I don't know why people think competing against Apple is hard or impossible. Apple steers a pretty clear course, and they don't like to take risks on niche products. They're not like Google where, on a whim, they might make an IoT blender for reasons that nobody can explain.
Apple also maintains very high margins, something that leaves more than enough room for another company to compete on price. The trap here is instead of producing a good product at a better price, most companies produce the cheapest possible product in a race to the bottom.
Find something Apple isn't doing, doesn't seem interested in doing, and use their business model. It worked for Nest. It worked for Tesla. It could work for you.
>I don't know why people think competing against Apple is hard or impossible.
You just made the case against this. The fact that they won't make an IoT blender is why they're hard to compete with. Apple is the polar opposite of Google in so many ways but a big one is their organizational mentality of "more wood behind fewer arrows". Apple is the batter in a no-called-strike baseball game (to borrow a Buffett quote) which means they're only going to swing when they get a fastball right over the plate.
Apple has high margins but it's actually because they have the lowest costs. In Porter's terms, they're the benefit leader but they're also the cost leader.
>Find something Apple isn't doing, doesn't seem interested in doing, and use their business model.
This isn't actually competing with them then. This advice is probably valid, but you're basically saying to just be the Apple of some other industry (like Keurig or Nespresso).
I'm saying Apple has amply demonstrated a business model others can apply, and it's not like Apple singularly invented this model. They just borrowed from other companies like Nikon, Canon, BMW and others, adapting and perfecting it for the business they operate in.
There's a lot of other industries that need an Apple and don't have one. Who is the Apple of toilets? Of mattresses? Of 3D printers? Of blenders?
Some have leaders that succeed not based on having a superior product, but by their crushing marketing pressure and omnipresent retail representation. They're ripe for disruption.
> Apple also maintains very high margins, something that leaves more than enough room for another company to compete on price. The trap here is instead of producing a good product at a better price, most companies produce the cheapest possible product in a race to the bottom.
This is what companies like Anker and Aukey are doing too. They compete directly against Apple's accessories with a well polished, durable product, for cheaper than what Apple charges.
“We fucking started a hardware business! There’s nothing else to talk about. We shouldn’t have done that.”
What I hear in this quote is "we went into a technological field we know nothing about with the intention of making money. We shouldn't have done that".
Nothing wrong with going into hardware if you have experience end-to-end bringing a hardware product to market. It's just that less people have that than in the software world, and us software people get used to being able to figure out everything JIT.
Probably don't go into chemistry either, or algotrading, or practicing law, thinking "eh, I'll figure it out as I go".
This is what the mgmt of Twitter/Youtube/FB/Uber etc are internally crying about helplessly. This "we'll figure it out as we go along" mentality, is deeply rooted in leadership culture at the moment whether its politics, wall street, silicon valley etc.
It did some useful things, but it needs to evolve.
Also, the current trendy management practices in software like: "Move fast and break things," "Release crap early and iterate," and "QA is a useless cost" don't transfer over to manufacturing hardware.
Nope; a lot more rules and regulations you have to respect, certifications for e.g. electrical safety, etc.
Mind you there's a few similar things in software, things like security requirements when handling payments, EU privacy laws, cookie banners, that kinda thing. But they're a lot easier to either outsource (use a payment processor service) or implement (cookie banners range from a notice that says "by using this site you agree to cookies", which probably isn't enough for the law but reasonably user-friendly, to full off-site cookie walls.)
I wonder if there's manufacturers that are offering this kind of techniques to other parties nowadays. I did hear that Apple bought all of the output of a German CNC machine maker for the iphone and/or macbook bodies for a long time just to be able to ramp up production.
This article doesn't mention it, but if you follow all their guidelines, not only did you lower your expectations if fit and finish, but you also lower the expectations of the whole product for people who buy it.
That's actually a good thing. First revision hardware is going to have issues and building a beautiful product and putting in a beautiful box will set you up for disappointed customers.
Bolt's blog is a treasure trove. I link to another of Ben's pieces in the story.
Either do what MakerBot did, and embrace the DIY enthusiasts as your first market, or do what Misfit did, downplay the tech features in favor of a more fashion forward aesthetic. But running headlong into the buzzsaw that is Apple seems like a recipe for failure.
