Still, hardware is, well, hard. Silicon Valley has produced a raft of world-changing software startups, from Airbnb (a YC veteran) to Facebook. But it's a whole lot easier to beta-test an app than to prototype and then manufacture a gadget with a bunch of moving parts. Then you have to market the thing to the masses, who are already enamored of their Apple and Samsung products. Exhibits A and B: Fitbit, the fitness tracking company, and GoPro, maker of rugged little cameras. This week, both cut their sales forecasts for the crucial holiday quarter and watched their shares plunge. (Neither started life at YC.)
Fitbit and GoPro's downturn is mostly fueled by mismanagement and poor decisions than hardware being intrinsically hard.
The hard part of hardware is building something that works well enough and is in demand. When you have your device on people's wrist or mounted on their helmet, you have overcome the "hardware is hard" part. (for the most part).
After that it's building on your original success, expanding your market and creating value for your customers; all of which are true for any other business, not strictly hardware. For fitbit and GoPro, the latter has proven more difficult.
I don't know abut GoPro, but Fitbit deserves to fail. Their hardware is terrible. The warranty only lasts one year, yet I've had to replace both Fitbit devices I've had twice within that one year. I'm now Fitbit-less because the latest, a Charge HR, failed for the third time, and it's now out of that one year warranty.
To give you an idea of how bad the Charge HR is, its charging plug attaches with metal prongs, but it plugs into a slim plastic housing that is only glued to the rest of the device. The obvious result is that the plastic housing eventually gets pulled off with the plug. That's how two of my three Charge HRs have died. It'd be like if your laptop were ruined within four months because the power cable pulled off an essential piece of the laptop housing rather than detaching cleanly, and now it won't attach in order to charge. The third Charge HR bricked itself during an unprompted firmware update, which started randomly and then I guess got interrupted midway through when I moved my phone too far from the device. It never recovered.
I don't know what the hell they're optimizing for over at Fitbit HQ, but they're trying to sell $10 throw-away quality devices for >$100. It's outrageous. I actively discourage everyone I know from ever buying anything from them, and many other people I know have also completely sworn them off for the same reason. Their product just sucks.
I forgot to write my conclusion because I got so livid and carried away about how much Fitbits suck, so here it is:
I don't think Fitbit's troubles are necessarily an indictment of the hardware industry as a whole. I think they're particular to that company having a horrible product and horrible hardware engineering, to the point where poor build quality actively repulses users rather than creating repeat customers. GoPro similarly has lots of issues that I've heard from others, but I don't have first-hand experience like I do with Fitbit so I won't spend any further time on it.
Contrast with, say, Apple, which is still a hardware company more so than it is anything else, and has become the largest publicly traded company in the world by doing so. I'm not a huge Apple fan by any means, but at least the devices don't suffer from such horrible flaws. My next laptop might not be another Macbook Pro, but it's at least in the running as an option because I've liked the one I currently have. Fitbit does not have that same charity from me.
> they're trying to sell $10 throw-away quality devices for >$100
I can say their software is in the same category from personal experience. Their Android app was written through minimum cost outsourcing, has no comments (since the developers didn't speak English well), a half dozen layers of abstraction when two would work, and is riddled with countless poorly documented strange edge cases and hacks throughout that make maintainability and enhancement a disaster.
On the user side, this is why features are rarely added and the software tends to lock up and crash a lot, and can brick devices during firmware update. Fixing even trivial bugs or adding basic features like social or counterfeit detection right takes much more work than it should since you have to navigate all that and may break something else while trying to do what you are doing.
They hired a few internal people around the time of the IPO, but they universally don't do much. The tech lead of the team generally came in, complained about some random thing that wasn't even accurate to make it seem like he did something, then left at 4PM. He wasn't in the office enough to know anything about what his team was doing.
A typical internal developer played Clash of Clans all day. When showed logs of his code not working he called meeting after meeting with management to get the guy with logs of the failing use case fired. The culture there is that there's no reason to do work there if you can office politics with your friends to get out of it.
Even given a reasonable commit adjusting timeouts to match the time taken for Bluetooth operations to the slowest observed device and fix several long running file transfers timing out by resetting the timeout each block transferred, he blew up and removed all timeouts for file transfers (bringing back the potential lock up issue) and called a half dozen meetings to fight it.
The culture there is just noxious and actively fights anyone trying to do work.
Wow, thank you for the insider perspective. That's truly horrifying that a $2B company (previously $10B) that depends heavily on software outsourced so much of its core competency like that. Talk about being penny wise and pound foolish. I've worked at two ~$10M companies that used all in-sourced development (i.e. me and my coworkers) and, while not without their faults, were significantly higher functioning.
Also, how the fuck do you brick a device by updating it? Wasn't this problem solved decades ago? You upload the new image and then, only once it's completely transferred and verified, do you atomically switch over to it. Are you telling me that they really made it so that the firmware update overwrites the existing firmware, so that any interruption in the process causes it to go poof? Ugh. I know that requires double the internal memory, but c'mon, this is 2016. The cost of that is some very tiny insignificant fraction of what an RMA costs. Never before in my life have I ever irrevocably bricked a device by updating its firmware.
we build hardware (flair.co) and have just started shipping (woot!) and I think the trick with hardware is figuring out how to do the 'software wrapped in plastic'. If goPro was quicker to move into video storage and editing to go with their cameras or fitbit managed to push on fda approval and share data with your doctor (or their own doctors as a service) they could do some really cool stuff (and have really cool revenue...). Its not that as a stand alone product its not compelling - reminder, they IPOd... Looking at them and calling them failures is such a myopic perspective if you talk to any preseed, seed or A stage company. The trick is building a killer product and not worrying about the service if you can. Then transition over to a service model, selling leads, data, tools etc. If you can layer on services early on, thats fine, but not needed. If you never do it, it will come back to bite as challengers show up and erode margins and you saturate your 'ideal customer' market.
Yeah. Anyone remember WakeMate? They had nonstop problems with their hardware. And Lockitron had yield issues for two years on their deadbolt operating robot before cancelling it. Hardware is hard.
Personally I don't mind some jokes here but I can't see what the joke is you're trying to make. If it was a clearly funny comment it perhaps wouldn't fair as badly.
Which is funny to me, but the parent might as well have been referencing something their mate told them; how much of the global audience on HN knows that skit. Anyway, we're well OT.
Based on some analysis I've read, GoPro's growth is somewhat limited by market saturation: a significant proportion of the people who are likely to be enthusiastic action camera users already have an action camera.
Fitbit and GoPro's downturn is mostly fueled by mismanagement and poor decisions than hardware being intrinsically hard.
The hard part of hardware is building something that works well enough and is in demand. When you have your device on people's wrist or mounted on their helmet, you have overcome the "hardware is hard" part. (for the most part).
After that it's building on your original success, expanding your market and creating value for your customers; all of which are true for any other business, not strictly hardware. For fitbit and GoPro, the latter has proven more difficult.