The Unique Difficulties of Kickstarting Hardware
“Truth is, hardware is extremely hard.”
OnePlus cofounder member Carl Pei said in an apology post on their forum.
In Carl’s case, he was apologizing for slipping the launch date of their new phone, the Two. Eagerly anticipated by gadget enthusiasts, its predecessor the One’s main disadvantage was how hard it was for a potential customer to get their hands on one.
After promising to manufacture “20-50x” the quantity of the new OnePlus Two, they were going to miss the initial shipping target.
Morale among the fan community was low.
OnePlus is not a team of brand new hardware engineers. They are made up of people who have been making consumer electronics (mostly at Chinese company Oppo) since the end of 2013 and still have ties to the resources of their former employer.
But in this glorious age of Kickstarter, the trend of underestimating hardware is one we see time and time again amongst experienced and inexperienced teams alike.
Hardware is hard.
According to Ethan Mollick, professor at The Wharton School (via the New York Times) “more than 80 percent of hardware projects ship with significant delays… [and] 14 percent of the projects studied have, since 2012, shipped either nothing at all or something too shoddy to use.” And besides the most obvious hardware development issues like high complexity or time consuming iterations, successful hardware crowdfunding campaigns create their own perils.
Team ZPM is a group of people who dreamt of an inexpensive espresso machine that could match what people get from commercial restaurant models. They had invented and built working prototypes of their device and were ready to sell more.
The initial plan was to raise $20,000 and buy a small milling machine to hand build 50 units for $200 a piece. However their Kickstarter accidentally hit the jackpot and they quickly raised well past their initial target, ultimately to the tune of $370,000 in pledges with 1500 machines sold.
Sensing that they might have trouble meeting such strong demand, they sold additional units and raised angel funding, finally gathering $1.2 million.
ZPM quickly suffered a series of setbacks due to crowdfunding. For instance, vendors using happy backer comments to justify schedule slips. More critically there was also a misalignment of goals.
ZPM had originally planned to hand-build all units, but they were now committed to producing thousands of them. Building them by hand was no longer an option; the task would take eons. But making their initial designs manufacturable consumed more resources than expected.
For example, ZPM paid $100,000 for DFM. Ultimately, ZPM actually had the product they wanted to sell, but the Kickstarter process had committed them to volume manufacturing.
Mistakes made out of inexperience ultimately chewed through their finances and drove the founders to sell what was left of the company.
I heartily recommend reading this New York Times article for more detail.
Another example is Central Standard Timing’s CST-01, a 0.8mm thick flexible E Ink watch on a spring steel band that they called “the world’s thinnest”.
The founders are ex-IDEO designers whose first published post about the watch were on the IDEO blog describing their design process and displaying prototypes.
Their Kickstarter page had all the credentials of a serious hardware effort: there were photos of real working prototypes, people handling mechanical samples and discussion of effective assembly partnerships and fulfillment. Like the OnePlus team, these were not people new to the process of hardware development.
But, hardware is hard.
On February 4th, 2015 they posted an update indicating that units were finally shipping after a 1.5 year delay, and included photos of operators performing final testing and packages ready to ship.
At the very bottom there was a single sentence about “making sure all the kinks are worked out.”
Two months later, they posted that yield was lower than expected. Much lower.
The post opened by reporting they had produced between 6% and 10% of their targets in the first two weeks, and that, at their current failure rates, they would have a 66% yield.
It turned out that manufacturing extremely thin flexible electronics required new innovations in packaging and assembly techniques which, impressively, CST had accomplished.
Their final updates include photos of bespoke glue pouring machines and other custom designed manufacturing equipment they were hoping to sell to recoup costs.
Ultimately the crowdfunding process had forced them to commit to a selling a certain number of units at a certain cost before they had established the true price of manufacturing.
This discrepancy consumed their cash reserves, and on June 14 they announced they were parting out the company and backers would see no return on their contribution.
Hardware is hard.
Doubly so mass produced hardware. But even more difficult is weighing the desire to produce the perfect product for your customers with the realities of engineering and manufacturing.
Producing the perfect product is impossible (cyan Nokia N9 notwithstanding) and even getting close on a first try is hard. It’s difficult to predict the types of problems that may await along the road to production without years of experience.
At Mindtribe we recommend working with an excellent engineering partner both before committing to crowdfunding (to help figure out the true cost) and after funding (to help with prototyping and the ramp to manufacturing). They can provide the support necessary to do it right.