* Above all. Remember. It's just a website.
* Understand who your audience is.
* Understand who you want your audience to be.
* Research and write personas.
* Constantly refer to your personas. Know them better than your colleagues.
* Know your colleagues. Be visible. Evangilising your ideas is your number one goal.
* Constantly talk, survey, analyse feedback from your audience.
* Listen openly to your colleagues ideas (then ignore them, the customer is your main concern.)
* Find designers and engineers who share your passion (and it has to be passion. There's no place for politics in steering successful products.)
* Don't listen to cynics. There's enough of those to stop us all doing anything.
* Don't get bogged down in requirements that nobody will ever read. Make sure at least one of your counterparts in  design and engineering understand what you want.
* Work on the solution in a small group. Create quick mocks to visualise your idea. (It's ok it's not your idea. It's not about you.)
* Evangilise, evangilise, evangilise.
* Create a prototype for your idea as quickly as possible. (It doesn't matter how.)
* Test. Take your idea on to the street if needs be. Ask people outside your organisation what they think.
* Colour IS important. everything is. You probably want people to use this product frequently, if not every day. This stuff is important to your CUSTOMERS.
* Listen to what feedback you get. Iterate your product accordingly.
* Guess what? It's almost ready to release....so RELEASE.
* What's the usage? Monitor feedback.
* Give it a bit of time.
* Is this thing losing you money? Not making you money? If it is, it's back to the drawing board. (Its code and .jpg's, not your child. Be a big boy/girl. End it.)
* People like it? Makes you money? You did good. Start again.
* Worried about innovation? Don't be. Innovation isn't a ready formula in my book, it's a journey (in most cases a very long one.)