IoT has had a bad case of expectation inflation for a few years now. It feels like we are seeing a bidding war between analysts to determine who can offer the largest predictions of connected devices. Many of these estimates were promised by the year 2020, which is now just over two short years away. The industry is certainly growing and loaded with opportunity, but I suspect that there will be a scramble of downward revisions next year.
Remember the famous Ericsson prediction of 50 billion devices by 2020? It is now revised to 18 billion IoT devices by 2022 (https://www.ericsson.com/en/mobility-report/internet-of-things-forecast). It is great that everyone is excited about the possibilities of IoT (as am I), but the enthusiasm got a little ahead of the challenging reality of deploying lots of new sensors and devices. In order to monetize data from sensors, you need to first build and deploy these physical devices generating the data at scale. Connecting the physical and digital worlds involves getting your hands dirty.
Imagine if an enterprise customer went to a vendor and expressed interest in purchasing 100,000 smartphones. Then imagine if the answer was something like this: “Sounds great. Why don’t you pick the chip, module, device maker, antenna partner, connectivity, and the entire software stack that you think you need on a device and application layer, and then get back to me.” You probably would not sell too many phones this year, and you would also scare the customer from thinking about buying smartphones for quite a while.
This is how much of the industry is selling IoT today. If you are a connectivity vendor, you first talk about how to connect things. If you are on the hardware side, it is all about solving the device problem. Application development platform vendors just assume the data will come to feed the applications being built. Very few organizations are beginning with tightly defining the business problem, and then selecting the correct technologies. More importantly, they are finding that the software and hardware solution that they usually need does not exist, and that they need to create it.
I know of a partner in Europe who has been working on an IoT solution to detect forest fires. We are reminded of how important this is with the recent fires in California. After over eighteen months of development the sensor works perfectly, but the problem is that birds keep nesting on the device and occasionally cover it with bird droppings. The droppings then keep the sensor from working properly! The point is that until you run trials, you never really know what will happen in the field.
Multiple partners in the logistics space are in the process of deploying millions of sensors on low cost assets. Working through the challenges of deploying hundreds, thousands and then tens of thousands of devices a week regularly runs into the reality of rough environmental conditions, aggressive forklift drivers and faulty electronics. Many great electronics manufacturers are simply unaccustomed to the challenges of deploying at scale in challenging physical situations like asset tracking, agricultural and facilities management applications. It will come in time, but there will be many lessons for all of us along the way. In the short term, believe in the promise of IoT, but expect the big adoption curve to continue to drift to the right.