It isn't one or the other. For starters, look for a paper by Stark and Contreras in the Feb 1998 ASCE JGGE. It provides an example of back-analysis of a slide in sensitive clay. The issues are peak and post-peak undrained strength, and the amount of strain (monotonic and/or cyclic) required to reduce the strength from peak to post-peak to residual/remolded. Refer also to more recent work by Boulanger and Idriss in the Nov 2006 issue.
I would be very hesitant to use U-U tests for much. Sample disturbance, drying, etc. can make the values highly uncertain. I'd prefer in situ VST and reconsolidated (SHANSEP-like) triax and DSS tests, as appropriate for the geometry of the slide, preferably including cyclic testing. VST can get you peak and remolded strength for very little $. I'm assuming from your previous posts that you are talking about high-falutin' numerical analysis, rather than small consulting projects that have to be done on the cheap. If it was the latter, the VST would be the #1 tool.