If robotics aren’t inherently capital intensive, does management in robotics just suck? Yes. Here’s why…
2012/09/03 3 Comments
I was harassing my asset management friends to get them to help me develop the synthetic short instrument I want to put into the robotic stock tracker and we started discussing capital use in robotics. Their question, was, “Okay, if robotics are not inherently capital intensive, why does it take more money to get a robotics company up and running? Isn’t that initial expense an inherent characteristic of the robotics industry?”
In a word, no. The fundamental problem with robotics companies is that management doesn’t have a well developed process for synchronizing customer and product development to use Steve Blank’s terminology. Or put another way, a lot of robotics companies spend a fortune on unnecessary engineering when they frankly suck at discovering what customers want. iRobot has a whole museum dedicated to their market failures. I contend that much of this engineering effort is not necessary to development of viable robotic businesses–this same learning could be done with vastly less expense.
Much of this problem comes from the difficulty of porting over rapid-cycle software development best practices for discovery of true customer needs. Most of our hardware development methodology comes from environments where customer needs are relatively well understood and engineering improvements require a lot of time. Robotics companies still have engineering cycle times (the amount of time to go through the engineering build, test, analyze, decide cycle) that are much longer than pure software companies at least 3 times longer and often much more as best as I can estimate from anecdotal evidence. Companies are very reluctant to reveal this information, so my estimate may be off by several factors, but it is clearly much longer for robotics companies.
I believe that we in the robotics industry need to tailor the customer and product development methodologies to the peculiar challenges of robotics. We will need to reduce cycle times of engineering teams down closer to software levels. 3D printing and continuing improvements in supply chain should make this feasible. Management should make it a priority and a reality, and be willing to incur some expenses to do so. Even more, management needs to do a lot of work to reduce market risk much earlier in the product development cycle.
iRobot’s museum show that it is proceeding to engineering while far too much of what is required to make a viable commercial product remains unknown. This isn’t to pick on iRobot, they may be among the best in the industry, but it is just to show that even the most advanced practitioners in our industry are not very good at understanding customers compared to other industries. Yes, for some customers, such as the government, just doing research can be a viable business model, but this won’t grow the industry. We need to develop ways to reduce market risk and we need to get good enough that we’re showing the software industry how they could learn about customers more and code less.
I don’t propose to give a complete answer on how to do this here, but it is clear that there is more than one path to reduce market risk in a product. Both Intuitive Surgical and Liquid Robotics seem to have taken the approach of building a robot that is so awesome and widely applicable that it will find a use even if it isn’t in the application that management originally intended. Other robotics companies, like Kiva Systems and RedZone (since Eric Close took over), seem to have taken a more traditional minimally viable product approach and iterated upon the original product. Both strategies appear to win in certain circumstances and companies that took the opposite approach in the same markets failed. How do we distinguish which set of market and technical circumstances we find ourselves in?
This interplay between technical and market risk and how it applies to robotics management is only beginning to be understood. Few people have proposed measurable distinctions that would allow management to make decisions about what risks to accept and what risks to mitigate before committing capital to a project. This area of research more than any other will unlock the potential for robotics to become the next tech boom.