This is the year of the IoT Hockey Stick according to Computerworld in “Soaring use of IoT in corporate networks” while citing Vodafone’s IoT Barometer.  On the other hand, Enterprise IoT Insights asks Nokia “What’s holding back the IoT?” and hears “there is a fundamental structural problem in the industry.”  General Electric finds the IoT platform business is hard, as mentioned in “GE is shifting the strategy for its $12 billion digital business.”

Looking for a firsthand account and insight on the Machine Economy, I turned to Mike Coffey, a veteran of the IoT C-Suite, to get an insider view.  In case you don’t know, Mike’s technology C-Suite experience is deeper than IoT, whose recent C-stints include KORE Wireless and Wyless, having spent his prior life in SaaS-enabled companies.

What is gaining traction in the IoT business?

There is a lot of activity right now and islands of success depending on where you look.  We are certainly seeing early enterprise uptake increase as more firms find success with initial deployments.  There are also more IoT enabling devices coming to market, more networking technologies, and of course a never-ending stream of applications which all lead to more creative business model experiments.  It’s an exciting time to be involved in this space.

Where are the big deployments that will lead to the billions and billions in the next decade?

It’s always “smile a little” when we talk about these billions and billions of devices because right now, 100k endpoints is a big deployment that could be driving a lot of economic value for those involved.  Some companies, such as utilities, automakers, and the largest of OEM’s are now pushing millions of devices deployed.  But a utility solution is entirely different from a fleet solution or a smart city solution.  We have a very fragmented ecosystem of technologies and companies, and where we find big successes are in particular markets.  There is no standardization or commonality in the way IoT solutions are implemented, and each vertical remains unique.

Each one is unique? Some in the industry claim over 400 IoT Platforms are on the market, suggesting a lot of professional services.

Yes, that’s right.  The ecosystem is complicated and not everything will communicate right out of the box.  Every IoT deployment has the same basic components which start with the devices.

The device needs to connect to something by wired or wireless means.  The connection carries the data to some place, typically to a cloud, and stores the data, also typically in a cloud.  We need an application to make intelligence out of the data (and there are many layers to storage, AI, processing, etc. in that application) and lastly, there is the user experience which sits on top of the whole stack.

If you examine each of these layers, starting with the devices, you will find there are thousands of components in the ecosystem.   It is complicated to decide which to use, whether it is an Off-the-Shelf device or custom-built component.

As for the connections, there are really only a dozen or so ways of connecting.  It starts with are you wired or wireless? If it’s a wireless connection, then you’re asking questions such as does my application move, does it involve far or close communication, or is it even more short-range, perhaps inside only or device to device?  From there you can choose from a host of ways for communicating – cellular of course which has a tremendously developed ecosystem, and the non-cellular Low Power WANs (Sigfox, LoRa, and others) of which the ecosystem is still finding its momentum, and short-range connectivity (Wi-Fi, Bluetooth, and Zigbee for example).

Considering the applications – there are many of those, and deployed on over 400 platforms sounds right.  There are firms that store data, that manipulate data, that provide development toolkits and security toolkits, and on and on. The point is that it’s all very complicated, and everything is hard to put together and solutions are not simple.  It all has to fit together, talk together, work together, and most importantly create end-customer value in a repeatable way.  Where the mistakes have been made over the years is that all too often companies end up recreating a specialized wheel.

What do you see as the solution to all this?

Time and some market consolidation will drive standards, and winners will emerge.  As time goes by, more large deployments will gain more traction which will drive more standardization and homogenization of technologies.  More large consolidators will get involved in pulling a lot of this together.  Of course, we’ll have tons of new startup as well, but one thing about IoT is once you’ve built a solution of scale it’s awfully hard to swap out.

What part of the value chain is the best place to be right now?

Most of the value generation happens close to the data, the application, and the end customer.  There are a lot of challenges facing the hardware vendors, device vendors, and even on the CPE routers.  Hardware manufacturing is a tough business, tends to be one-off in nature, with a heavy R&D investment.  They are all looking to provide greater value to the industry and find recurring revenue business models.

