Over the past couple of years, many enterprise IoT applications have produced disappointing financial results. Several high-profile industrial IoT projects (which I will not mention here) got bogged down in implementation and were pared back substantially.
Some IoT products and solutions have done okay, but most have underperformed expectations. As of June 2019, Gartner’s highest-rated IoT Platforms — PTC, Software AG, and Hitachi — still had not risen to the ‘Leaders’ (Magic) quadrant.
What happened? Were the revenue models wrong? Or was the tech to blame? And what will happen next?
In Feb 2019, I wrote an article (The Strategy of Selecting an IoT Platform) that explained why this was likely to happen with IoT Platforms. But these same problems carry over into any IoT application and most IoT products.
First, there were too many undeveloped moving parts — edge devices, sensors, AI chips, protocols, software components — and far too many customizable solution possibilities for anything with the name ‘IoT Platform’ to make much sense.
The idea that any particular company, say GE, could throw a net around the entire IoT market, brand it a ‘platform,’ and make it all work as a product should was impractical. The fact that much of the technology was also rapidly changing and maturing at the same time made the effort that much more grandiose.
This is still true today for the most part. But, as IoT and edge components continue to mature, certain types of projects are beginning to deliver great results. More on that below (or better yet, sign up to receive our newsletter if you want regular updates on emerging technologies in this market).
Next, it is easy to underestimate the size and complexity of IoT. IoT market projections are massive — $1.2 trillion by 2022 growing at 13.6% CAGR. That is roughly double the size and growth of the global cloud market in nominal terms. (Note that more recently, some cloud projections have been updated to include edge data centers and other edge components that cloud service providers now offer — this makes the projections higher and blurs the lines).
my latest book
Both the technology set and the market for ‘IoT’ are extremely broad. The term is being used to describe just about anything that is not a traditional web application.
In fact, it is so broad that the first step I recommend for most people is to remove the term ‘IoT’ and replace it with the problem to be solved: e.g., In-Home Diabetes Health Revenue Model.
The reason for removing IoT from the discussion is more than just semantics. IoT is an enabling layer; an architectural layer. AI, edge computing, 5G, machine learning — these are all building blocks that enable new business cases. As an example, in the 1980s, databases were a new architectural layer that emerged from ordered lists.
Databases did offer efficiencies and new ways of conducting business, but these advantages were available to everyone else too. As databases were adopted by more businesses, they carried declining strategic value and became something that everyone basically had to use to stay in business.
The same is true for IoT. And since a revenue model is a framework for generating income, not an architectural design, it should focus on the offering, not the technology.
The exception to this are companies that gain their advantage through a deep understanding of a well-defined technology, such as 3D modeling, augmented reality, deep learning, computer vision. But these revenue models write themselves.
At this point, you may be thinking, ‘Hey! I thought we were building an IoT revenue model.’ Okay, I lied in the title. We’re building a revenue model that is based on new opportunities that IoT may be providing. Focusing first on the problem you are solving helps avoid the biggest mistake I’ve yet to encounter in IoT — unanticipated alternative solutions.
Step 2: Avoid Unanticipated Alternative Solutions
There are two ways you can build an IoT product. You can start by looking at the problem you are solving for the customer, considering all possible solutions to that problem, and picking the best one. Alternatively, you can start by considering the technologies first and building a solution that can be applied to certain customer problems. The latter approach is more vulnerable to unanticipated alternative solutions.
As an example, consider the city parking problem. In many cities, private parking garages that use static pricing currently maximize revenue around 50% occupancy and cost significantly more than public parking. This causes a problem for cities.
One official told me that 35% of traffic in the city was attributed to people driving around looking for cheap public parking spaces to open up in lieu of parking in a garage. Clearly, a better solution for all parties involves on-demand pricing.
In this case, several IoT companies bid solutions that would place sensors in each parking space to detect cars. This would allow parking facilities to know how full their garages are and to adjust prices accordingly.
This would work great. The problem, however, is that AI computer vision is a better solution for this particular problem. It is cheaper and easier to use an AI computer vision system to tally cars as they enter and exit.
Whenever you start with the technical solution, say sensors that measure factory equipment, and then look for problems to be solved, you run the risk of being circumvented as technologies mature and as new options become available. This should be reflected in your revenue model.
That reminds me. I promised to mention certain types of IoT products that were beginning to generate great results. I spent the last quarter speaking with a number of IoT and Edge companies in various markets. There are several companies reaching a critical point in their offering, and computer vision companies are notable in this regard.
Michael at Chooch.ai provided an incredible set of business cases for their platform and hardware products. Training times are fast — a few hours — and implementation is relatively easy. Chooch.ai is offering cost-effective solutions to problems that could not be solved at all just a few years ago.
Another notable company in computer vision is Adlink. Paul Wealls showed me some excellent real-world examples where their products are solving previously intractable problems
Step 3: Value Proposition Guides The Revenue Model
The value proposition your product delivers should guide your revenue model. Too often, whatever form the technical solution takes ends up determining the revenue model.
IT projects, in particular, have a tendency to begin as technical solutions rather than problem definitions. Many IT projects that I see start something like this: “We want to put sensors on X and let our customers see everything on a dashboard. ” As you know, that approach did not work well for the parking folks.
Once you commit to the solution and lock-in costs, it can be very difficult to change the way you generate revenue from it. A more effective approach in my experience is to pick the game your company gets to play.
To illustrate, consider this graph:
This graph is only intended to be an example to help you generate ideas. As you design your own revenue model, pick whatever axes work for your particular offering.
I hope this article has given you some thoughts on developing your own brilliant revenue model. If so, or if you need a sounding board, please feel free to contact me.
Finally, for good measure, make sure to remove ‘IoT’ from anything that has the word ‘Revenue Model’ on it!
If you love emerging tech, feel free to connect on LinkedIn or sign up for our newsletter. Also, see my latest book on Amazon: AI, IoT & the Intelligent Edge: Building Your Enterprise Tech Strategy.
Daniel Sexton is a Founding Partner at RedChip Ventures. Prior to RedChip, Daniel was a Managing Partner at a private investment fund for 6 years where he helped lead and manage investments in technology and product companies. Daniel has over 15 years of experience leading large-scale, technology solutions for Fortune 500 companies, such as Genuine Parts Company, CitiGroup, and Blue Cross Blue Shield. He has founded two tech companies and worked with a number of tech startups both as a partner and an advisory board member.
Daniel graduated from Georgetown University and completed graduate studies in CS at The University of Tennessee.