I agree that raw scores have a significant potential for one judge to have more influence than another. However relative placement also has problems, as it doesn't allow a judge to say just how much one couple is better than another, or how close they are.
My personal preference is to use normalised scores, where the judges award raw scores as they see fit, and the scrutineer then makes sure that the average and spread of each judge's marks are the same
In practice it makes very little difference - at least in my experience. When I scrutineer a competition, I use the normalised scores to produce the result. However I also use the raw scores and relative placement as a way of checking the results. It is very rare that there is a difference in the result from the three methods.
Bookmarks