Challenges With Small Scale Manufacturing Via Kickstarter, Reflecting on the Public Radio

Over the next three weeks Spencer and I will be building up and shipping out the remainder of our Kickstarter rewards and pre-sale orders — what amounts to roughly 2,100 individually-tuned radios. Our two-person team began down this path about two years ago.  What started as a side project increasingly becoming front-and-center in both of our lives.   

People frequently ask if we had any idea how much work the project would actually take, and the short answer is definitely “no.”  I can think of a few aspects of the project in particular that at one point or another buried us in work, crippled our output, or slowed our fulfillment abilities. Supply chain management is one, minimum order quantities is another, and working with individualized rewards using the current Kickstarter fulfillment tools was a third.  And, in one way or another, these all relate to the particular challenges of undertaking small-scale manufacturing via Kickstarter.  

The early days on Grand St Beta.

Leading up to the Kickstarter campaign, we worked hard to develop the beta version of our radio, and we generated a decent amount of feedback from our first customers on Grand Street.  We knew that there was a fair amount of re-engineering left to do, but we decided to Kickstart between these steps anyway.  Kickstarter was to serve primarily as market validation, and we knew that we were taking a gamble; if we funded, or overfunded for that matter, we’d have to reengineer and redesign our radio for manufacturability and quality (our original radio reception was pretty crap).  As a two-person team with relatively limited experience in embedded systems and electronics design, our learning curve would be steep to say the least. 

It wasn’t a surprise to us when we funded, and it wasn’t even such a surprise when we overfunded either, but the degree to which we all of a sudden had to literally become a business, form an LLC, and handle customer service inquiries, while figuring out how to begin refactoring our design, did come as something of a shock. Ultimately, we were not unique in this — Kickstarter has a way of doing this to projects — but unfortunately, as an electronics and hardware undertaking, we sort of landed in no-man’s-land.  

We’d raised enough money to build up our radios, fund a handful of revisions to our designs, and buy ourselves pizza on nights we were working on the project, but not enough to have any real purchasing power with our contract manufacturers.  We couldn’t take advantage of the economies of scale we would have benefited from had we 10X overfunded, as the costs per unit (with electronics components especially) fall off precipitously with larger orders. 

Fulfillment party #1 !  

At the same time, 2,100 radios was about 35 times more than what we’d previously put together by hand, and there was no way we could undertake the entire assembly ourselves.  Which is to say that, in order for any of this to work, we’d need to reduce the cost of our own labor to zero and call in big favors from friends.  We started doing it for the glory, for the learning, and maybe bragging rights. Whatever it was, we had started something we had to now follow through on.  

Luckily for us, Spencer had experience from his previous job working with overseas manufacturers, and had a pretty good idea of the hoops we’d have to jump through.  Yet in the end, we still had major issues with certain suppliers.  Our speakers, for example, were about two months late delivering, over budget, and required hundreds of email exchanges, and, in the end, the quality was just okay.  And that was starting with a product that the manufacturer already had tooling for! When things were going too slowly overseas, we found that we didn’t have any clout or leverage with these factories — we were spending less than $10,000 with each of them — and we couldn’t honestly forecast what future orders would look like in hopes of persuading them to prioritize our order.   

Throwing more money at the situation might have helped, but we were really experiencing another problem: we didn’t have any established supply chain that we’d fully vetted.  We just didn’t have the time or the resources to do this.  Our process was this: find suppliers who can make the thing affordably, on time, and will respond to our emails relatively quickly - bonus if they had reviews or other noteworthy customers, or were willing to provide suggestions along the way.   

We went overseas for a handful of components that seemed straightforward enough to take this kind of gamble on. But really, we didn’t have a choice, as our cost of goods would have been through the roof had we not done so.  Having parts custom made, like our knob, ended up saving us 80% on that line item alone.  For our most important contract manufacturer, the printed circuit board assembly house, we chose to stay local, which was a great decision — there was a ton of back and forth with the factory, and it was crucial to the projects success.   

There have been a few other annoying logistics to suss out along the way.  Like, where do you store a couple thousand mason jars or pallets worth of shipping boxes from Uline?  How do you receive these items if you don’t have a loading dock?  Because we didn’t quite have the funds to outsource all of the final assembly, this stuff was up to us to figure out.  Our apartment closets are stuffed. 

Early motivations brainstorming.. 

In the end, we started this project to accomplish a few things: add to our portfolios, learn new skills, make something our friends would enjoy or even buy, and have the experience taking on a completable project. We have learned a tremendous amount along the way: everything from which packaging tape works best to why class D amplifiers don’t play well with RF projects. And in that regard alone, our project has been a huge success.  

It’s interesting to note that as more and more ‘maker’ tools are developed, and the barrier of entry to developing hardware projects gets lower, the level of risk project creators take on is not necessarily following the same trend.  This kind of hacking>honing>kickstarting workflow can lead to trouble. In our case, if we hadn’t been working on The Public Radio pretty much constantly over the past 8 months, we’d be much more than two weeks late in delivering and much less than just breaking even with funds.  

Aside from planning on nothing but sheer dedication to your project full time, my bit of advice on Kickstarting a small scale hardware project is to be careful how hard you push to overfund. Chances are you might get more than you bargained for, and will have plenty of late nights ahead.  If you’re going to swing hard, swing really really hard — and know where that next manufacturing plateau is.