I think the competition is on par with Apple, maybe it should be titled "Copying Apple is hard". Ultimately, have been a lot of innovation by the competition long before Apple did it, like QI charging, large screen, multiple usability improvements etc.
The point is that starting a new company to compete with Apple is effectively suicide, not that an existing business with billions in capitalization like Samsung et al can't do it.
As a HW designer I’d like to contest the title; hardware is hard. Mistakes always happen on any project, but with hardware they’re very expensive to find and fix. Test equipment like oscilloscopes and bus analyzers are in the $100k range, furthermore building your hardware will cost $100k++. Hardware is hard. Software tools are generally free and fixes only cost engineer’s salaries.
Not that this keeps me from thinking HW ideas, and SW is great too.
Great points. Hardware is definitely hard. There are more variables, abstractions, and processes to worry about.
In software, you can write a function that doubles big numbers perfectly until it hits limits of CPU and memory in about one line of code. If you try implementing the same function in an analog circuit, you'll encounter compounding noise depending on heat, ambient EMF, adjacent components, etc.
Upon identification of a critical bug in your SaaS product, one `git push HEAD` is all it takes to get the fix to your millions of users. Hardware is not anywhere near as forgiving--your batch is usually lost.
These characteristics require a different approach for developing hardware-leaning products. Everything needs to be far higher quality before release, because small hardware mistakes can have several order of magnitudes higher of costs than software bugs. That's why it's harder: you have to land a rocket on a planet with fewer chances to correct your course.
Hey there, I'm the writer and I'm definitely guilty of exaggeration on the title. My goal was to push back against this popular narrative that it's something intrinsic about hardware that kills these startups, not strategic errors or poor execution. Hardware is less forgiving, but I'm not sure it's meaningfully harder to build a new gadget than take on Facebook or Google in social or search.
My first job was at an HW startup and I launched a dozen hardware products. It's tough. A screw-up at a single soldering station in Shenzen can screw-up a deal years in the making with Walmart. Overordering materials means you have to lay off friends.
But saying "Hardware is Hard" is an oversimplification.
When we say that to explain the failure of HW startups, we miss a lot of important nuances. Hardware is hard, but that's not the primary reason most HW startups fail. Almost all of the high-profile failures we see in hardware are companies that are trying to be the "Apple of X."
In this post, there are two hardware startups, valued at nine-figures, that are essentially bootstrapped. Hardware was hard for them too, but they managed to win with very little in the way of resources.
A quick note to the author: The article names the company as "Dopplr" but it should be "Doppler". The former seems to be a defunct social networking service.
A bit of hyperbole in the title is ok; you make some great points in your article. I’d like to see a welcoming environment for HW startups so I appreciate your work to clear up the misconception that HW is a recipe for failure.
Also, the idea of taking the B2B route and avoiding consumers is great; consumers have high expectations and are a very well served market.
I'd also add that debugging is much harder in hardware. You don't get an error message or have the opportunity to run lots of tests. So you better have a strong handle on the fundamentals.
Anything that has any interaction with physical world is hard in comparison with SaaS-only business, hardware is not special in this. And in my opinion hardware is significantly easier than various shared economy or service startups, because in hardware your problems are still mostly about engineering and can be analyzed, measured and affected in meaningful ways.
> Dopplr’s product was ambitious, gorgeous, and by all accounts, it worked well, but it didn’t provide the magical experience the Airpods did. Almost no startup could! That “magic” is the cumulative product of decades of R&D and manufacturing expertise.
Nope. The "magic" of Airpods (I own a pair) is quick synced pairing. Startups like Dopplr can't provide that "magic" for iOS/Mac users because Apple doesn't enable the requisite access on their software. Dopplr, Bragi etc are better than Airpods but Apple controls many of the devices they need to interface with so they are fucked.
The lesson is don't design your hardware to match Apple and primarily at Apple users. They will cut the legs under you
Is it so hard to sync up bluetooth headphones? I've done it once per headset I bought on my phone and it was a simple process. Children do it everyday. I suspect Doppler Labs failed to thrive for other reasons.
