In this short series of blog posts, we’ll explain how we successfully bootstrapped and launched a smart home gadget that we manufactured in China.
It’s not typical to hear the terms ‘bootstrapped’ and ‘hardware startup’ in the same sentence. The vast majority of new hardware companies either raise large funding rounds from investors or use a crowdfunding platform like Kickstarter to pre-sell their product.
We decided to go our own route. The goal was to prototype and then build our initial production run on a modest budget without any outside investment or the high expectations of a crowdfunding campaign. This frees us up to run the company on our own schedule and terms.
We’re the father-and-son team behind QTIME, a popular smart home gadget for games consoles. It lets parents manage their kid’s gaming time in their home.
Aside from a brief visit to China, the whole process took place on my (Keith’s) farm in rural Wiltshire, England.
James was addicted to videos games as a teenager, eschewing homework and exercise for his obsession. He would gladly spend every waking hour gaming if he was left to his own devices.
I would struggle to get him off the console, even resorting to pulling the house power breaker in desperation.
Many years later, with a sudden glut of time on my hands having sold my TV consultancy company, I started to wonder if this remained a problem for parents. To my surprise there was not a single suitable product on the market.
As I discussed this with James and as we reminisced about those fond memories of teenage battles, we eventually decided to build a hardware solution that would work on any games console. The timing was fortunate; gaming addiction has become something of an epidemic with the WHO now officially recognising it as a disorder.
The first prototype
I should note that at this early stage I hadn’t developed a hardware design for almost 25 years. But I found that apart from integration and miniaturization, the foremost development was the quality of free embedded development tools that are now available.
The very first prototype was based on an off-the-shelf radio-controlled relay board bought on AliExpress ($50). It used an 868MHz remote that could be used to replace the power lead of the games console entirely.
It was good enough to prove the concept.
We need a guinea pig
My 13-year-old stepson Isaac, who had an determined interest in playing Fortnite, was our unfortunate and long-suffering test case for all of my first designs.
The first issue was getting the timing right. Our gadget needed to detect exactly when the the games console is switched on. To do that I developed (through many iterations) a current detector system using a toroid design.
It took a while to get it right and in the beginning it would often power off the console unannounced. Sorry, Isaac.
At this stage we decided that the gadget needed a smartphone app and some wireless connectivity to make it easy for parents to set up.
Enter the ESP8266
Luckily for us we stumbled across a fantastic low-cost chip called the ESP8266. It’s popular with hardware hackers, comes with WiFi onboard and it fit our needs perfectly.
I pieced together the second prototype using basic stripboard and the Arduino toolchain ($600). I wrote some eccentric hardware-engineer C code and shipped it to James (who was at that time living in Indonesia) to write some better firmware. Amazingly the Heath-Robinson construction survived thousands of miles by courier.
Despite the severe resource constraints on such a basic chip, it worked very well for our use case. It was just about powerful enough to run a basic webserver to facilitate pairing with an app.
At this point word of our project had spread through our friends and family. It appeared that almost everyone we knew with teenage boys was struggling with getting them off their gaming consoles and this gave us confidence in going to production.
Running a pre-production trial
With sufficient confidence in the prototype, we decided to build 15 pre-production units. After doing our own schematic capture we had the PCB laid out and the prototypes built using a local surface mount manufacturer for around $2500. The costs came to around $3500.
We wanted to use a local company for close communications and so that we could sort out any manufacturing issues quickly.
Using basic off-the-shelf plastic casing, we shipped out our prototypes to some local families for a two-week trial. James wrote some monitoring software so that we could see detailed analytics on how the units were being used. At the end of the trial period, we interviewed the parents in depth to understand how they used the product, its utility and any improvements we could make.
It was encouraging that most of them didn’t want to give their prototype back.
The trial taught us two things. First that pairing the gadget to home WiFi needs to be as frictionless as possible. Smart home devices are notoriously tricky to set up. We researched a bunch of other products and found that only Google Chromecast does this well.
The second issue turned out to almost be a showstopper.
A fatal flaw
You’ll remember that at this stage the product worked by cutting the power source to the console when the child’s time is up.
The problem is that if you suddenly cut the power to a games console there is a non-zero chance that you will corrupt the hard disk.
We instituted a 10-minute audio and visual warning but found that most kids were so engrossed in their game that they would continue playing up until the very last second.
At any scale this could turn into a disaster.
We had to go back to the drawing board. It was a difficult decision to jettison this design as we had sunk hundreds of hours of development time and thousands of pounds into it. It would also mean another three months of delay.
The big redesign
In the next part of this blog series we’ll explain how we quickly iterated to a suitable design and manufactured the first production run in Shenzhen, China.