4 ms·

Dealing with suppliers and Chinese factories may present a logistical challenge, but it's still easier and less expensive than doing the manufacturing yourself.

Dealing with suppliers and Chinese factories may present a logistical challenge, but it's still easier and less expensive than doing the manufacturing yourself. If you have experience developing a hardware product, I'll defer to your experience, but what I've seen from being on the periphery of a couple hardware projects doesn't jive with what you're saying are the lessons learned from trying to ship a polished hardware product with a small team on a limited budget.

msandford 12y ago
> Dealing with suppliers and Chinese factories may present a logistical challenge, but it's still easier and less expensive than doing the manufacturing yourself. I sort-of agree with you there and sort-of don't. IF (and this is a HUGE if) you already know EXACTLY what your product is going to be THEN you MIGHT be able to save a bunch of money using Chinese suppliers and factories. But when you're doing a hardware Kickstarted kind of product you don't necessarily know exactly what the product is and how it will work. You might well be inexperienced too. Manufacturing in-house gives you the ability to rapidly iterate and close the feedback loop (which any good software engineer will tell you is critical) so that you can go from sorta-working to working to working perfectly. 3d printed parts are not the same as injection molded parts for myriad reasons. The screws for assembly you buy off of mcmaster carr will almost assuredly be different than the ones your Chinese supplier gives you, especially if you don't know to ask what the material is, what the tolerances are, etc. A screw is a screw right? Not even close. Software guys often have a lot of trouble understanding hardware because software is so pure. A bit is a bit and it always will be. An object has a type which often fully defines it. Bits can largely be counted on not to flip. If it passes the tests in dev and test it should work in production, and if it doesn't we can revert it. When you do all this stuff in China your feedback loop might be 3-6 months long. That means that you might spend two years (4 to 8 iterations) getting from an idea to a worthwhile product. Doing manufacturing in-house can bring your feedback loop from months to hours, days or weeks. For the same 4-8 iterations you might only spend a few months. That's the difference between "very nearly success" and "quite late, eventual delivery way behind schedule and expectations"
krschultz 12y ago
I've worked at a couple companies that build hardware. In my experience you need at least 1 mechanical engineer on the company payroll. I don't think you need to own the tools or the factory, but building hardware is as difficult if not more so than software. No one would fund a software startup where they didn't have a single FT software engineer, yet for some reason people think that "it's just plastic" so you can contract it all out. Nothing could be further from the truth. There is a lot of nuance to getting an injection molded design right.
bsder 12y ago
> Dealing with suppliers and Chinese factories may present a logistical challenge, but it's still easier and less expensive than doing the manufacturing yourself. Not true for mechanical things. Injection molding machines are now all electrical and sit in the quarter million dollar range. CNC machines similarly. A million dollars gets you a very nice mechanical production line and machine shop that lets you fabricate precision metal parts AND high volume plastic parts. Now, your only expenses are salary and production inputs. I deal with this every day. How many tooling charges at $5,000 to $10,000 per charge do you have to take before that machine shop makes sense? Not many. I've eaten that charge 22 times in the last 3 years. I'm starting to accumulate the machine shop...
- msandford 12y ago
  I am doing what you're talking about but with old, used equipment. You can get old injection molding machines for less than $50k each, maybe much much less. It's the same for used CNCs. If you know a little bit you can tell if it's still in good shape or if it needs work. I will put together a decent production line for much less than a million, probably much less than $100k in capital equipment. At that point the rent and air conditioning start to become substantial (Texas, so it's HOT). Every time I read someone saying "software is eating the world!" and all the smart people rushing into software I always have a little chuckle. I really love it. It means that anything that's not software has a dearth of smart folk and that puts me at a huge competitive advantage.
  - bsder 12y ago
    I don't find many all-electric injection molding machines used, yet. Otherwise, yes, lots of used stuff is available.
    - msandford 12y ago
      Yeah I don't care about the specific technology. Hydraulic is fine by me because my needs are relatively speaking low volume. One piece a minute is plenty for the foreseeable future and that isn't challenging in the slightest. I'm also partial to having a couple of machines instead of just one because of the reliability. If I spend $250k upfront (which I don't have) I get one machine that's going to be extremely reliable but if it does break I'm losing you know $500 an hour or whatever it is. If I spend $20k on one machine and $30k on another and $50k on a third I end up with no single point of failure. Yeah they will break more but I'll (probably) never be stuck with no ability to make parts. It's easier capex too since revenue from the first finances the second and revenue from those two finance the third. What kind of stuff do you do? I don't interact with many people on here who have mechanical aptitude. Obviously products of some kind. Can you talk about it? I'm intrigued.

