When discussing strategy for technology companies, there is this debate that seems to come up once every five years or so: do we chase the latest fad/trend trying to capture near-term revenue, or do we slog it out in segments with longer design-in cycles and slower time-to-money?
In the Web 1.0 phase, which roughly coincided with a shift from analog to digital cellular technology, it seemed every embedded firm started chasing telecom and datacom infrastructure deals. This was usually at the expense of pursuing defense, industrial, and medical segments – labeled as “cash cows” and marked for minimal investment. It worked for several years, until a bunch of companies with high stock valuations and divide-by-zero multiples imploded. A plague of layoffs hit embedded suppliers that had geared up for bigger business, and an industry-wide pullback was on.
Web 2.0, the e-commerce push, left out many embedded firms who were now gun shy, recovering from the bubble burst and retreating into the more traditional markets. Cash cows were back in style, and in fact grew at a healthy clip. There was a slight problem developing, however: the surge in inexpensive PC hardware technology that accelerated during Web 1.0 put enormous cost pressure on the OEMs selling more durable, but much lower volume gear. This drove companies to create new niche form factors and drive further into segments with more ruggedness, bigger price tags and less cost sensitivity.
A change came with Web 3.0, the social, mobile, and post-PC era. With fabless SoCs, open source software, and 4G infrastructure hitting stride, the flow of ideas swung heavily to the other side. Consumer electronics topped $1T as chips showed up in everything from music players to TVs, with the smartphone and tablet at the head of the parade. Mobile replaced the PC as the volume driver, and while high-end SoC development costs swelled, low-end MCU and SoC unit costs plummeted.
This ravenous technology-creating engine we built in the first three phases of the Web has to be fed, or we risk a slow death by starvation. The PC segment is flat, no matter what Intel says – while a complete collapse is unlikely, significant growth in the old familiar places is not coming back for anything but perhaps a brief curtain call. Many observers have also declared the smartphone as mature, locked up by Apple, Samsung, and Qualcomm.
We enter Web 4.0, the Internet of Things, looking out at a transformed landscape and somehow knowing we are at the point of no return. The new world is out there, and we must go in search of it, but nobody knows exactly how to get there, how long it will take, or what we will actually find.
IoT technology is in place, perhaps too much of it, with ample options and little agreement on standards so far. Devices can connect to each other and the cloud easily. Big data can be stored, sifted, and shaped. The maker movement, powered by inexpensive hardware, learn-to-code, and a new wrinkle of crowdfunding means anyone with a prototype and a hashtag can get in front of consumers. The barriers to access have largely disappeared – but not necessarily the barriers to success.
The first adventurers have plunged out into the void. Nest, Oculus VR, and Pebble are among the pioneers finding treasure, and danger. Nest, now flying the Google flag after being boarded, has suffered its first mass recall. Oculus VR, on orders from the King of Facebook, isn’t even sure they can make a profit on their first million yet-to-ship units. Pebble has so far been the exception to an alarming wearable abandonment rate, navigating side-by-side with the mighty Samsung armada and trying to stay ahead of the Apple fleet lurking behind the horizon. Following them are a myriad of startup landing craft of all shapes and sizes.
Meanwhile, some conservative voices are calling for tech companies to stay at home, resisting the temptation to sail for consumer IoT markets. They cite the obvious hazards: unrealized volumes, undefined or unproven use cases, unbridled hacking, and unattractive ROI. Most embedded types still snort at mobile and social, saying they have nothing to do with embedded, portraying John Chambers as some kind of heretic. (Cue my walkup tune, courtesy of Avenged Sevenfold.)
The traditional embedded segments – now encompassing automotive, aerospace and defense, industrial automation, medical diagnostic and imaging, and similar applications – certainly beg for an approach loosely referred to as the “industrial IoT”. It looks very familiar, much more of a sure bet, and a lot easier to sell to existing customers and prospects like them with existing channels and support.
Nobody has repealed the embedded design-in cycle, however. Many of these applications require significant capital investment, lengthy qualification or certification, and a lot of patience for returns to materialize. Fads need not apply. There is definitely always going to be a niche for safety and complexity when it comes to technology.
But, we need to face facts. Taking nothing away from the embedded community, particularly the microcontroller and FPGA types, the substantial progress made by embedded has mostly been possible by travelling in the wake of the PC dreadnaught and the smartphone heavy cruiser ahead. Without those volumes, we’d still have comparatively expensive DRAM, FLASH, and LCDs, and technologies such as ARM, Ethernet, USB, Wi-Fi, MEMS sensors, GNSS, and Bluetooth wouldn’t be nearly as ubiquitous.
These enormous projections for the industrial IoT from GE and others? Possible, if we wait long enough, but the catalyst of consumer demand for IoT devices is needed to speed the process and draw investments and more ideas. Volume demand is the difference in propulsion between nuclear power and oars, and without volumes from consumers the IoT will move slowly and tentatively.
What about the failures, the breaches, and the just plain daffiness of some consumer IoT efforts? It’s easy to yell FAD in a crowded auditorium. Unfortunately, many IoT pioneers have mistaken “maker”, someone just like them with a higher tolerance for acts like configuration and maintenance, for “consumer” who can drive a smartphone and a Tesla but not a BeagleBone. Bad UX design, lax security protection, lousy battery life, and other technology blunders abound. Falling in love with cool tech without really tapping into the consumer’s frame of reference usually generates a miss.
Launching into the consumer IoT is not for the faint of heart, or for those seeking guaranteed ROI. For those businesses, the industrial IoT shore – where the embedded business will rally next – is an OK place to be. For the technology sector to regenerate and grow again, we have to take consumers along for the journey. As Andre Gide wrote:
One doesn’t discover new lands without consenting to lose sight, for a very long time, of the shore.
On the IoT, everybody knows you’re lost – and that is a good thing. Wave if you see me, agree or disagree.