Commercial products are, first and foremost, designed to meet a business need. They exist to make a profit for those people who design, build and sell them.
If that business requirement can be met by lashing together a few off-the-shelf boards and hiding the mess in a box, that may be just as valid as something engineered to the 'n'th degree and signed off by a dozen committees before going into a million safety critical products. It's simply not possible to say that "commercial products all require... <thing>", because "commercial products" is such an enormously wide category.
For some, the component cost is a complete non-issue. In others, shaving off every fraction of a cent is essential. There's no universally applicable rule.
The same goes for documentation, process and sign-off. If you're making an automotive ECU with authority over engine and brake control, then of course the design must be completed according to prescribed procedures, documented, tested and so on. On the other hand, if you're making (say) a protocol converter for laboratory use, with the expectation of selling a few hundred a year, then no, it won't have been through any of that.
It's perhaps more useful to look at what defines a hobbyist product, and there are some common elements:
- designed primarily for the fun of it, not because it's expected to be commercially viable. Profitability is a nice-to-have, not essential.
- overall scope limited by the time and resources available to the person developing it.
There are plenty of other attributes which people might be tempted to apply only to hobby projects, but they're not really differentiators...
- likelihood of containing egregious design errors due to lack of knowledge and experience - definitely not limited to hobby projects only.
- poor quality of documentation - though commercial projects tend to suffer more from bad translations and an excess of unnecessary warnings, rather than simply not existing at all.