I think UX is often miss-interpreted as design. Design is often an acquainted taste and therefor it is pretty much untestable, but usability testing is something different. It is important that user can use the product with only minimal training, but more important it should feel logical and non repetitive within the users mental model.
During evaluation of a new products, I spend less time thinking about possible security and or performance issues, then if it delivers what I expect from it. If the user experience is bad, I do how ever question if the product will also lack in the other departments. I think that UX therefor also functions as a form of marketing of the products reliability, although it might be irrelevant, it might set the tone of my thought process.
Just like security and or performance testing, usability testing is often not performed at all unless a client asks for it or when issues have arisen in the field recently. Its relatively expensive and the return on investment feels very low. The sales team can't sell a product without the needed features. They can sell stuff that is insecure, slow and which has a horrible user experience.
The user experience is the one thing that is most in your face and will make users happy far more then security and performance, unless your a bank. Since users do not care about security. Performance is only relevant if they know better and faster systems, but if its the only tool for the job often they will probably accept slowness as well. But extra clicks, ohboy! A chance of getting RSI, help! This is a real issue for most user groups. :)
A good read is Rocket Surgery Made Easy: The Do-It-Yourself Guide to Finding and Fixing Usability Problems, it suggests a solid strategy to find usability issues with only a fair amount of effort.
I am not really sure you can compare security, performance, reliability and usability testing. They all cover a different aspects of the software product lifecycle. Also the people who are concerned about these "requirements" topics have different roles and responsibilities within most companies.
- Security: System administrators and public relations (due to privacy issues)
- Performance: System admins, users and managers
- Reliability: Legal department
- Usability: Users and (efficiency) managers
I would wonder how can we estimate the risk of each group and where should we focus our efforts on. Development teams are always under time pressure, making their main similarity a question of can we afford to skip this type of testing or can we not.