On the connectivity side, there is increasing competition and more large players, even the mobile operators themselves getting involved.  Whether it is with licensed or unlicensed spectrum, the mobile operators are very aggressive on IoT.  However, in the near-term, the emergence of various cellular IoT focused solutions, such as LTE Cat M1 and Cat NB1 is having a real impact on the alternatives.

In general, everyone is trying to integrate up the value chain seeking recurring revenues, and larger firms are entering the connectivity space.  Both the hardware and connectivity markets are fundamentally challenging over the long run as they inevitably face commoditization.  The best place to be is close to the data, application, and customer.  This is where the greatest value is built, and built with software and solutions

If the hardware is getting commoditized, what is the risk of picking a technology, and the vendor that goes by the wayside?

We have seen it time and time again over the last 10 years.  A firm making a risky technology choice can get burned – badly.  When it builds a technology into the IoT solution and the underlying technology goes bust, it has a major impact, especially if the failure is in the hardware or in network connection.

When it comes to IoT, take metering for example.  These may be put in peoples’ homes, on the streets, and deployed globally.  A truck roll might cost $100 and if the need to service 100K or a million assets arises, it can get rather expensive.  One must make the right technology decisions, or else there will be serious ramifications.  My advice is work with big players, work with proven technology, and be slightly cautious of cutting-edge technology.

Can you mention an anecdotal IoT catastrophe?   

We have had a lot of firms building personal technology devices like smartwatches and things of that nature for protecting or tracking people and then had battery failures.  Put one on a patient or loved one that you need to track, expecting a multi-year life and 1,000 recharge cycles, only to have it fail prematurely, and you have lots of anxiety.  There might even be significant legal and regulatory ramifications.

One of the worst I can think of from several years ago was a deployment of ankle bracelets for criminals.  They used a cellular partner to connect and monitor those ankle bracelets, but because of a series of bad agreements and decisions throughout the whole value chain, the cellular partner went bust.  They had something in the order of 10,000 offenders and criminals roaming the streets that could no longer be tracked.  This is a problem not unique to IoT: build a business (like tracking people) and if the technology decisions turn out badly when everything is in the field, you’ve got a serious problem.

Could this network risk be solved with an LTE eSIM, which can switch among telecoms? 

That is the ultimate promise of eSIMS, so yes.  The challenge is that eSIMS are really only being implemented for the largest of players at the moment (auto OEM for example) and have yet to get down to the mid-market (100k and under deployments). However, the emergence and progress of the GSMA standards is a reason to stay with cellular.

What about 5G and the mmWave spectrum for IoT, 28 GHz for example?  

There are a number of IoT use cases that work for 5G.   LTE Cat M1 and Cat NB1 works for 90% of what we see for cellular in the market today, so I am not a fan of saying “wait for 5G,” but 5G will be great for last mile use cases.  In the last mile, you have rich data and rich content delivery over relatively small distances.   Take a connected car for example.  If a self-driving car needs to stream lots of data about what’s on the road, then 5G is perfect for that.

Or consider residential last mile, like the cable ISP, which has put a lot of fiber and wired infrastructure into homes.  In a reasonably dense city environment, a 5G micro-tower on the end of a street can supply a wireless connection without a truck roll to install in the home as primary internet connections can be replaced by 5G.  The CLECs, T1 wired businesses like banks and retail institutions, can all switch to 5G, running everything from the 5G micro-tower.

However, because 90% of today’s IoT use cases can be served with LTE, 5G IoT will tend toward the heavy data applications (until 5G becomes ubiquitous and then we could perhaps be talking about LTE-sunsets – oh my, I really did say that…).  But in-short, organizations should deploy LTE IoT solutions now, there is likely a long 10-year life ahead of them, and then 5G will be great for the ultra-high data demand use cases and we’ll watch it evolve

….

This interview was conducted as part of the Machinomics research service at SAR Insight and Consulting, where we anticipate the game-changing shifts the Machine Economy will bring to the enterprise.  Current Machinomics research for the enterprise includes IoE Networks, IoE Topographical Business Intelligence (aka Wardley Maps), Blockchains, and Machine Learning.

Contact the author: Joe@SARInsight.com