We use Optimizely for all of our A/B webpage testing across both our main website and any campaign landing pages. Only the Marketing department uses this tool. We use Optimizely to test out theories in order to see their effect on conversion rates and other relevant KPI's.
We went with Optimizely over Lift because it was more flexible with creating variants outside of the "blocks" of content that Lift handles. So for instance, if you wanted to test an entirely new area of content you cannot do that with Lift. You can only test images/text in existing content areas.
Optimizely is well suited for your standard A/B testing scenarios. It does not do multivariate testing (at least the version we have does not). I would not look to Optimizely for any type of personalization for a website. There are better tools out there for that.
3 - Various marketing positions from Digital Marketing Manager to graphic designer. We usually have everything tested before its made live to ensure that there is no unexpected interruption in user flow on the website. The people who us Optimizely are almost always within the same department.
2 - You will need a designer who has a solid working knowledge of A/B testing to support it. Keep in mind that all designs must work across desktop and mobile and there are unexpected differences in the user experience that a designer can account for. You should also have a front end web developer on hand to make sure that what is being tested can be implemented if it proves to provide a lift.
Overall it it one of the best if not the best tool in the market. It has all the features we need and seems to be growing functionality so that there is a future with them as we grow our needs. The only issue could be price increases that we keep an eye on.