Information Systems Users don't know what they want... So, don't ask them what they want: and definitely don't write some software and then say "how's this look?".
On the other hand, don't have someone spend weeks talking to various people about what they want ( and not get to talk to other people ), write it all down and deliver a document, saying "build this"... and then spend months building "this" and then say "here you go".
But, something must be done; start with a change of focus - Information Systems should not do what users want, Information Systems should do what an organization needs to be successful.
Isn't that the same thing? It should be, but users are people and one person seldom wants what other people want; and people are notorious for wanting one thing now, and something else next month... or even tomorrow.
So change focus - from people to organization, and from wants to needs.
The User Myth
Please keep in mind that the scope of my opinion is Information Systems, mainly the type of systems that people are paid to use, people otherwise known as 'employees'.
A lot of systems and software are not in this scope. For example, what I have written here was first drafted using a notes app on my mobile device; that is a whole different thing, one I always want to explore... But when I go out to work for clients, it is almost always for Information Systems: "we need a new system to process more orders" ... "our loan approval system has to be expanded" ... "the State is changing the rules for UI eligibility" ...
So let's address the User Myth.
First, I am not the first person to note that the only other people who are referred to as "users" are drug addicts, but it seems to be a term we are stuck with. Seriously, when I drive, am I a car user?
Now consider those employees I mentioned up front. When they are hired, they always get trained on the existing systems they need to do their work. Given the multiple-year lifespan of a typical information system, and the usual level of turnover in a department/function of an organization, a lot of people will use a system over time. Very few of this number will have been there for the introduction of the system, and usually none of those people will still be around when the system is retired and replaced with something else.
And yet we are supposed to assume the group of people using a system at it's retirement are automatically the experts who should be asked what they want in a new system? Think about this for a moment or two; that should be all the time you need to conclude "well, no".
That's because users come and go, but systems (seemingly) go on forever.
The current group of employees who are expected to use a new system are indeed important stakeholders and participants in creating that system, but they are not the only ones. (And still, some of them could be gone before the system is finished.)
So change your focus to the needs of the organization; stay tuned for my approach on doing just that.