Making Sense Of The Industrial Internet Of Things (Lessons From GE Digital)

by: Yves Sukhu

The Industrial Internet of Things (IIoT) market is evolving and holds great potential for technology companies.

But the market is still nascent and nebulous, as evidenced by GE's stated exit from the market despite a $1 billion+ investment in its Digital unit.

With just about every technology company claiming to be an IIoT player, there are some companies which arguably have a better chance of success than others.

1.0 Executive Summary

Originally, the intent of this report was to examine what went wrong with GE (GE) Digital such that the company now plans to auction key pieces of this business which "…loses money despite billions in investment". My initial impetus for that analysis was to reflect on what I wrote in a 2017 Seeking Alpha article on the unit, determine how beneficial the potential sale would be to GE, and to evaluate how the sale would impact an investment thesis for the company.

However, as I was writing the article, it transformed into a consideration of the broader Industrial Internet of Things (IIoT) market; the market which GE Digital had set its sights on. I think there is value in this, given the aura of hype that surrounds IIoT, and its companion or parent market (depending on your point of view), the Internet of Things (IoT).

I, for one, have compared the current IIoT market to the early e-commerce market in the late 1990s when, suddenly, every technology vendor was an "e-business". Using that comparison, there was clearly a significant opportunity for electronic commerce, but it took far longer to materialize than most were willing to admit or aware of, and many technology companies that claimed to be e-commerce leaders were niche players at best.

Accordingly, this report attempts to make some sense of what the IIoT market is and what it isn't in an effort to determine which "players" may be worth investing in. Further, I carve out a section for a "look back" on GE Digital to ascertain what went wrong, and how that information may also help us decide which IIoT technology companies have the best chances for success.

2.0 IIoT: Hard to Nail Down

In the most general of language, IIoT refers to the hardware, software, and services associated with the control, management, and analysis of devices and machines. Companies may use IIoT solutions to become more efficient, drive quality, and even bring new products to market via lower-cost, improved manufacturing processes. In the same "IIoT-breath", we can be talking about a sensor used to monitoring pollutants in a water pipe and those technologies used to operate and automate an entire factory. This link from IoT Analytics does a good job highlighting some of the top IIoT use cases, as segmented by geographic area:

Source: IoT Analytics

The basic story underpinning the IIoT market is that the number of connected devices/machines is expected to grow exponentially, along with the data thrown off by those machines.

Source: IoT Analytics - IoT Platforms Presentation

Simply, the scope of IIoT is very broad, which is both good and bad. It's good in the sense of a large (some might say unbelievably massive) total addressable market (TAM), but at the same time, it's "bad" because it can be very difficult to pin down exactly what the market is. For the purposes of this article, I make one overarching distinction between IIoT and IoT with the former referring to industrial applications and the latter referring to consumer applications, even though the enabling technologies may be the same or similar between the two markets. I have seen other authors/analyses refer to IIoT as a sub-market of an "all encompassing" IoT market; however, let's just assume they are distinct albeit with some degree of overlap:

Diagram Source: Yves Sukhu

There are good reasons to be excited about the potential for IIoT applications, and that has a lot to do with the underlying technologies. A PR News Wire article on the subject suggests "…technology trends which are expected to have a huge impact on [IIoT] evolution are IPV6, sensor proliferation, cloud computing, Big Data, and faster communication standards such as 4G-LTE…" In other (perhaps oversimplified) words, technological advancements will allow us to control, manage, and analyze more things than we have in the past. It will also enable entirely new operating models; as an example, the same PR News Wire article briefly mentions the exciting prospects around machine-to-machine communication and management.

Suffice it to say that IIoT solutions, as they are described by most analysts and experts, encompass a multitude of technologies. IoT Analytics offers a simple diagrammatic representation of the key components comprising an IIoT platform:

Diagram Source: IoT Analytics - IoT Platform Presentation

The software component is broken down further into building blocks as follows:

Diagram Source: IoT Analytics - IoT Platform Presentation

So, just how big could the IIoT market be? I quoted Forbes contributor Louis Columbus in my 2017 Seeking Alpha article on GE Digital (Note: As we'll discuss, I actually argued in my article that investors should be cautious in regard to GE's forecasts for GE Digital and the IIoT market). In Mr. Columbus' article, he notes:

  • The Industrial Internet has the potential to deliver up to $11.1 trillion in value on an annual basis by 2025, with 70% or $7.8 trillion of that value captured by business-to-business solutions. Source: GE Digital, The Emerging Industrial App Economy.