I just stopped by their site and saw their farewell message. I could 50+ people in the photo. That's 50 salaries this startup had to absorb every month. If we assume 100k real cost per head (recruiting, pay, benefits, etc) then we're at $5m per month on salary alone. If it took 2 year from idea to product, then you're in the hole $120m. If you sell these headphones with a markup and post-everything margin of $30 per set, which seems reasonable, then you need 4m in sales to break even. That's a serious amount of sales to push through.
The numbers look more reasonable now, but ultimately its a hard sell especially when Apple doesn't actually need to make a profit for its accessories, but you do. Even then, what's the market look like for this without an Apple. I suspect not very large and its Apple's own marketing efforts that have gotten people interested in these tiny wireless earbuds anyway.
>Apple doesn't actually need to make a profit for its accessories, but you do.
Do you have data to support that claim?
Amazon might sell hardware at cost or a loss, but Apple most certainly only sells hardware when two conditions are met:
a) they can make a better device
and
b) they can make money
There's a substantive difference between "simple" and "just happens". It's why my BT headphones sit in a drawer, while I gladly anticipate paying 16x as much for AirPods (and no, the money isn't freely available). The error mode for "simple" is "hard", while the error mode for "just happens" is "simple".
Syncing is usually relatively painless but reliably and "instantly" reconnecting is a long ways off. I have maybe a 50% success rate without going to Settings>Bluetooth and clicking on the device. That inexplicably fails about a quarter of the time and I have to powercycle the speaker and toggle bluetooth off/on a few times until it decides to work again.
That rate drops to basically 0 if the speaker is paired to multiple phones, or there are multiple speakers in range of my phone.
If I had to deal with that multiple times per day like with headphones I'd just give up and return it.
I have one absolutely reliable connection (turntable>bluetooth transmitter>stereo) but even that one takes about 15 seconds to establish after turning everything on.
I have airpods and although the _initial_ sync is magical, it's just like every other set of bluetooth headphones when you've got more than 1 device to choose from. If I have my macbook and phone side by side and put in my airpods, they always connect to the phone. I have to go to the mac's bluetooth menu to choose airpods if I want to listen from my mac (icloud sync doesn't seem to work).
Many other Bluetooth headphones have a variation of a feature called 'multipoint' which allows you to be connected to both devices at once so you don't have to choose and switch choices each time
You can choose the AirPods from the volume icon on the menu bar, or are you saying the AirPods don’t show up there? I switch between iPhone and Mac audio a few times a day.
I see. I think I would like it if they could play all audio from all devices at once. Similarly, it would be nice to control phone audio with the Mac volume and pause keys while using AirPods. I didn’t expect any of that, though.
They don't show up there. I've deleted them and re-paired, but they never have. I also have Beats with the W1 chip, and they don't show up there either.
It's a little of both. For sure being designed into the OS layer is a help, but aside from Doppler, plenty of startups have failed because they're trying to ape Apple's fit and finish, rather than finding a way to build a sustainable business.
Android user with Airpods. Magical syncing doesn't do anything on Android. It's the battery life, ear comfort, and the fact that the case is designed so well. There isn't a single set of wireless headphones that does the same thing as the Airpods.
Definitely the case quality and the fact that it serves as a supplemental charger, even when not connected to a power supply. Wireless charging along with lightning charging as well as super audio quality that is indistinguishable from wired headphones of a similar form factor and price.. that's why AirPods win.
The pairing part is just a one-time thing and yes, the experience is nice, but pairing ease isn't the primary selling point. My portable Bose speaker paired almost as quickly as AirPods.
Criticism of AirPods is almost universally from people who've never used the product.
Two of my colleagues were in on the Bragi Kickstarter. They were buggy and went through many firmware updates, but those couldn't fix the poor BT connection between the two.
The 'Magic' of 'Airpods' is the design, marketing, branding, distribution, eco-system advantages.
The exact same product under another brand ... well, we may not have ever heard of them.
Apple is still a 'highly functional incumbent' - and since it's always a bad idea to go up against incumbents even when they are dumb and lazy (usually they are, this is what incumbency does) it's a doubly bad idea to step in front of the Apple Mack Truck.
