Founding hardware companies: what happens when you build something that lasts?
It wasn’t until I jumped off the cliff that I realized there were things I hadn’t considered about starting a hardware company. Contrary to popular belief, the hardest part of hardware comes well after the product is designed, manufactured, and fulfilled. It was only when we put our product in the hands of customers that the real challenge emerged. We stepped into the realm of supporting a product that people loved, that lasted well beyond the event horizon for a startup, and that challenged us with issues that couldn’t be answered with analysis or by following best practices.
We founded Other Machine Co. with technology that was developed on a government grant. The government’s goal was to build desktop-sized manufacturing equipment that could be used to train the next generation of the U.S. manufacturing workforce for less money. When the funding unexpectedly got cut, our goal was to keep our jobs.
If there’s one thing I wish I’d known, it’s that technology built for the government is a product for the government and isn’t necessarily great for anyone else. However, there was something special about what we’d created. We’d built the Othermill, a desktop CNC milling machine, but we’d also radically reduced the barrier to entry for manufacturing high-quality parts. The Othermill was a safe, easy-to-use, and inexpensive version of a CNC milling machine — one of the most common, complicated, and expensive tools in hardware prototyping and production. Any time you lower the barrier to a complex technology you open up new markets, and that’s the part we just couldn’t let go.
CNC machines are typically very difficult to use and definitely not welcoming for casual hobbyists, high school librarians, and girl scouts. Even though this machine was easy to use, we made a classic mistake: believing that this product was for everyone. That belief is shorthand for “we have no idea who this is for or what problem we’re solving.” The original intent was to meet the government’s needs, but if we wanted the company to survive, we didn’t have time to find out which commercial market it was for. So we did the dumbest/smartest/riskiest/best/only thing we could do on short notice and ran a Kickstarter campaign back when it was still pretty novel. We used the campaign to get quick cash and do some minimal validation of a potential market.
Nine months late. 4x over budget.
Just over 200 people paid about $1,500 for our product in our crowdfunding campaign. Even though we raised $311,657, this ended up being about a quarter of what we needed to actually deliver those 200 machines. The machines we eventually shipped were essentially minimally tested, glorified prototypes.
The strangest and most unexpected thing was how reliable they were. About five years later, after I sold and left the company, many of our Kickstarter Othermill owners were still using their machines. But eventually after years of use, the original Kickstarter Othermills started to have problems.
In the years that passed, we moved on from those original machines. We were now on our radically redesigned, fourth generation product. We found our customer (professional electrical engineers) and our problem (rapid PCB prototyping). We even had a new brand and product name (Bantam Tools Desktop PCB Milling Machine). The Kickstarter Othermills contained some custom parts that we could no longer fabricate, several hobby components that were not intended for heavy use, and a few commercial parts that got discontinued. As the machines started to wear out, we sent replacement parts when we could, but the options were dwindling. When our extra parts eventually ran out, we were left with one of the biggest moral decisions that every hardware company faces: what do I owe my customer as the product ages?
The hardest part of hardware: ongoing product support
It always feels like the design stage is an eternity. Getting production going is like an endless marathon of marathons through minefields beyond your control, but the real test of character is actually supporting a product that is in the hands of customers for years on end. This is especially difficult as you improve the product and come out with new models.
For most hardware companies, you’re lucky if customers actually open the box and use your product at some point. In our case, early customers grew to rely on their machines. They’d grown attached to them. As the machines started to wear out customers wanted to repair rather than replace their Kickstarter Othermills. This became a problem, since many of the parts we used in those early machines were basically custom parts.
Parts were prototyped in small batches and essentially bespoke because we were still iterating on the design to improve manufacturability. Because of the way the machine was built, those parts lasted for years, but bearings wear out and motors get dust in them. Eventually, the machines needed parts that we no longer remembered how to make and maybe never documented in the first place.
As much as our early customers loved their machines and wanted to repair them, there was no way we would have the resources to support them long-term when we couldn’t even get or make the parts. Yet we really cared about these people. It was a tough spot.
Spoiler alert: there’s no good way to support old products
There’s no magic way to support old products while creating new ones. But awareness of the entire product lifecycle is important if you’re thinking of getting into hardware product development. We ended up approaching it from the following perspective:
- Without these 200 early adopters, our company never would have existed.
- The value of capital equipment (like a $1,500 laptop) depreciates to zero over five years.
- Customers had grown attached to their machines and wanted to keep them. (And we had unknowingly encouraged that by giving the machines fun serial numbers like Sassy Saxophone and Brave Bassoon.)
- Having outdated equipment in the field is a liability and a time pit for support and service.
- We wanted our company to survive and simply couldn’t afford to replace or repair the Kickstarter Othermills.
- Supporting a unique version of software for 200 machines added a lot of complexity to the development process.
- Without these 200 early, passionate, patient adopters, our company would never have existed (this gets listed twice because it was important to us).
The decision of how to handle the older versions of our product is a moral decision because there is no way to make a 100% rational choice. You have to pick a path based upon what kind of company you want to be.
Ultimately, we decided we wanted to be the type of company that values its customers, but limits the cost to something that we could absorb. We offered all the Kickstarter Othermill owners $1,000 credit towards the purchase of a brand new $3,199 Othermill Pro, but that they’d need to use the credit within the year. It was a thank you, but it also acknowledged that we are a business, and told customers that if they really did need to prototype PCBs on their desktop, they’ll need to spend some cash. We also decided to focus our software support and new features on our newer products.
In hindsight, I wish we’d just used more off-the-shelf components and made our early products self-serviceable and easily upgradeable. That’s what we did for all future models. But it’s not always possible to build in backwards compatibility and a path for replacing discontinued components. But if you care about your customers, you’ll at least give it some thought and be intentional about the path you choose. From my perspective, the hardest thing about hardware is product support and answering these moral questions, not product creation.