Lost: 19 Billion Connected Widgets

I used to do a lot of writing. Back in my days at B&B Electronics, we peaked at over 50,000 opt-in subscribers to my "eConnections" monthly newsletter of tech tips, industry observations and general ranting.

One of those past subscribers reached out to me recently and forwarded me this eConnections issue from 2016. It's still relevant today, and a good reminder to use caution and a bit of skepticism when a technology is at the top of its hype curve. A number you might keep in mind is that in 2023 the forecast for cellular modules is in the neighborhood of 400 million units, so it does seem that when Ericsson cut it's initial forecast by 20x, it got them into the right ballpark. Excerpts from my 2016 newsletter below.

Maybe I read headlines with a slightly different lens than most, but I read some recent updates from the pundits at Ericsson with great interest, and maybe even a bit of self-satisfaction.

You see, about 5 years ago as fuel was being gathered for the IoT marketing fire, Ericsson called the market at 20 billion cellular connected devices in 2020. Not to be outdone, a number of other pundits jumped on and numbers from 30 to 50 billion connected devices became common forecasts. Now, depending on who was calling the ball, not all of those were going to be connected via cellular, but no matter how you sliced it, the numbers were big.

I love reading this stuff. It inspires me to imagine vast potential use cases, value propositions, technical possibilities.

But then I filter it against what I think I know about technology, about the rate of market adoption, about product development cycles, and about various markets’ speed, willingness and ability to spend money in order to save money.

And when I applied my set of “Mike’s filters” to the 20B…I was struggling a bit.

So a few months ago when Ericsson updated their numbers from 20B to 1B, I couldn’t help but feel a little validated.

Now 1 Billion is still a pretty respectable number. And remember, we’re talking only about cellular connected devices. And maybe that’s where the devil is in the details…

I've been in the connectivity space for a long time. And I can tell you beyond doubt that not every sensor needs, wants, or can afford a data plan, the associated power or cost of a cellular connection. That’s the obvious part. Less obvious, but maybe more important is the cost to provision and maintain every sensing device as a unique IP connection.

A car makes a nice example. If we look in my garage (and eliminate anything we find there older than me), we’d meticulously count somewhere around 80 sensors per car. And at various times I might like to know what they all have to say. But I can do that without 80 data plans. I can do that through a single gateway port, wired or cellular, if you prefer.

So I think what was missing in the original numbers was a healthy dose of reality to say that our IoT network of the future will contain data from tens of billions of devices, that data will travel over local sensor networks. Wired networks, wireless mesh networks, long range sensor networks. Most of it will aggregate through central gateways, or even networks of gateways. Many gateways will have wired Ethernet connections and many will be cellular. So, as so often happens, somebody forgot to divide. So now Ericsson decided to divide by 20.

Mike’s napkin math filter

Anyone placing bets? Talk back, love to hear your comments.

-Mike