Apple is still doing the 'evolve iPhone' thing and they can only do that for so long. The 'let's build a big fancy building' step is often sign of peak-ego, we'll see how long they go from here.
Oracle, SAP etc. - look how long they can remain relevant just by doing small iterations in their software - they can hold business clients to the bone.
I wonder if a consumer company can do the same. Their brand has a lot of built up value, and I don't see either Google or Samsung stepping up to the plate ... but cycles are fast in consumer land and consumers are fickle ... and unlike Nike where the 'tech' really is just for show ...
"Their brand has a lot of built up value, and I don't see either Google or Samsung stepping up to the plate"
They do seem to be trying, though, at least if their ad campaigns are any indicator. Google in particular seems to be positioning itself as a viable Apple competitor both in terms of the device itself (in this case, the Pixel line and its myriad of capabilities) and the branding around it (by tying in to the rest of Google's product line, particularly around search).
Samsung, meanwhile, seems to have the first part of that down (and has done so for as long as the Galaxy line has existed, much to Apple's chagrin), but the second part is kind of missing there.
Yes, the marketing piece is hard to operationalize and a 'creative lead' who leaves can ruin it - unlike regular ops where a 'finely tuned machine' can run for a long time with just responsible leaders.
The new CMO is from Burberry, and I think she will keep things humming well.
Johnny Ive is still participating, so they have a design front established.
Engineers can think of cool new things.
So they have a lot of momentum.
But still ...
There was a time before Apple, there will be a time after Apple :) ...
1. Cost and pricing models. This is very difficult especially when you're not big enough to place larger production runs. You don't have economies of scale here. It's expensive to manufacture just a few boards at a time.
2. Manufacturing/supply chain points of failure.
3. Hardware points of failure (buggy vendor chips etc.) which could easily derail a project. Hardware isn't as malleable as software too.
4. Long product development times. Your first version of the product can quite easily stretch out to many months or even years. This is a huge risk from a shareholder's perspective.
To gain some context into hardware manufacturing, I recommend reading:
The Hardware Hacker: Adventures in Making and Breaking Hardware by Andrew Bunnie Huang.
Never mind that Apple has a long running inside track with MSM, and do some serious marketing.
I recall being virtually unable to go a commercial break without seeing at least one iphone feature demo ad right after launch in my part of the world.
With Apple it is just as much about attention as it is actual products. They have managed to make the choice of phone or computer into a lifestyle signal that their competitors can only dream about.
Oh, and i recall the Palm CEO talking about how when they were designing the Pre they constantly ran into Apple having bought up whole factory outputs of part Palm wanted to use. End result was that Palm constantly had to go with their second best choice.
I was part of a hardware startup for nearly a decade.
Most real hardware companies that I recall failed for reasons other than hardware. Which fits with the author's overarching argument.
When people imagine a hardware company failing, I think most imagine something going terribly wrong with prototypes, some failure in the field, manufacturing failures, overbuying, etc, etc. But I think those events are rarely fatal for real hardware companies.
I say "real hardware companies" because those kinds of failures are endemic with crowd funded hardware startups. And I think that's why most people assume that all hardware companies fail for those reasons, because crowd funded hardware startups get so much public attention.
But outside of the crowd funding bubble, my experience has been that those kinds of events are just normal bumps in the road. We were on razor thin angel funding, and we hurdled all those kinds of problems. What really kills hardware startups is what kills _any_ startup. Bad product-market fit.
Over the decade of my involvement with my previous hardware startup, we watched _tons_ of fellow hardware startups die. Not because they hit some manufacturing snag, or shipped a million units of explosive doorstops. Every single one died because no one wanted their product. The classic startup tale.
Perhaps I have a unique perspective on the whole software startup versus hardware startup thing because I'm a jack of all trades. I'm just as comfortable working software as I am hardware (and boy does that duality come in handy). So I've seen both sides from the trenches. Hardware people will tell you it's harder, because they don't know how frustrating real software development is and assume it's some kind of greener pasture. And software people will tell you it's harder, because hardware is a mystery to them.
