13 years ago, Tesla started to produce Roadster, marking the beginning of the next generation for cars. On Aug. 5, President Biden announced a new ambition of 40–50% of all vehicles sold by 2030 is electric. The U.S. aims to catch up with Europe & China on the EV transition & maybe even lead again. This marks the beginning of the mid-game. This hockey stick growth will be driven by the early adaption & early majority, accounting 47% of TAM. This decade will be the golden age of EV. Even better in this round, innovations are blossoming everywhere. It’ll go far beyond replacing combustion engines with electric motors. A prime time to redefine the car as we know it.
The “ghost” in the new breed of cars will be a key differentiator in this new age of software-defined. And the game will be very different from what established carmakers used to. For example, Tesla is “playing to keep playing”: Full Self-Driving subscription model. So the question is how carmakers can transform to play better and avoid the Nokia traps early rather than later. For example, Biden’s E.V. Plan may squeeze Toyota hard as NYT suggests. As far as I observe, it’s not just the HW, but also SW. Which is much harder to catch up. Psst Toyota, it’s still not too late to pivot.
To lean better
30 years ago, The Machine That Changed the World popularizes the “lean production” concept from Toyota. The great work is based on a $5m + 5 year MIT study for the future of the automobile. Since that, the lean concept continues to inspire many transformations beyond manufacturing, e.g. Lean Software Development: An Agile Toolkit. Today, Agile, Scrum, Kanban, etc. are commonly applied in SW development. And, it’s not just in Silicon Valley, but also most devices & even car makers I work with. To go for the next level, even the question of how to be leaner is the same, the answers are different if to ride the next big wave. Let’s scratch the surface of 2 paradigm shifts.
Pull in user value
Back in my Moto/TTPCom feature phone days, we spent more than 50% of the time on the feature parity of predecessors & carrier requirements. Sure, those “customer requirements” were important & blocking releases. But, they did not bring too much value to the end-users. So no matter how much time & cost we put into those features, they did not only fail to put a smile on a user’s face. Often, they made the products worse by adding too much complexity 🤯. Today, I observe most carmakers are still in a similar hole, and digging harder is a wrong answer.
The best way to be lean is to eliminate waste, and cutting features typically trumps any optimization strategy. Sure, I know it’s scary because no one gets fired by adding features, and someone most certainly will be by cutting the wrong one. It takes courage to eliminate the fat, aka features with low user preceded value. But that’s the best if not only way to keep a product lean or prolong its life. Remember, even SW marginal cost is low, it’s not net-zero either. For example, Windows Vista was not born as an 800-pound gorilla. It grows into one by gradually gaining the weight toward a path to collapse on its own weight.
To mitigate that destiny, you’ve to be mindful to “pull” more user value than adding what you told. For example, the Lean-Startup cycle is a popular process in Silicon Valley. Because Build-Measure-Learn helps teams to be more certain on what users really want in a highly uncertain environment.
Deploy Frequently
The adaptive evolutionary theory inspires us a way to predict which product may survive in a rapidly changing environment. tl;dr a higher deployment frequency means higher adaptability to fit better. For example, when a carmaker still takes 2+ years to ship a new product & features, the product development will always be a “mass production” style instead of a “lean production” style.
Just-In-Time is the lean way to reduce the end-to-end response time from suppliers to customers. When it’s properly executed, it significantly reduces inventory for production & distribution systems. The same spirit can be translated into a high deployment frequency in SW development to enable:
- Deploy as early as possible. So you can iterate faster on the Build-Measure-Learn cycle. You can ship minimal viable products early, enhance them & add new ones as you go. Furthermore, if you are not in the “one person workshop”, you also reduce the shadow of the future problem complicating the collaborations.
- Keep the options open as long as possible. Your product decision complexity is dramatically reduced simply because you only need to predict the next quarter rather than the next year. More importantly, the option to change at the “last minute” is priceless.
- Recover fast. Typically, carmakers take 6–9 months to test & fix issues in their SW. Which is great. At least for the good old day, the SW was much less complicated, features were static and pretty isolated from the rest of the digital world. But, is that the “feature phone” world you think the next decade will be? In a fast-changing world, trying to achieve a “zero” bug produce may be a myth. So the game may be striving to discover issues early & recover quickly.
Full Disclosure
The opinions stated here are my own, not those of my company. They are mostly extrapolations from public information. I don’t have insider knowledge of those companies, nor a whatever expert.
