Ancestry uses Sorenson Laboratories (SL) for their testing. They have been testing Y-DNA since 2000. We have used them since then. They had errors initially, when the process was more manual. Usually, one marker would be off by 1. However, they reran all of the samples in 2003. The current crop of analyzers are very well automated, so there little chance of error at that stage.
A sample mix-up is one of the most common errors. However, SL started using bar coding before they sold Relative Genetics to Ancestry. the only other likely source of error is entering data to the web site. They may still do this manually, but there should be digit checks by now.
The chances of an error are so small that I would wait to see if any Rudd matches show up with further testing. I would also test cousins who branched off about 1850 or so. It is possible that there was a recent adoption in which the adoptee was not informed.