  • Investment in the Industrial Internet of Things (IIoT) is expected to top $60 trillion during the next 15 years. Source: GE Announces Predix Cloud - The World's First Cloud Service Built for Industrial Data and Analytics

These are some big numbers, yet they may be possible given all the potential applications with IIoT. Even so, consider that the U.S. economy is expected to generate around $20 trillion in goods and services for FY '18. So, the second bullet suggests 3 times the economic output of the entire United States in FY '18 could be invested in just this tech area over the next 15 years, although it's not entirely clear what "investment" means here, just as the definition of "value" in the first bullet is nebulous. Even so and despite current challenges with the Digital unit, why would GE exit a market with this kind of value capture and investment potential? After all, GE has invested billions thus far in its Digital unit. If these forecasts are even just "slightly" correct, is there not a reasonable opportunity to recover that investment over the long-term?

You might counter that GE doesn't have a choice; it has to sell GE Digital to raise cash and concurrently reduce its expense outlay. I get that. But still if those forecasts above - with numbers anchored by a "T" for trillion - are "in the ballpark", then it seems like they are exiting a massive growth area. For example, GE's Aviation business unit is one of its largest by revenue, bested only by the company's Power unit. Boeing (BA) noted in an aviation market outlook report from 2014 that it expects $5.6 trillion to be spent on aircraft through 2034. So, again, if the IIoT numbers above are right, they dwarf the entire expected worldwide spend on commercial aircraft over the next 15 years!

Let's take a step back in time for a moment. Here's a great analysis from IoT Analytics from 2014, where the author attempted to make sense of multiple, divergent forecasts:

The author states:

"…Gartner's numbers appear extremely conservative…On the other hand, IDC's [numbers] appear extremely high: 7 trillion in 6 years…I have not considered [GE's] numbers here because I find the forecasting method questionable."

It's notable that as far back as 2014, the author of the IoT Analytics study had problems with "all over the map" IoT forecasts and, specifically, the way GE was developing its market predictions. Fast forward to 2018, and we still have an endless supply of growth estimates for the IIoT market. For example, here's a recent GlobeNewswire article estimating a market size of $232.15 billion by 2023 at a CAGR of 8.06%, and then here's a MarketsandMarkets report estimating a total market of $91.04 billion by 2023 at a CAGR of 7.93%. Obviously, there's some significant difference between the two sizing calls. But it is notable that both are significantly lower in their projections as compared to the forecasts from the IoT Analytics report from 2014. It is also notable that they are nearly in agreement in their CAGR estimates. Maybe the deviation in various expectations is decreasing, which could be a good thing.

Still, it's probably fair to say that no reliable estimate for the IIoT market exists because:

  1. It's really not possible to agree on what the market is and what it isn't.
  2. For a variety of reasons, including the immature nature of the market itself, it's been difficult to analyze the market over recent years, and to build forecasting models off those analyses.

Of course, that leaves investors who have interest in this area in a bit of a quandary. Naturally, there's quite a bit of risk investing in a market that lacks definition, not to mention one where one of the biggest players (i.e. GE) appears to be heading for the exit.

Let's take a look at what may have happened with GE Digital in an effort to identify IIoT players that might be better positioned to succeed.

3.0 Lessons From GE Digital

GE was cruising full-speed ahead with its GE Digital ambitions as recently as its FY '15 Annual Report where the company noted "…GE can become a top 10 software company by 2020" and GE Digital is expected to "…expand [the company's] growth rate…[and] improve [the company's] margins…". What happened?

