I've been the person doing that testing. Checking interoperability with a range of other devices is one way you can look for problems but when it comes to signing things off the standard is the final word. Do you know what happens to the devices that don't comply with statutory radio transmission standards? They don't get sold.

Typically the cost of fixing the problem and going through the whole certification process again from the start is significant. There are some direct financial costs but mostly the damage is increased time to market. That's a great incentive for the maker to get their act together and comply with the standards like everyone else as quickly as possible. If you screw up too many times your product might be obsolete before you're ever allowed to sell it.