Fake it before you make it
The last two companies I have worked for were bringing a 'product' out from proof of concept and into production, not thinking ahead (perhaps willful negligence) would apear par for the corse
Neither company had any idea what compliance would be required *before* they had sold "production units"
Each company believed that turning something that (mostly) "works" in a tightly defined test setup would be fine for production.
* Repeatable production was never considered
* Assembly time and cost had not been considered
* Production scale up and logistics were so far over the horizaan as to be laughable.
* Production volume was assumed to reduce cost but didn't take into acount changes that would be needed to realise these savings for volume manufacture
* EMC was never considered
* Laser Safety (one product used Class III Lasers)
* Intended final performance (at this point neither company had hardware capable of performing at the speeds they were stating would be achieved in the product - 1 was limited by memory bandwidth and the other by working voltage on stepper motor drive FETs BOTH thought this was a software problem)
* Neither company had a working remote update for firmware (or even host OS and application software)
* Both companies were courting international sales but were struggling to support a 'test/trial' customer site a few miles away
Both companies needed a new round of prototype hardware before they could consider a sale
Neither company had budget for this nor the time scale
Both companies were pushing new features before they even had hardware or software capable of delivering a minimum viable product.
Budgets were wishful thinking and weren't even considered. Items that were estimated at one price turned out to be magnitudes more expensive...