I've wasted days dealing with Google Cloud's SDK issues. And I've wasted just as many days chasing a defect in a board remotely with a customer half way around the world.
In other words: The grass is not greener on the software side.
Count how many Chinese companies just jack into Apple lookalike market or market for goods designed along the same line of thought.
Take Rapoo for example, once an unremarkable company. What elevated them to the current level was a big bet of proper design. Their first red dot winning product was designed by an intern that they almost had fired for insisting on design that "does not look enough like apple".
I also see no surprise if a company like Jawbone gets into "last mile operational impediment" what in reality was just the usual result of company not having engineering expertise.
Here, I'd say lies the greatest rift in between SV style software startups that take quality compromise in code as something usual and a normal hardware company that knows that a broken product is literally simply broken with all resulting material losses
59 comments
[ 4.8 ms ] story [ 98.0 ms ] threadSo what is it? Hard or not?
1. Apple products are glossy and polished, aiming to capture the premium end of a large consumer market for very complex products.
2. Doing this better than Apple is very difficult.
3. However, other less difficult business opportunities exist.
Apple also maintains very high margins, something that leaves more than enough room for another company to compete on price. The trap here is instead of producing a good product at a better price, most companies produce the cheapest possible product in a race to the bottom.
Find something Apple isn't doing, doesn't seem interested in doing, and use their business model. It worked for Nest. It worked for Tesla. It could work for you.
You just made the case against this. The fact that they won't make an IoT blender is why they're hard to compete with. Apple is the polar opposite of Google in so many ways but a big one is their organizational mentality of "more wood behind fewer arrows". Apple is the batter in a no-called-strike baseball game (to borrow a Buffett quote) which means they're only going to swing when they get a fastball right over the plate.
Apple has high margins but it's actually because they have the lowest costs. In Porter's terms, they're the benefit leader but they're also the cost leader.
>Find something Apple isn't doing, doesn't seem interested in doing, and use their business model.
This isn't actually competing with them then. This advice is probably valid, but you're basically saying to just be the Apple of some other industry (like Keurig or Nespresso).
There's a lot of other industries that need an Apple and don't have one. Who is the Apple of toilets? Of mattresses? Of 3D printers? Of blenders?
Some have leaders that succeed not based on having a superior product, but by their crushing marketing pressure and omnipresent retail representation. They're ripe for disruption.
This is what companies like Anker and Aukey are doing too. They compete directly against Apple's accessories with a well polished, durable product, for cheaper than what Apple charges.
What I hear in this quote is "we went into a technological field we know nothing about with the intention of making money. We shouldn't have done that".
Nothing wrong with going into hardware if you have experience end-to-end bringing a hardware product to market. It's just that less people have that than in the software world, and us software people get used to being able to figure out everything JIT.
Probably don't go into chemistry either, or algotrading, or practicing law, thinking "eh, I'll figure it out as I go".
It did some useful things, but it needs to evolve.
Mind you there's a few similar things in software, things like security requirements when handling payments, EU privacy laws, cookie banners, that kinda thing. But they're a lot easier to either outsource (use a payment processor service) or implement (cookie banners range from a notice that says "by using this site you agree to cookies", which probably isn't enough for the law but reasonably user-friendly, to full off-site cookie walls.)
That's actually a good thing. First revision hardware is going to have issues and building a beautiful product and putting in a beautiful box will set you up for disappointed customers.
Either do what MakerBot did, and embrace the DIY enthusiasts as your first market, or do what Misfit did, downplay the tech features in favor of a more fashion forward aesthetic. But running headlong into the buzzsaw that is Apple seems like a recipe for failure.
* The cost model is different.
* The hiring model is far more interdisciplinary.
* You can't apt-get your supply chain.
* There's near zero real world A/B testing.
* You're at the mercy of far more external factors.
Relative to SaaS, hardware seems Sisyphean.
Not that this keeps me from thinking HW ideas, and SW is great too.
In software, you can write a function that doubles big numbers perfectly until it hits limits of CPU and memory in about one line of code. If you try implementing the same function in an analog circuit, you'll encounter compounding noise depending on heat, ambient EMF, adjacent components, etc.
