@mike,
when I look at your maps (both, from this post, and from Map of University Undergraduate Education), it becomes apparent what kind of challenges are in front of universities. While I do not live in UK, the major problem is the same - we had a university boom, when it was much easier to get a reasonable job with a degree than without it.
This caused 50% of population to study, which caused an educational bubble, and it is about to burst as people realised that you no longer have to have a degree to get a reasonable job. Naturally, a degree often stays a critical requirement in some sectors (health, law), but increasingly it becomes unnecessary effort if you can follow an alternative route.
Across the past couple of decades, some study fields become a job trainings. They were not (and still are not) expected to let you study a field of your interest, but rather focus on practical skills that can be monetised immediately after graduation, or even before. This user need is a subject to disruption (nanodegrees, different career paths) and I am afraid this is nothing we can do. Frankly, I do not think we should do anything at all, as majority (I think) graduates does not need to study.
But that means universities main income source will cease to exist. In the case of UK - private tuition fees are being redirected elsewhere, in Poland - state pays for a student, less students -> less income for the university. Which is a huge problem.
One direction here would be to embrace the no-degree path. Start mooc, and employment-oriented short and affordable courses. Of course, there is a question of the market size, user needs, etc… plenty of learning.
Also, this disruption forces universitites to the position where I think they should be - doing true research which later can be commercialised. But I can see a huge pitfall there - an expectation that the university will be capable to earn enough by selling innovations is a very bold one. It assummes that the university is capable of predicting which research can be commercialised and, what is more problematic, the time between research and commercialisation is relatively short.
I would say that if targeted research could be achieved today, we would already had companies doing it repeatably. But I fail to observe them - businesses appear around an idea, and live as long as the market niche is there. Then they fail to reinvent themselves (find a new niche) and die. There is a couple of exceptions (Nokia), but even those went out of businesses eventually. I do not see it working for universities.
Universities, however, bring enormous value in building contact network (if you do not have one). I am not sure, though, if there is any competition in this space from non-university organisations. I can think of meetups and other gatherings serving exactly same purpose.
Finally, I think we should get back to a central funding of a research. Decide apriori how much risk we are willing to take (% of GDP), and how much of this should go into research expected to improve common well-being in a distant future (Genesis), and how much should go, if at all, into improving existing solutions (easier to commercialise).