msandford 12y ago

> Dealing with suppliers and Chinese factories may present a logistical challenge, but it's still easier and less expensive than doing the manufacturing yourself. I sort-of agree with you there and sort-of don't. IF (and this is a HUGE if) you already know EXACTLY what your product is going to be THEN you MIGHT be able to save a bunch of money using Chinese suppliers and factories. But when you're doing a hardware Kickstarted kind of product you don't necessarily know exactly what the product is and how it will work. You might well be inexperienced too. Manufacturing in-house gives you the ability to rapidly iterate and close the feedback loop (which any good software engineer will tell you is critical) so that you can go from sorta-working to working to working perfectly. 3d printed parts are not the same as injection molded parts for myriad reasons. The screws for assembly you buy off of mcmaster carr will almost assuredly be different than the ones your Chinese supplier gives you, especially if you don't know to ask what the material is, what the tolerances are, etc. A screw is a screw right? Not even close. Software guys often have a lot of trouble understanding hardware because software is so pure. A bit is a bit and it always will be. An object has a type which often fully defines it. Bits can largely be counted on not to flip. If it passes the tests in dev and test it should work in production, and if it doesn't we can revert it. When you do all this stuff in China your feedback loop might be 3-6 months long. That means that you might spend two years (4 to 8 iterations) getting from an idea to a worthwhile product. Doing manufacturing in-house can bring your feedback loop from months to hours, days or weeks. For the same 4-8 iterations you might only spend a few months. That's the difference between "very nearly success" and "quite late, eventual delivery way behind schedule and expectations"

krschultz 12y ago

I've worked at a couple companies that build hardware. In my experience you need at least 1 mechanical engineer on the company payroll. I don't think you need to own the tools or the factory, but building hardware is as difficult if not more so than software. No one would fund a software startup where they didn't have a single FT software engineer, yet for some reason people think that "it's just plastic" so you can contract it all out. Nothing could be further from the truth. There is a lot of nuance to getting an injection molded design right.

bsder 12y ago

> Dealing with suppliers and Chinese factories may present a logistical challenge, but it's still easier and less expensive than doing the manufacturing yourself. Not true for mechanical things. Injection molding machines are now all electrical and sit in the quarter million dollar range. CNC machines similarly. A million dollars gets you a very nice mechanical production line and machine shop that lets you fabricate precision metal parts AND high volume plastic parts. Now, your only expenses are salary and production inputs. I deal with this every day. How many tooling charges at $5,000 to $10,000 per charge do you have to take before that machine shop makes sense? Not many. I've eaten that charge 22 times in the last 3 years. I'm starting to accumulate the machine shop...

msandford 12y ago
I am doing what you're talking about but with old, used equipment. You can get old injection molding machines for less than $50k each, maybe much much less. It's the same for used CNCs. If you know a little bit you can tell if it's still in good shape or if it needs work. I will put together a decent production line for much less than a million, probably much less than $100k in capital equipment. At that point the rent and air conditioning start to become substantial (Texas, so it's HOT). Every time I read someone saying "software is eating the world!" and all the smart people rushing into software I always have a little chuckle. I really love it. It means that anything that's not software has a dearth of smart folk and that puts me at a huge competitive advantage.
- bsder 12y ago
  I don't find many all-electric injection molding machines used, yet. Otherwise, yes, lots of used stuff is available.
  - msandford 12y ago
    Yeah I don't care about the specific technology. Hydraulic is fine by me because my needs are relatively speaking low volume. One piece a minute is plenty for the foreseeable future and that isn't challenging in the slightest. I'm also partial to having a couple of machines instead of just one because of the reliability. If I spend $250k upfront (which I don't have) I get one machine that's going to be extremely reliable but if it does break I'm losing you know $500 an hour or whatever it is. If I spend $20k on one machine and $30k on another and $50k on a third I end up with no single point of failure. Yeah they will break more but I'll (probably) never be stuck with no ability to make parts. It's easier capex too since revenue from the first finances the second and revenue from those two finance the third. What kind of stuff do you do? I don't interact with many people on here who have mechanical aptitude. Obviously products of some kind. Can you talk about it? I'm intrigued.