As an employee at 90's startup trying desperately to be relevant in the 2000's, I have access to the corporate idea wiki. Most of them are pretty bad ideas. Some people neglect researching their ideas (a protocol like ICPM quench was suggested), some try to port the latest trends to the corporate world. In all these ideas (no comment; proprietary crap is still proprietary), I've noticed three recurring problems with ideas:
Sure you can do it. You can put a "reply to all" button on keyboards, but that doesn't make it a good idea. Before adding the page, make sure people will realistically use the technology.
So you're going to immerse users in your beta world. Make sure you have users' interests at heart. Users might not want use all the features in your world, so don't make them jump through hoops there for someone else.
These ideas are made by people who spent too much time next to the kool aid bowl. They use the latest acronyms and buzzwords, and usually "leverage" the companies latest pet project.
Far from mutually exclusive, these usually go hand-in-hand. These problems aren't unique to large companies, either; even startups need to keep it in mind. If your product isn't the useful, shows copious ads with minimal functionality, or spends more time touting web 2.0, ajax, and xml, you've lost focus on the most important thing: users.