User Acceptance Testing is often a practice to help reduce the chance of rejection of your product. It can be a great thing, if anything happens because of it.
Ideally, UAT happens throughout the release, so there are not huge surprises at the end that there is no time to fix.
Who are U? Are we using representative users to do user acceptance testing? If not, how do we make sure the proxy users represent real users well? Have they used this kind of product in the past? Are they in contact with current users, so common issues and abilities are well understood? If we are just having people trying to follow a test protocol without insight into real user needs, we might miss opportunities to make the product great.
And if if we're not doing UAT at all, heaven help us...
