We need a new name for the Internet of Things (IoT). The term has become so widely used that it is starting to lose its meaning. Postage stamp-sized bluetooth tags on every UPS package. A $5,000 refrigerator. A sensor network covering a 10,000 acre oil field. Personal massage devices. Connected dog collars. Drones. Thermostats. Burglar alarms and door locks. Someone, somewhere has labeled all these things as “IoT”, despite having almost nothing to do with each other. We need more than a new name, we need a new way to describe and categorize all these things.
Often when we discuss IoT with clients and colleagues we resort to describing them it in terms of industry verticals – Home Networking, IoT for energy, IoT for logistics, etc. But this is problematic. For the companies supplying the systems for all these things, this approach presents a hopelessly fragmented market. How do you know what your market share is across three dozen verticals? We have tried to build these market models. They are possible, but they are also very hard to build. Let us know if you would like to have one.
To make matters worse, there is considerable disparity within these fields. If we talk about IoT systems for energy are we referring to the wide area sensor network in the desert or the system used to keep track of delivery trucks? No IoT customer has a Vice President of IoT, instead they have titles like VP of Digital Transformation, or just head of IT.
So really this is a problem for the suppliers, especially for their marketing departments (who hold a lot of responsibility for the problem to begin with).
At this point, we should admit that we do not have a new name. This is very much a work in progress. That being said, we do have a few suggestions for an improved taxonomy. One way to look at the market is to divide the systems up by their connectivity method. We touched on this last week when we classified all the ways to connect IoT modules. We could also classify the market by the type of digital logic used in the system, or even the way the module interfaces with the real world. This is handy, but also risks the blind men and the elephant problem – with a tendency towards myopic focus on very small sub-segments.
Another way to look at IoT is by price point, or more elegantly, by system complexity. The Wi-Fi and Bluetooth connectivity modules that Qualcomm sells in its IoT division are very different from the $0.10 Bluetooth tags coming to the market soon. Digging a layer deeper, there is a growing distinction among the various IoT solutions on offer. An advanced consumer video drown that sells for $1,000 needs to have advanced location capabilities to comply with aviation regulations, to optimize for power, and provide advanced low latency connections to the controls. By contrast, a tracking module for a shipping container attached to a car battery needs only some fairly rudimentary functions. The difference here is the former requires some real design work, while the latter can make do with an off-the-shelf Quectel module.
We think there is a useful classification system in this. Divide the market by the amount of engineering resources required. High end, semi-custom solutions need the support of large chip makers like Qualcomm. Low end, off-the shelf systems that come with no design work and minimal support can purchase from the catalogs of the module makers like Quectel and Sierra Wireless. And then the systems in the middle can make use of Silicon Labs burgeoning IoT platform. This is not a handy system for industry analysts as the dividing lines are very blurry, but we think it is a much more useful framework for companies as they assess their product roadmaps. Some companies will want to operate at the high-end, others at the low-end – allocate engineering resources accordingly.