Good Robot, Welcome and Why

Good Robot founder Jon Buford with The Minder

As of today, COVID-19 has been something in everyone’s lives for at least seven months. For me, I had plans for bouncing back and forth between the US and Asia via Europe for that time. I ended up in my hometown in Georgia in the US when international travel was disrupted. Like everyone, the worst case I expected was that it might be a couple or three months and everything would be back to normal.

In that time, I had to reassess what was next for me, but also the general consumer electronics industry. The practice of developing new products in one place, making them in another, and selling them globally has its problems, but many of them end up being connected to how things are made, and not necessarily the cost of labor so much. For consumer products, the major parts tend to be injection molded plastic and electronic circuit boards.

Within the last few years, inexpensive and reliable 3D printing using UV cured resin has made great progress both in speed of output as well as reducing the cost of the equipment. It doesn’t completely replace thermoplastics, but for basic casings and things like that, it really is more than good enough.

I started developing The Minder assuming that I would probably have the casing injection molded and the circuit boards done in China with vendors I’ve done projects with before. It would have worked, and I could still do the final assembly and testing here in the US, but that had a few real limitations. First, the cost of doing molded parts has a big hurdle for the first few thousand parts. Even inexpensive tooling is in the thousands of dollars, so for small runs that means that your parts cost a lot until you have made enough to spread that cost out more. This pushes production towards huge runs and that in turn means that products that might be amazing and needed for a few thousand people probably would never get made.

Second, with this process, once you have made a mold, you tend to not move beyond that basic design because of that initial cost of making it. This means that products tend to be either well-funded and destined for mass production or they are niche with casings that have some real compromise because you can’t afford to create more complicated and expensive tooling or because it is done already and you can’t go back and make major changes that would improve the product without starting over on that tooling.

Long story short, I picked up an inexpensive resin printer to use it for final pre-tooling checking of my 3D designs, and I realized that it really is that time where it makes sense to use that for the production of real products, not just for prototyping. This isn’t a new idea, it has been promoted for years, and it made sense for very specific applications. What has changed is that the technology is basically now at a commodity level, so offerings from different vendors have very little differences and even feature the exact same main components and core operating system. The cost per printed unit for the printers themselves is pretty low and the consumables are all open market, so those are both predictable and reasonable.

I could make a similar case for the PCBA, but I’ll put it more concisely, the further down the supply chain we turn raw materials into finished goods, the less time there is spent with finished goods getting shipped places and it reduces the cost of inventory and risk of not selling that inventory into the market. If I have a stock of a bunch of components I purchased, it is possible to create something different than they were originally bought for or to even resell them to someone else for their use. If I have a bunch of something I purchased, a finished product, I don’t have that flexibility and it is only the demand or lack thereof for that specific product that determines if I can sell that for a profit or a loss.

That is the core of what is broken with traditional retail supply chains. Purchase orders are made for products 6 months in advance of when they are possibly going to be sold. If there is not as much demand for that inventory as desired and planned for, usually the company in the middle ends up eating the difference between the desired price of selling them and the final clearance price. This is a very inefficient system that can neither easily ramp up or down with any speed and it promotes making millions of the same thing.

We are on the cusp of being able to produce things on demand with a short enough time to where it would be similar to what we are getting used to for an online order that takes a little while to be delivered. There are limitations on cost, materials, and overall fit and finish of things made this way, but expect that gap to be reduced quickly as we learn from our first products and figure out what questions we need to ask in order to improve them.

The future of globalization is aggressive localization. Making things that are for you not everyone. Welcome to the future.