Should we automate existing work or unlock new paradigms?

Original post on LinkedIn

My conclusion: only once robots are deployed to unlock new paradigms of customer value will we see rapid proliferation. Robots to automate the classic “dirty, dull, and dangerous” tasks of today will still be useful, but the value proposition is often thin and the implementation path difficult.

A story I find useful is the pursuit of “The Everything Store” in the early days of the internet. Many innovators realized the web would transform retail, with online-based stores able to deliver superior convenience, selection, and price - the retail trifecta. The key question: how & where to start? Webvan and Amazon took two very different paths.

Webvan decided to start with groceries. Their vision started with convenience: offer seamless online ordering with delivery to your doorstep.

They built a website and an expensive automated warehouse. The service was up and running in 1999, but they quickly learned 2 things: 1) automated warehouses and delivery of perishable goods are very expensive and 2) customer willingness to pay for convenience was low. After burning through hundreds of millions the company shut down in the dot-com bust of 2001.

Despite Webvan’s innovations they didn’t deliver a new paradigm: all of their grocery products were already available locally. Superior convenience was inadequate, and it was expensive. 

Amazon started with books. They went searching for a retail concept that was only possible with the internet - using technology to unlock a new paradigm. There’s an illuminating Jeff Bezos quote from 1997: “... there are more items in the book category than any other category by far… with more than 3 million books in print. With that many items you can literally build a store online that couldn’t exist any other way.”  

While physical bookstores carried only a few thousand books, customers were delighted that Amazon offered all 1.5 million English-language books. They didn’t mind that, initially, prices were average and shipping times were long. In contrast to Webvan competing with local groceries, for many books Amazon offered the only way to buy them! Once Amazon perfected books they expanded to other categories, eventually becoming the Everything Store.

There are parallels in robotics: many companies start with expensive, fully-automated robots that do tasks the same way they are done today. Humanoids and general-purpose robotics control algorithms, when focused on existing tasks, fit this category. The value proposition is thin, and the cost to deliver is high.

Instead I’m excited by robotics companies who find a new paradigm of value in a niche of the home, factory, or field, build a profitable business around it, and pair it with a clear path to their full future vision. iRobot may be the best-known example here, but they are a cautionary tale: they executed in their niche but failed to continue to innovate and expand to their full vision.