Obviously, something didn't work out as planned. Alex Moazed wrote a fantastic article "Why GE Digital Failed", which is included in the Appendix. He provides an excellent historical summary at the beginning of his article:

"GE's digital transformation efforts have been ongoing for many years, dating back to 2013 with references to Predix, which GE intended to be its software platform for the industrial internet.

Then, one year later, in 2014, GE released a press release that said it was generating more than $1 billion in revenue from productivity solutions, highlighting Predix.

In 2015, GE announced the creation of a new business unit, GE Digital with a CEO, Bill Ruh, who was previously the VP of GE Software.

And, finally, in 2016, reports came out that GE Digital had more than 1,500 employees in its San Ramon office in California."

Noting that the picture painted by GE's management appeared "rosy", Mr. Moazed notes:

"Despite its grand aims, in practice GE Software was created as an internal development shop... Much of the revenue GE Software [and eventually GE Digital] was generating came from billing other GE business units, not external customers"

I made this same point, albeit a bit more indirectly, as per my 2017 SA report:

"Not only are IIoT analyst forecasts all over the map, but so are the numbers for GE Digital. In the company's FY 2015 Annual Report, GE reports $5 billion in revenue for GE Digital for 2015:

But, in the FY 2016 Annual Report, GE reports $3.6 billion in revenue for the unit:

This is confusing (to say the least), but perhaps explained by GE's modified definition of 'what' counts as GE Digital revenue.

Here is the definition from the FY 2015 Annual Report:

'Digital revenues - revenues related to software-enabled product upgrades, internally developed software (including Predix) and associated hardware, and software-enabled productivity solutions. These revenues are largely generated from our operating businesses and are included in their segment results.'

And here is the definition from the FY 2016 Annual Report, with the definition modification underlined:

'Digital revenues - revenues related to internally developed software (including PredixTM) and associated hardware, and software solutions that improve our customers' asset performance. In 2016, we reassessed the span of our digital product offerings, which now excludes software-enabled product upgrades. These revenues are largely generated from our operating businesses and are included in their segment results.'"

So, the company was perhaps "playing a bit" with GE Digital's numbers. Still, why did the unit presumably fail to demonstrate sufficient external growth, leading to management's "swing of the axe"? Mr. Moazed suggests the following in his article (I apologize in advance to Mr. Moazed if I have misinterpreted/misrepresented any of his points):

  1. GE Digital was established as an independent unit from GE Software in order to give it more autonomy and freedom to innovate. But that independence came with P&L responsibility and "heavy" organizational structure which drove the unit to seek short-term versus long-term results.
  2. GE Digital's short-term focus manifested itself in the form of projects with internal (GE business unit) customers and partners who may have been incentivized by GE to develop Predix solutions.
  3. GE Digital may have morphed into more of a consulting services organization versus a software technology unit, as it began selling services to other industrial manufacturers to help them with their own digitization efforts. Financial returns on these efforts may have been subpar.

Mr. Moazed suggests that both GE Software and GE Digital were/have been "set up" for failure. They didn't/don't have a chance to succeed due to the inflated expectations of the parent, their large (organizational) size, and a slow-moving GE process structure. He states very clearly that GE Digital would have needed to be spun out as a separate company in order to succeed. Maybe that is so; the phenomenon of "innovation" units straining under heavy corporate structures is nothing new.

But I draw a few corollaries from Mr. Moazed's points above, though they are strictly my interpretation:

  • GE Digital may have felt a partner-based approach would be more efficient getting Predix technologies into the "hands" of end users as opposed to building out a large salesforce to achieve the same result. Of course, the latter can be quite expensive.
  • GE Digital's external marketing may not have been sufficient.
  • GE Digital may have begun focusing on consulting revenues over software revenues because Predix technology adoption was slow/limited.

