A couple of years ago I was talking with a company that wanted to build robots. Their proposed price point for the robots was about $20K per unit. It would be a brand new product, that people had never really seen or experienced before, and it would replace some amount of human labor.
After thinking about this some, I was extremely puzzled. Putting myself in the shoes of the customer, here’s how I would think: “So, you want me to pay you $20K cash up front, for a product that has never been tested before. If it breaks down, I can’t go to anyone but back to you to have it serviced. And meanwhile, I’m stuck as if I’ve replaced the human labor with robotic labor, and the robot breaks, then I have to scramble to find a stop-gap solution, while the robot is repaired or serviced, which could take at least a few days, if not weeks. Nope, I’m not buying it.”
There had to be a better solution. And the better solution is to really think of delivering Hardware-as-a-Service (HaaS). In the example above, instead of charging $20K up front to sell a robot, the company could instead propose a very low threshold for the customer to buy by saying: “We’ll deliver you the robotic labor as a service. You’re probably paying $3,000 per month for the human labor at this point. We’ll provide you the robots-as-a-service for $1,500/month. We own the robots, so you’re still expensing the cost of the service. If a robot breaks down, we will replace it with a new unit within n hours (could have a backup unit on-site, or deliver a new unit within 24 hrs). There is no risk for you to try it. You could start with 1 unit, phase it in, and then expand by adding additional robots over time.”
If the company adopted this model, not only would they have more revenue ($24K/year vs. $20K one-time) from the sale of the robot, but they would end up with a recurring revenue model, which would make for a much better business model for the company rather than selling complex widgets. This model does introduce a cashflow problem where the cost for producing the initial units has to be borne by the company and probably won’t be recovered for at least 6 months into the contract. However, it reduces the commitment required by the customer to buy (potentially shortening the sales cycle), results in more revenue for the company, and in general leads to a better alignment of incentives on both sides to ensure that the project succeeds. The customer can also be certain that now they won’t be left hanging with some complicated machine that no one else in the world knows how to service, maintain, repair, and most importantly upgrade.
More and more as I’ve met with and invested in different hardware companies, I keep coming back to this Hardware-as-a-Service model. I think HaaS can be an elegant approach for companies producing complex hardware products, or hardware products which require complex software backends/services to deliver real value to the customer.
The best of both worlds is when the customer is willing to pay the upfront cost of the hardware, and is willing to pay for a subscription service on top of it. Some of the companies that have done a phenomenal job on this are Meraki, Dropcam and Ring. Kudos to Sanjit Biswas (Meraki), Gregg Duffy and Aamir Virani (Dropcam), and Jamie Siminoff (Ring) for figuring this out and getting it right.
The hardware-as-a-service model is what I loved about Better Place. The company didn’t succeed in implementing its vision (for reasons that I have absolutely no insight into), but it did paint a picture of the future that I think is inevitable. Tesla is well positioned for moving into such a model (Transportation-as-a-Service) in the future and Uber would possibly be well positioned for it too.