I often joke that if I got $1 every time someone told me that "Hardware is hard", I would have become a unicorn myself.
As an ex-engineer who spent the first decade of his career in semiconductor R&D – arguably the hardest in all hardwares – I have to say that this sentence is a lot more about misunderstanding than real wisdom, despite the fact that it came from one of the smartest person in our era, Marc Andreesen.
My conclusion first: yes, hardware is hard, but it's not much harder than trying to come up with another taxi hailing service to compete against Uber or Didi, which respectively raised multiple billions worth of cash for cut-throat competition.
In the modern startup world with abundance of information and resources, nothing is particular harder or easier than another. Yes, it's much easier to get a mobile app off the ground, but since it's so easy it also means that there would be tons of copycats as soon as a certain initial product-market fit is achieved. The competition would quickly shift to the hard part – raising as much money as fast as one could to stay ahead of the competition.
On the other hand, if a startup is doing something that's fundamentally hard – or more accurately, cumbersome – to do, it also means the competition would be less fierce due to the entry barrier.
In fact, "hardware is hard" is a mere statement of the perceived difficulties in operations of a hardware startup. It does not represent the general difficulty for it to achieve success compared to its lean software/mobile brothers.
Most of the time when people think about the "hard" parts of a hardware startup, they are talking about "manufacturing" and "retail".
For a lean startup, the general assumption is that everything could be done in-house: from coding to growth-hacking, from launching to operating, from marketing to sales, all in the coziness of one's own office. The image of a bunch of hippie coders sitting behind 27" iMacs with out-of-size headphones on their ears trying to disrupt the world is so strong that just the thought of handling physical semiconductor components, multiple supply chain vendors from China or Taiwan, as well as the various sales channels in each and every single market to break into scare the hell out of lean evangelists.
However, are manufacturing and retail really that hard? I mean, if that notion is right, wouldn't it be true that there should be really few players dominating these two fields by massively outperforming the others?
That obviously is not true.
In fact, manufacturing and retail are two very old games, so old that most lean entrepreneurs don't even bother to work on them simply because they seem old and heavy. However, the fact that they're very old games also mean that they behave more like observable "risks" than unobservable "uncertainties".
"Risks" and "uncertainties" are two very different concepts in financial languages.
Risks are variability that are observable, foreseeable and sometimes even quantifiable – although don't stretch this last claim too far lest we forget the bloody consequence of the 2008 Financial Crisis. In other words, risks could be managed, mitigated or even controlled with the various tools, either financial or operational. The spread will always be there, but the ability to change the spread could differentiate one venture capitalist from the other when looking at the same startup. A typical risk of a startup would be recruiting enough talent to grow to its potential. This is not 100% sure until talent are on-board but they're visible given the talent pool of the area or even the country, assuming enough capital funding for such national recruiting process.
Uncertainties, however, is not observable, foreseeable, let alone quantifiable. For example, one could not control whether the girlfriend of a co-founder/CEO would dump him just before the day of an important product launch and that the later would lose all strength to work for 2 weeks. One also could not foresee the entrepreneurs being hit and killed by a DUI driver when riding his bicycle home in the mid-night after a long day of coding. These are pure uncertainty and beyond the reach of even the wisest investors.
By nature venture startups are both about risks and uncertainties. Risks are more or less visible to all experienced VCs in their respective realms. As long as VCs could do a fair assessment on all the visible risks, it won't stop them making investments. On the other hand, uncertainties are what keep VCs awake in the middle of the night – things that are out of control but that are crucial to the success of the startups.
In the world of hardware startups, manufacturing and retails are actually more like risks than uncertainties.
In terms of "manufacturing", an experienced VC would be able to quickly spot all the weaknesses in a hardware startup in doing manufacturing: Is there already DFM or DFT in prototyping phase? If not is it salvageable? Are key components single-source? Are certain components srouced from Taobao.com during prototyping phase actually EOL ones? Is there any part of the product that requires special reliability test? Or standard electronics tests suffice? etc.
These are all very mature questions in manufacturing and after evaluating all the risks, a real hardware VC would be able to assess if he or she could help mitigate them by either connecting the startups to suitable EMS firms or helping them hire qualified engineers in-house. As a result, even though "manufacturing" is hard, it's where the right VCs could add values to differentiate themselves (and therefore their investments) from the others.
The same thing goes with retails. In all the developed countries, retails is a slowly-evolving game that has been going on for more than a century. As dominant and successful as Amazon is, it took them (and all other e-tailers) a good two-decade to reach a mere 7.4% market share of all retails:
This means that handling the channel mix in a large and mature market like US or Japan is nothing magical. There should be enough talents in the pool that could do the job well enough. A VC with retail means would therefore be able to assess quickly the suitability of the product by a certain hardware startup to go on any particular channel and whether he or she has the means to help speed up that process by doing intros or referrals.
In other words, with the right VCs that dedicate themselves to hardware startups like us, the two parts that are most often perceived as "hard" in hardware startups are mere mitigable risks, far from enough to stop us from doing an investment and definitely not enough to keep us up at night.
So what is the really "hard" part about hardware startups?
Above all, product-market fit.
Or more specifically, the ability to sell hardware products to consumers at a price that's not determined by competitions, therefore be able to reinforce the brand values more and more and grow long term revenue and profit streams.
Note that this is not as obvious as what most people believe. The real ground-breaking hardware products are usually beyond consumers' imagination before they appear, as Steve Jobs famously said "Customers don't know what they want."
For example, there was no real smartphone demand before iPhone was invented. There was no sector called Action Camera before GoPro rose to dominance despite the dozens of copycats. The term "wearable" was created after the reception by US consumers of products by Fitbit or Jawbone became evident.
On the other hand, even the largest electronics brands with the most marketing resource could get a product wrong and end up selling zip. My friends who have bought the $1 WebOS tablets by HP when the later was forced to dump all inventories at such a nominal price would know that.
As a result, a startup could be expert in all aspects such as design, electronics, softwares, channels, etc but if the consumers choose not to buy their products, then all bets are off.
This is what keeps us up at night. This is the really hard thing about hardwares.