So, if those corollaries are correct, it actually may be that (1) there was/is something about the Predix ecosystem such that adoption was slow by both partners and end-users; (2) GE Digital may have had an over-reliance on partners as opposed to a direct sales team/organic marketing to push Predix into the market; (3) the unit's partner strategy may have been ineffective for whatever reason; (4) IIoT market growth did not match expectations; and (5) some combination of (1), (2), (3) and (4). In regard to the (1) first inference, it's noteworthy that Google Trends data for the term "predix" demonstrates more of a downward trend since 2016:

Source: Google Trends

If, indeed, Predix adoption was sluggish, it could have been due to the proprietary nature of the Predix platform, as well as a lack of open standards for the IIoT market. Dr. Tom Bradicich, Vice President and General Manager of Servers and IoT Systems at Hewlett-Packard Enterprise (HPE), notes "open standards are key to expanding IIoT and innovation in the marketplace". One might argue that in its attempt to dominate the IIoT market early, GE was a bit "ahead of the times" in that sufficient penetration of standards and underlying technologies has not yet occurred. It certainly would not be the first technology company to falter in this regard.

In any event, if all the inferences above may be correct, what IIoT market players can offer investors an opportunity by avoiding GE Digital's missteps?

4.0 A Simple Analysis of Players

If you were to put the names of most enterprise tech companies on a dart board, and then toss a dart at those names, chances are that you'll hit one with a page on their website dedicated to their "role" in IIoT. With so many companies claiming to be part of the IIoT revolution, who should investors consider as the most relevant players? IoT One provides a "Top 100" ranking of IIoT companies, and the website Postscapes offers another ranking scheme for publicly-traded companies, although Postscapes' analysis is more broadly for IoT companies, not specifically IIoT (For readers who are interested, Postscapes also provides an elegant listing of startups in this space).

Simply put, these lists are long.

It should be obvious from a cursory examination of both lists that a broad swath of companies are "making a run" at IIoT, from traditional technology companies (e.g. IBM (NYSE:IBM)) to telecommunications leaders (e.g. Verizon (NYSE:VZ)) to semiconductor manufacturers (e.g. Intel (NASDAQ:INTC)) to industrial manufacturing giants (e.g. Siemens (OTCPK:SIEGY)). A comparison of the top 50 companies between the two lists is somewhat revealing in that only 11 companies are shared among both:

Granted, one list (the IoT One list) is IIoT-specific, and the other encompasses a broader portfolio of companies. Even so, perhaps this is to say that there is still no clear consensus on who the leaders in this market are, and that may be, coming back to an earlier point, related to a market that is poorly defined at this stage.

For the purposes of this article, and this section in particular, I chose to deliberately focus on the IoT One list which is IIoT-specific. In fairness, the list is dated from 2017 and does not include (arguably) newer players to the space such as Alibaba (BABA). Nevertheless, I put together a simple spreadsheet of the top 15 companies from that list, where I married IoT ranking data along with P/E, P/S, P/B, and dividend yield (where appropriate) data. Certainly, this information is high-level, but my approach here is to establish a basic guide that provides insight into the top players. I repeat that the analysis to follow is only the top 15 players from the IoT One list which might lack some companies that readers might be expecting to see; Amazon (NASDAQ:AMZN) Web Services for example, is on the list but not within the top 15.

If I make the broad assumption from the previous section that GE Digital may have run into trouble partially due to the proprietary nature of the Predix platform, the spreadsheet data allows us to examine IoT One's ranking in this area against financial data. Here is how the spreadsheet data looks, with the companies ranked by "ecosystem openness":

Data Source(s): IoT One, Yahoo! Finance

Table Source: Yves Sukhu

As we see, Intel is at the top of the pile when we rank by ecosystem openness. I also like Intel in this space given its ability to play in multiple areas of the IIoT platform stack, as supported by its experience both in hardware and software. This position is bolstered by IoT One's spider chart of its capabilities:

Chart Source: IoT One

While not necessarily cheap in terms of price-to-book (P/B) and price-to-sales (P/S), the company is perhaps more sensibly priced based on these metrics as compared to many other tech companies today. For these reasons and along with its 2.6% yield, Intel is my personal favorite on the list in terms of making a play on the IIoT market. Frankly, if there is one company on this list that should play well within the IIoT market, it is Intel in my view.

Arguably, Cisco has a similar value proposition in this space as Intel:

Chart Source: IoT One

I am already long Cisco, and I think shares are a little over-heated right now. So, I am not adding to my position at the moment. But readers who are hunting for solid IIoT players might wish to do their own analysis on the company to determine if it's a worthwhile addition to their portfolio at the current valuation.

