As someone recently participated in an EMC measurement, I really don't understand how anyone passes these tests without some kind of cheating (using double-shielded, very expensive industrial cables + hacking with functional earthing).
I'm sure many devices on the market are not compliant. I follow lists of products removed from market for non-compliance and there are plenty each week. But the fact is that, if you're not in one of the categories that are under special scrutiny (aerospace, automobile, medical, etc.) or do something grossly incompetent (e.g. interfere with a mobile operator or someone else with a similar power to put you into the regulatory spotlight) you're unlikely to get into trouble for shipping a non-compliant device.
Make someone's Wi-Fi a bit slower and a bit more packet-lossy? Chances are nobody will care. It's a sad state of affairs really because pervasive radio interference is just making things worse for everyone.
Another classic example is early cellphones that you would pick up on stereo systems etc. - "dat-dara-dat-dara-dat-dara-dat-dara-daaaaaa" going out in full blast.
Some years back when I had some small involvement in new EU regulations that case was actually given as an example of how, even after many reviews, some forms of harmful interference can only become apparent after a technology is already widely deployed.
The RF breakthrough is caused by the Amplitude Modulation in GSM, but then there's nothing fundamentally wrong with AM.
Unlicensed equipment in the ISM WiFi band are operating under Part 15 (in USA) so must accept any interference.
And the phone interference to a stereo is caused by cheap and nasty design of the stereo. The phone itself cannot be blamed, and there's nothing which can be done to the phone to fix the interference (other than switching it off). Again the stereo is probably operating under Part 15, so must accept any interference.
The first two options are somewhat expensive, the NRE for the power supply design isn't attractive to manager types and conductive coatings for plastic or a metal enclosure are not the cheapest options. But if you're dead set on compliance it's better to frontload the design costs instead of iterating 3 times before you get to market.