As someone who used to work at an IoT company the consumer space is brutal. Unless you've got something you can sell for 10x BOM, or there's naturally a subscription based model then it's not worth the blood, sweat, and tears.
QA/QC is very expensive, especially for low volume. You send something out to a customer, they claim its broken, you ship out a replacement at your expense, maybe the customer eventually sends back the broken one, then it sits on the shelf for 2 years because you don't have the manpower to do a root cause investigation. Now you've sold 300 units and have replaced 50 of these and you have no idea why they're broken, what went wrong, and have to a a full validation of each one from scratch. Most of them seem to work fine in your test lab (maybe it was the customer's fault? But maybe your design is incompatible with their production environment?) and the rest all have unique problems which each take 1-5 full days of investigation to nail down a root cause.
Maybe 2-3 share a single root cause, and it's not clear how you'd prevent this from happening in the future.
If you're shipping large volumes, this can add 10% to the MSRP, but if you're shipping low volumes, $50-100/hr of weeklong diagnosis and investigation can easily add 50-300% to the price of each unit.
And that's if everything else is going perfectly. Which means you lucked out and found an amazing Chinese contract manufacturer who works with you hand-in-hand to fix any design bugs and manufacturing issues, and ensure parts availability.
If you're doing such low volume, yet your service and customer relations is based on big corp style, sure you're going to have big problems.
If you sell only 300 of something, and returns are coming in, you should be figuring out why ASAP. And probably talking to the people before they attempt to send it back.
That's pretty much what they are saying as far as I can tell? My experience is pretty similar to theirs, root cause analysis takes a lot of time per device, and there's rarely once single failure cause for all of them.
The math is this:
BOM + cost to acquire a customer + cost to sell the unit < sale price
Otherwise you don't have a business. Please don't ask me how I know this.
10x is a rule of thumb, YMMV. You just need to make the above inequality evaluate to true. You can (maybe) raise money to deal with the various non-recurring expenses. Oh and you should pay your employees too.
It's not that it's hard to do the math, but it's hard to find product market fit where that inequality is satisfied. And sometimes people think they are gods gift to hardware design and can't be reasoned with to do a cost down.
NRE into the production line, especially hard tooling for plastic is huge. Plus all those things you mentioned, especially inventory risk are often underestimated.
Don't forget that any manufacturing run itself is a huge risk. Tiny changes that need to be made to address anything from RF performance to fit and finish can invalidate a run and just burn up hundreds of thousands to millions of dollars (depending on scale)
Even for something like HDPE? I did a project a while ago that was effectively a pontoon workboat. We used HDPE pontoons in lieu of stainless steel and the cost savings was about 25%. I assume that the raw materials (basically large diameter HDPE pipe) and plastic welding tools were just that much cheaper/easier.
I was interested in doing a new HDPE product development on my own, but I'm a bit nervous this is a deep tooling rabbit hole.
Supply chain into consumers is brutal: retail outlets mark up products by 100% (ie you sell them something for $100, they sell it for $200). If you use a distributor, that's another 100% markup. 10x BoM on consumer price means that some of your biggest sales channels to consumers (BestBuy, Walmart, Target) are paying you 25-33% of what they get from customers. To get your own markup, you need 8x BoM.
