Given the well-known bias humans have against change - to be fair, there are good reason for this bias - I think it's important to give a design a good while of use before evaluating it against the predecessor, and only using objective criteria to evaluate the successor. Any form of 'it looks worse' is not useful.