Is it because of UX's intangible nature and therefore is difficult to measure? Or is it because businesses simply do not understand or care about it enough?
What is your opinion?
1. Cost focus. In a cost-focused organization (as opposed to revenue-focused), such as universities and certain enterprises (mostly, IMNSHO, large SaaS enterprises), much of the development is performed by lowest-cost developers, meaning those without UX experience or training.
2. Training. Project managers don't know better. In order for the PM "customer" to evaluate good UX, (s)he must know what good UX is. Without some exposure to heuristics to evaluate a good US, it won't ever be a consideration in planning or accepting the project.
3. The larger the organization, the more likely the PM "customer" is not an actual user of the product. Corollary: the larger the organization, the greater the likely distance between the developers and the customers.
4. Inertia. Some large companies still use terminal systems or require IE6. It's more difficult to change than it should be.
5. User lock-in; lack of substitutions. In environments where users don't get to choose which system to use (e.g., students using a university's learning management or registration system) and can't turn to substitutes, the incidence of poor UX seems higher. These are also situations where users have little bargaining power relative to the developers.
In my personal opinion your assertion that it's overlooked by business is not really accurate... it's more that business people don't know how to approach UX design, nor how to recognize talent in UX design.
Just my 2 cents.
Enterprise UX isn't about making the product easier to use, it's about making it more efficient to use, in most cases. I remember working for a large Fortune enterprise company many years ago, and they had a product that, simply put, was for customer, bug and issue tracking. There are a dozen or so REALLY good startups in this scene now (there weren't then) and they're all radically easier to use than the product we had, because they focus on the kind of UI/UX you're talking about.
That said, there was a laser sharp focus on making the product more efficient to use. We hired consultants to use the product extensively, and determined that moving some of the buttons around was a more efficient workflow. We determined that DEcreasing the font size allowed for more information to be on the screen at a given time, which allowed highly trained personnel the ability to do work more quickly than if they had to constantly scroll back and forth. Things like this fly in the face of what is largely considered good now, but was not only a priority, but a highly flaunted (and appreciated) selling point.
The biggest difference is the customer. Most of the SaaS customers are for things that aren't used religiously, or by a lot of undertrained / underqualified people. If you have a help desk for a 50,000 person company, chances are you're not going to use FogBugz and the like to run your company with. You're going to have something in-house, and you're going to train your people to use scripts, and you're going to have a really efficient (if not necessarily GOOD) help desk that uses these tools day in, day out, repeatedly. They get to learn the product well enough that they don't have to look at it to use it. They learn it as well as most of us know how to type. They know that they can enter last name, tab, first name, tab tab tab phone number enter and voila, there's the customer record.
Edit: This was supposed to be a reply to http://news.ycombinator.com/item?id=3052236