Upon identification of a critical bug in your SaaS product, one `git push HEAD` is all it takes to get the fix to your millions of users. Hardware is not anywhere near as forgiving--your batch is usually lost.
These characteristics require a different approach for developing hardware-leaning products. Everything needs to be far higher quality before release, because small hardware mistakes can have several order of magnitudes higher of costs than software bugs. That's why it's harder: you have to land a rocket on a planet with fewer chances to correct your course.
My first job was at an HW startup and I launched a dozen hardware products. It's tough. A screw-up at a single soldering station in Shenzen can screw-up a deal years in the making with Walmart. Overordering materials means you have to lay off friends.
But saying "Hardware is Hard" is an oversimplification.
When we say that to explain the failure of HW startups, we miss a lot of important nuances. Hardware is hard, but that's not the primary reason most HW startups fail. Almost all of the high-profile failures we see in hardware are companies that are trying to be the "Apple of X."
In this post, there are two hardware startups, valued at nine-figures, that are essentially bootstrapped. Hardware was hard for them too, but they managed to win with very little in the way of resources.
Also, the idea of taking the B2B route and avoiding consumers is great; consumers have high expectations and are a very well served market.
But all of these are hard -- it only looks easy in the books.
Nope. The "magic" of Airpods (I own a pair) is quick synced pairing. Startups like Dopplr can't provide that "magic" for iOS/Mac users because Apple doesn't enable the requisite access on their software. Dopplr, Bragi etc are better than Airpods but Apple controls many of the devices they need to interface with so they are fucked.
The lesson is don't design your hardware to match Apple and primarily at Apple users. They will cut the legs under you
I just stopped by their site and saw their farewell message. I could 50+ people in the photo. That's 50 salaries this startup had to absorb every month. If we assume 100k real cost per head (recruiting, pay, benefits, etc) then we're at $5m per month on salary alone. If it took 2 year from idea to product, then you're in the hole $120m. If you sell these headphones with a markup and post-everything margin of $30 per set, which seems reasonable, then you need 4m in sales to break even. That's a serious amount of sales to push through.
The numbers look more reasonable now, but ultimately its a hard sell especially when Apple doesn't actually need to make a profit for its accessories, but you do. Even then, what's the market look like for this without an Apple. I suspect not very large and its Apple's own marketing efforts that have gotten people interested in these tiny wireless earbuds anyway.
Do you have data to support that claim?
Amazon might sell hardware at cost or a loss, but Apple most certainly only sells hardware when two conditions are met: a) they can make a better device and b) they can make money
Syncing is usually relatively painless but reliably and "instantly" reconnecting is a long ways off. I have maybe a 50% success rate without going to Settings>Bluetooth and clicking on the device. That inexplicably fails about a quarter of the time and I have to powercycle the speaker and toggle bluetooth off/on a few times until it decides to work again.
That rate drops to basically 0 if the speaker is paired to multiple phones, or there are multiple speakers in range of my phone.
If I had to deal with that multiple times per day like with headphones I'd just give up and return it.
I have one absolutely reliable connection (turntable>bluetooth transmitter>stereo) but even that one takes about 15 seconds to establish after turning everything on.
The pairing part is just a one-time thing and yes, the experience is nice, but pairing ease isn't the primary selling point. My portable Bose speaker paired almost as quickly as AirPods.
Criticism of AirPods is almost universally from people who've never used the product.
The exact same product under another brand ... well, we may not have ever heard of them.
Apple is still a 'highly functional incumbent' - and since it's always a bad idea to go up against incumbents even when they are dumb and lazy (usually they are, this is what incumbency does) it's a doubly bad idea to step in front of the Apple Mack Truck.
Apple is still doing the 'evolve iPhone' thing and they can only do that for so long. The 'let's build a big fancy building' step is often sign of peak-ego, we'll see how long they go from here.
Oracle, SAP etc. - look how long they can remain relevant just by doing small iterations in their software - they can hold business clients to the bone.
