Works fine and is simple to use and makes small changes testing easier.
Almost drop to 4 stars as the instructions imply it is more complex and I wasted time trying to do all the steps listed. The whole discussion of the google tag manager is "optional".. it may provide more information but it is overkill for just doing basic A/B type experiments in a page. All you need is to install plugin, insert short code fill in experiment ID and run experiment.
Another important thing, mentioned on the cited google document pages, is that you ignore the warnings about URLs when getting the experiment ID. Don't add any google experiment code on the page, just make sure it has analytics.
I'm still using Google Content Experiments plugin as well, which is easier for full page testing (when the pages are not simple variations of each other, e.g. major different layouts, widgets, breaking a page into 2-3 pages, etc..)