Last week, I started a deep dive into new business models enabled by the IoT. This week, I’m going to continue with a few more, including some that companies want to develop, but need more buy-in from customers first. Literally.
In the “already here” category, I’m going to look at the pay-per-use business model in which a vendor uses connected data to charge customers based on when they use a product and/or how they use their product. The most popular example of this is insurance, where firms charge customers for coverage based on when and how they drive. Pay-per-mile options are familiar to most, but others include selling insurance based on the occupancy of a building or charging users each time they use a shared machine such as an electric scooter or even an MRI machine.
Most people are familiar with pay-per-use business models, but licensing digital twins is so new most companies probably haven’t encountered it yet. And the final business model I will cover — that of sharing in overall value — is at this point more myth than reality.
In addition to insurance, pay-per-use is an accepted model in cases where companies or individuals might share assets. That’s because connectivity gives the owner of a machine or the service provider two critical elements: data about how the product is being used and its location. An app can then provide information about who is using the machine or service. This model is great for standalone products that aren’t part of an overall ecosystem or complex process. For example, you probably wouldn’t want to pay for each use of an essential tool on a factory floor.
It can also act as a gateway to charging users a monthly service fee, which we talked about last week. The caveats to consider are that heavy users end up paying more under this model than they might under a monthly service fee. If they are really heavy users they will likely want to own it themselves, unless there is a complex maintenance and servicing aspect.
Another word of warning for vendors considering this model is that they need to deeply understand the costs associated with owning and operating the equipment so they can charge appropriately. If they can’t achieve some expertise or economy of scale, it may not make sense to operate a fleet of machines or to build out a service and offer it on a pay-per-use basis.
Let’s move on to digital twins, which can still be somewhat nebulous. Some define a digital twin as a digital model of a machine, a building, or even a process that is constantly updated (ideally in real time) as the elements about it change. Others define a digital twin as a set of mathematical models that dictate how a machine will behave. Both definitions acknowledge that a digital twin is an attempt to create a digital record of a machine, and that that record has value.
For example, GE used a digital twin of a wind farm to calculate how to optimize each individual turbine for maximum power generation. Newport News Shipbuilding sold the U.S. Navy a digital twin of a ship the Navy purchased so the Navy can track its repairs and maintenance in a digital record.
When licensing a digital twin, the license might be a one-time fee or an ongoing payment that’s associated with continuously updating the model to reflect the changing state of the machine. The digital twin might be presented as a proprietary file format or as an algorithm. Today, it’s generally sold as an add-on service.
If you’re selling a digital twin of an existing machine, building, or process, consider that you might need staff on hand to keep it up to date. And ask yourself who can capitalize on the data generated while using a digital twin. For example, if GE’s digital twins can help a utility generate more power, GE might want a cut of the revenue. Similarly, if a digital twin created and offered by a real estate firm helps save tenants money, the commercial building operators might want a cut.
And that gets us to the final, and frankly, mythical business model that vendors consider when connecting their products — sharing in the value created by the connected device. I remember sitting with the creator of IBM’s Watson onstage at a GigaOM event years ago listening to him talk about the benefits Watson’s AI could bring to an organization and the plans IBM had to charge customers a percentage of their savings for using it.
The idea that a company can come in, attach sensors to a customer’s equipment, perform some proprietary AI, and save them so much money or time that they will be able to share in the benefits through some sort of deal is still with us today. But as Jason Urso, CTO at Honeywell Process Solutions, says, what often happens is that customers recognize the value and just want to pay for the solution instead of forking over an ongoing payment that may change over time.
As he points out, understanding the data and putting in the equipment has high up-front costs. And once the vendor deeply understands the process, the customer often does, too. When the customer recognizes the potential for savings, they generally take the product as a monthly service, turning the beginnings of that business model into a long and expensive sales process for the vendor.
Such sales also rely on data, but in times of unprecedented events (like a global pandemic or an ecological disaster), those sharing-in-the-value models can break, which leaves the vendor relying on inaccurate assumptions. Vendors may also tie their financial returns to the wrong metrics, which can also lead to trouble down the road.
Another challenge is getting customers to agree that the vendor’s calculations of savings or revenue generation are correct. Plus, many IoT projects bring in multiple parties to build a solution, so figuring out how to apportion value among those players will be tough, especially if the solution combines technology from larger and smaller companies.
I’m not sure we’ll ever see this last business model become viable, but I know there are still plenty of executives trying to figure out how to make it work. Maybe one day, some customer will buy it.