I wonder if a consumer company can do the same. Their brand has a lot of built up value, and I don't see either Google or Samsung stepping up to the plate ... but cycles are fast in consumer land and consumers are fickle ... and unlike Nike where the 'tech' really is just for show ...
It will be interesting to see how this plays out.
They do seem to be trying, though, at least if their ad campaigns are any indicator. Google in particular seems to be positioning itself as a viable Apple competitor both in terms of the device itself (in this case, the Pixel line and its myriad of capabilities) and the branding around it (by tying in to the rest of Google's product line, particularly around search).
Samsung, meanwhile, seems to have the first part of that down (and has done so for as long as the Galaxy line has existed, much to Apple's chagrin), but the second part is kind of missing there.
The new CMO is from Burberry, and I think she will keep things humming well.
Johnny Ive is still participating, so they have a design front established.
Engineers can think of cool new things.
So they have a lot of momentum.
But still ...
There was a time before Apple, there will be a time after Apple :) ...
1. Cost and pricing models. This is very difficult especially when you're not big enough to place larger production runs. You don't have economies of scale here. It's expensive to manufacture just a few boards at a time.
2. Manufacturing/supply chain points of failure.
3. Hardware points of failure (buggy vendor chips etc.) which could easily derail a project. Hardware isn't as malleable as software too.
4. Long product development times. Your first version of the product can quite easily stretch out to many months or even years. This is a huge risk from a shareholder's perspective.
To gain some context into hardware manufacturing, I recommend reading:
The Hardware Hacker: Adventures in Making and Breaking Hardware by Andrew Bunnie Huang.
I recall being virtually unable to go a commercial break without seeing at least one iphone feature demo ad right after launch in my part of the world.
With Apple it is just as much about attention as it is actual products. They have managed to make the choice of phone or computer into a lifestyle signal that their competitors can only dream about.
Oh, and i recall the Palm CEO talking about how when they were designing the Pre they constantly ran into Apple having bought up whole factory outputs of part Palm wanted to use. End result was that Palm constantly had to go with their second best choice.
Most real hardware companies that I recall failed for reasons other than hardware. Which fits with the author's overarching argument.
When people imagine a hardware company failing, I think most imagine something going terribly wrong with prototypes, some failure in the field, manufacturing failures, overbuying, etc, etc. But I think those events are rarely fatal for real hardware companies.
I say "real hardware companies" because those kinds of failures are endemic with crowd funded hardware startups. And I think that's why most people assume that all hardware companies fail for those reasons, because crowd funded hardware startups get so much public attention.
But outside of the crowd funding bubble, my experience has been that those kinds of events are just normal bumps in the road. We were on razor thin angel funding, and we hurdled all those kinds of problems. What really kills hardware startups is what kills _any_ startup. Bad product-market fit.
Over the decade of my involvement with my previous hardware startup, we watched _tons_ of fellow hardware startups die. Not because they hit some manufacturing snag, or shipped a million units of explosive doorstops. Every single one died because no one wanted their product. The classic startup tale.
Perhaps I have a unique perspective on the whole software startup versus hardware startup thing because I'm a jack of all trades. I'm just as comfortable working software as I am hardware (and boy does that duality come in handy). So I've seen both sides from the trenches. Hardware people will tell you it's harder, because they don't know how frustrating real software development is and assume it's some kind of greener pasture. And software people will tell you it's harder, because hardware is a mystery to them.
I've wasted days dealing with Google Cloud's SDK issues. And I've wasted just as many days chasing a defect in a board remotely with a customer half way around the world.
In other words: The grass is not greener on the software side.
>Do the opposite of what Apple would do
Count how many Chinese companies just jack into Apple lookalike market or market for goods designed along the same line of thought.
Take Rapoo for example, once an unremarkable company. What elevated them to the current level was a big bet of proper design. Their first red dot winning product was designed by an intern that they almost had fired for insisting on design that "does not look enough like apple".
I also see no surprise if a company like Jawbone gets into "last mile operational impediment" what in reality was just the usual result of company not having engineering expertise.
Here, I'd say lies the greatest rift in between SV style software startups that take quality compromise in code as something usual and a normal hardware company that knows that a broken product is literally simply broken with all resulting material losses