How Manufacturing Practices Prevent Innovation
Some engineers defend the slow pace of innovation in hardware by pointing to the fact that atoms are inherently more unwieldy than bytes. It takes more effort and expenditure to make changes to atoms than to bytes, particularly in a firm that is manufacturing at scale.
But the atoms-vs-bytes difference is far from being the whole story. Existing manufacturing processes are also slow because they are designed to be slow. If a team of engineers in a large automaker wants to redesign a car door, they may need to wait ten years or more for the firm to pay off the $100 million that it spent on a door mold machine before a different design can be made.
Today, a $100 million manufacturing mold with a ten-year replacement cycle is usually not the only solution that is technically possible. With modern 3-D printing techniques, it’s possible to make rapid changes at a relatively low cost. The problem is a management problem: the unwillingness to do anything about the sunk cost in that $100 million door mold machine, and all the other hard-to-change manufacturing equipment and processes.
These hard-to-change processes were good enough for a 20th-century car. What the customer got on the day of purchase is what the customer got to use for the rest of the car’s life. The customer wasn’t asking for anything more. But it’s not good enough in a marketplace where customers are coming to expect what Tesla already delivers: a capability to improve their cars every month or so after purchasing through software updates. These updates aren’t just cosmetic changes to the display. For instance, Tesla owners might suddenly discover after an update that the rear seats are now heatable.
But the problem is not just manufacturing equipment like the hard-to-change $100 million door mold machine. It’s also the thinking and attitudes that accompany that equipment. Investment in hard-to-change $100 million mold machines and similar equipment can come to be seen as the necessary way in which manufacturing should be carried out. Anything else is questionable. You do the design and you install your $100 million mold machine and you are set for ten years. That is until you discover that your customer wants something different.
Hardware engineers and their managers can come to justify such thinking to themselves and find solace in the thought that they know what customers want in a car. There’s no need to test whether their underlying assumptions are correct or whether tastes may have changed. They are not seriously thinking about the possibility that customers who thought they were buying a transportation device have now discovered that what they really want is something different. For managers in a car manufacturer, it may be simpler not to inquire too much into what customers think: they bought the car, so they must be happy. In this way, the $100 million mold machine can come to be seen as an unchangeable aspect of the business.
Traditional software development used to have a similar problem. Large software companies with teams of specialists were working in dedicated facilities, going through many years of planning and implementation to produce big hard-to-change systems that might or might not come close to what they had hoped to build back when they started, years before. Often, they failed to be completed at all.
Experimentation with Agile software development practices in thousands of firms over the last two decades showed that it was better in complex rapidly changing situations to design products in a modular fashion and iterate rapidly towards steadily better solutions. The work is broken into small pieces. Self-organizing teams work in short cycles getting constant feedback from customers. Managers learn to refrain from investing anything like a $100 million system that is hard to change and will take ten years to pay off. Teams work in a modular fashion and continually innovate in the light of experience and customer feedback. And better yet, they find that it can be done at scale. They never arrive at perfect. They keep at it. The improvement journey never ends.