I make a good living working on large corporate sites (am fully overbooked, so not fishing for work) such as this pair and such madness is fairly normal.
The problem lies primarily in corporate funding processes relating to any form of IT/marketing activity and in the way the website is built initially and then updated in dribs and drabs. The effort put into getting next year's budget is so intense that people no longer seem to have the energy to do anything else in a sensible manner.
Even if someone in Marketing has heard about UX (user experience), it is hard to get an internal UX person on the headcount. And even with such a person on the payroll, large sites are maintained by large teams and things slip through the cracks. While major facelifts are done by external agencies, ongoing work is usually done internally.
Ongoing work (e.g. a "little widget on a page") is broken down into little projects and it would be unusual for those projects to step back and look at the whole page in case something else might break. There isn't the funding for such luxury. You want a widget, you get just that and the funding is usually not enough, so something always gets "descoped". If you mention regression testing, you're probably too old.
And there's usually some sad backstory such as some legacy systems that would cost too much to replace and Finance would not want the cost blowout to affect the share price.
So far I haven't even mentioned the calibre of the coders and how hard they work; the usual corporate politics and rivalry, etc. So, even in a happy corporate tower, UX has very little chance of winning. End users have to lump it.
In the case of VA and Velocity, I haven't any knowledge of their history, but the reasons for not being one site and not having "single sign-on" would be one of those sad stories.
