Why IoT hasn’t scaled
Why IoT hasn’t scaled
Pundits and salespeople have been talking for years about how the Internet of Things (IoT) will include billions of devices in just 5 to 10 years. While there has been growth, most of the opportunities for IoT have yet to manifest. There are a lot of reasons why, but here we’ll address one of the key reasons we see – the need for the device to be specific to the application.
Often, there’s a misunderstanding surrounding the “thing” in IoT. Particularly when it comes to the mass production of them. In general, IoT devices tend to be custom developed and specific to their exact application. This is required to achieve the business mission, and includes environmental factors, reliability, size, weight, and keeping the cost of the device within an allowable range for the business case. This custom development effort takes time, money, special equipment, and expert engineering minds to execute.
In this world of mass production, automation and instant consumer gratification, it’s natural to assume IoT products can easily scale like any other IT device. But these devices aren’t necessarily the same as “new” technology that came before them. Laptops, for example, are easier for companies to scale into 15 to 20 different models, as you then can make hundreds of thousands of each model. To get 1 billion IoT devices, the equivalent wouldn’t be as feasible. Instead of producing 1 million units of 1,000 different device types, you need to produce something like 100,000 of about 10,000 different device types. Moving that single zero means you need 10 times the number of development engineers, special equipment, money, and, the real killer, time. IoT is higher mix, lower quantity. It’s a wide (lots of SKUs) but shallow (thousands not millions of each SKU) product distribution.
There is currently, and will be for the foreseeable future (20+ years) far too few engineers with the skills to develop these products. IoT devices are fairly complex for a number of reasons and RF engineering isn’t a common speciality. These factors create one of the major constraints on ecosystem growth – development capacity.
Some other reasons why IoT hasn’t scaled:
- Most opportunities remain unseen. IoT is as applicable to any business as desktop computers or the internet. Most of the uses are not tech companies, but the rest of the business community. Most of these businesses simply don’t know what IoT can do for them.
- There aren’t enough IoT experts. Most companies implementing IT solutions for business automation either have internal experts or can go to a wide array of IT consulting companies to figure out exactly what they need. Companies that specialize in how IoT plays into those same needs are much less common.
- The “neato” gets in the way. Sometimes there IS someone who’s heard of this “IoT stuff” and they go and create the piece of the system they’re interested in. That could be the device, the cloud backend, a front end web site or whatever. The common situation is that a hammer is getting built before anyone knows if you’re building a ship, a plane or a space station. This results in a “solution” that often doesn’t address any real business need
Let’s take the application example of fleet tracking. Typically only one device type is needed to fulfill this service offering for a given business. The device is typically a box that goes under the dash or seat of a vehicle and it’s connected to vehicle power. The vehicle’s communications bus and antennas for the various radios used in the device. The number of copies of that box that are needed varies with how many trucks or other vehicles a customer wishes to manage. It could be 50; it could be 50,000. That box has a useful lifetime of at least 5 to7 years or longer. So if you’re a truck freight company with 10k trucks, (FedEx for instance had around 86k when this was written) you’ll need 10k units to deploy over the next 6 to12 months and then depending on if you’re recovering devices from vehicles as they go out of service or abandoning them, you’ll need to make a few thousand a year for new vehicles.
For experienced developers it costs around $750k to develop a fairly simple custom device and it takes 9 to18 months depending on the level of application software and cloud software that needs to be custom. We’ll round that up to $1 million. That development cost across 10k units is $100, before you get to the actual cost of goods. The advantage of this situation is that the trucking company owns that IP, so the cost of goods could end up being only $50 vs buying a sufficient device on the open market and paying $100 to $300 per unit. But you’re still talking about $150 per unit at 10k. If you only need 1k units because you only have a thousand trucks, that would be $1500/unit. If there is an off-the-shelf device you can use, that can save you a lot of time and money, obviously. For fleet tracking, there is a fairly wide array of options. But for many other applications there are few or none.
Now, why would you do any of that? Well, in the fleet management example you’d get automatic toll payment, automatic driver logging, predictive maintenance that would prevent breakdowns, increased fuel economy, and increased safety, as well as data on direct costs of specific services, routes, vehicles, drivers, and customers. Basically, you get hard data on which to make business decisions and justify those decisions to stakeholders and you get it without having dozens of humans whose job it is to obtain and manage that data. You also have the option of selling or trading some of that data with outside entities that may value it such as governmental departments, industry groups, or adjacent businesses, such as insurance and maintenance companies. Once you turn on a stream of data, it will provide continuous value over time.
Right now, IoT is still before the general inflection point where it becomes a default part of any business’s basic processes. We’re past the early adoption stage, the technology is all well proven and in some industries like transportation. They’re already on next generation devices to add increased value and capabilities.
Until we reach that inflection point, F3 will be here to help with any and all of your IoT development.