I am a bit hesitant to "jump on" what I consider to be the mostly software-centric players on the list, namely SAP (SAP), IBM, Dell (DVMT), Google, and Microsoft. As I've argued - in so many words - in several of my previous Seeking Alpha articles (and this article), every software company will tell you they are a player in IIoT because the scope of IIoT is so broad, they can easily "massage" an IIoT story. So, it's somewhat difficult to tell who is really serious about the IIoT market versus those who are just jumping on the "hype train" (And don't be easily fooled by the companies that tout their IIoT-focus after spending a few million to acquire some start-up that itself had jumped on the IIoT "hype-train"). With this in mind, if I had to choose among these 5 companies, I would lean toward IBM and Microsoft because of their cloud market share and existing enterprise market penetration. While it's beyond the scope of this article, I am concerned about what has been going on, what is going on, and what is going to happen with IBM long-term. That leaves Microsoft, whose Azure cloud-presence and broad technology stack theoretically enable the company to become a significant player in the software-oriented component of the IIoT stack. But the stock is a bit expensive at the moment in my opinion, so I personally would not touch it for that reason. Although (again) readers should do their own analysis.

Of the communication services companies on the list, AT&T, Verizon, Ericsson (NASDAQ:ERIC), and NEC, all should all clearly benefit from IIoT device proliferation, as the requirements for bandwidth and throughput continue to increase. I like Ericsson here for the same reasons that I like Intel: it has a broad portfolio of technology and capabilities, and its current valuation is (at least) not completely outrageous.

Chart Source: IoT One

And now we are left with the industrial companies on the list, ABB (NYSE:ABB), Siemens, GE, and Bosch. I don't consider Bosch, which is a private company. We've already discussed what happened with GE's Predix, and we all know what's going on with the company more broadly right now. So, from an IIoT-specific perspective, GE would seem to be a poor investment choice right now (Let's see how things play out with GE; its story within the IIoT market may not be over just yet). I like both ABB and Siemens. Both of these companies have been "doing IIoT" before IIoT was even a buzzword. It seems logical that their respective years of experience in industrial applications, systems, and operations should drive wider penetration as the IIoT market evolves. As with Intel and Ericsson, neither stock is cheap but neither is wildly overpriced either.

I reiterate that the analysis above is simplistic, and readers need to do their own homework. But, to summarize, my 4 IIoT company choices are Intel, Ericsson, ABB, and Siemens:

Data Source(s): IoT One, Yahoo! Finance

Table Source: Yves Sukhu

Given that I lean to dividend-paying large-caps, these companies obviously fit my traditional investment model. If we revisit some of the challenges that may have affected GE Digital, such as lack of system openness and poor sales execution, the companies above are not immune from the same mistakes. At the same time, we see Intel and ABB already ranking fairly high in terms of ecosystem openness according to IoT One, and it's entirely likely that the technologies offered by Ericsson and Siemens may trend in a similar direction as the overall market establishes standards (Again, GE might have been a bit early in the game with Predix). Also, each of the companies have well-established sales and marketing engines within industrial and enterprise markets. That, of course, does not guarantee success since GE also had such resources. But perhaps these companies might be able to examine GE Digital's missteps in an effort to focus their IIoT sales efforts more effectively.

As with many evolving markets, there is a good chance we might see a player that emerges from a humble start and grows into market dominance within IIoT, just as we have seen Amazon grow from an unpretentious book seller to a business conglomerate with worrisome ties to the United States intelligence community (Why Amazon's Collaboration With the CIA Is So Ominous - and Vulnerable). I wish I knew who the "Amazon of IIoT" will be. Until I can figure that out, I'll take my chances with the 4 companies above.

5.0 Supporting Documents


Disclosure: I am/we are long CSCO, IBM. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: I may initiate positions in INTC, ABB, ERIC, SIEGY within the next 72 hours.

Editor's Note: This article discusses one or more securities that do not trade on a major U.S. exchange. Please be aware of the risks associated with these stocks.