Here's what it boils down to: UBS has lots of data and everybody talks about turning Big Data into Big Money. This is exactly what UBS want to to. The problem: UBS lacks the expertise and know how to tackle this kind of problem. After all, they're a bank. So, what can they do?
They could go to one of the many new Big Data consultancies that spring up like mushrooms these days and ask them for a solution. Yet, that is risky. How can a bank determine if such a consultancy is worth the money, especially in a field where those companies are relatively young. UBS could contact many consultancies. Of course, that will drive prices even higher. There must be a better way.
Indeed, UBS has found a better way. They do the UBS Innovation Challenge. Everybody can participate with ideas tailored to the UBS problem domain. UBS screens those ideas, selects a "best idea" and gives the winner SGD 40'000. There are additional costs, such as flights to Singapore and accomodations for the final participiants, but the costs for UBS are manageable.
So, UBS benefits in two ways: 1) They are presented with not only one or two ideas but with many ideas. 2) The costs for these ideas is known in advance.