basis of 3G traffic carried, rather than 3G devices sold. This would reduce the licencing costs of typically low-bandwidth M2M devices. Qualcomm would win too, since a lower IPR payment received in relation to a 3G connected digital photoframe is better than nothing.

* Interoperability testing is expensive, particularly when expressed as a cost-per-device for devices with relatively low expected sales volumes. All devices with WWAN connections need to undergo interoperability testing of some form before they can connect to a carrier network, but it's complex and expensive.  A CE manufacturer can reduce the testing burden by using WWAN modules, rather than chipsets, but, again, these add to BOM costs. From the perspective of a CE manufacturer, all this just serves to make WWAN connection options less attractive relative to WLAN connection options. Unfortunately, the trend here seems to be for mobile operators to close internal test departments and outsource testing capabilities in a tightly controlled way. Yes, the closed-system/ walled-garden strategy is getting another spin. If mobile operators took a more open approach to interoperability testing, then the market might benefit from competition between platforms and approaches. There would be scale benefits too, as individual test houses could be expected to have relationships with a greater number of mobile operators.

Of these three factors, it's probably the cost of interoperability testing that's the showstopper, and we're unlikely to see any breakthroughs in IoT cost levels anytime soon because of the relatively closed nature of the market. Reluctantly, we probably have to conclude that the mobile industry has effectively closed the door on the vast bulk of the potential M2M consumer electronics market (at least where viable alternative technologies, such as Wi-Fi, exist).  Conversely, opening the market for interoperability testing might just have the potential to be a bit of an iPhone moment.