An argument can be made for replacing brittle, difficult to follow twisty spreadsheet programs (which is a lot of what people do with Excel that they can't do with anything else) with more conventional programs; maintainability benefits could be significant, just to pick one low lying fruit. But there's no money in it.

I was talking more about mainframes than Excel. I also posted that before the thread was full of comments that changed my view on mainframes.

As for Excel, I've read about a lot of situations where it's used for things it's not a good fit for, especially situations where a simple database and frontend might be more appropriate.

It isn't so much Excel as it is spreadsheets in general. The cycle I've seen is that when a group starts out, they use Excel because it is a rapid tool that they understand. But then they keep building on that until they have a very complicated mess on which they are dependent. Spreadsheets don't tend to have good tools for testing and can be very brittle, so they end up slowing things down and causing problems.

There are some interesting case studies here: http://www.eusprig.org