If working with a large data set, the convert function in the where clause will slow things down as compared to having a datetime column. It may be beneficial to import first the data into a temp table, especially if you are using a linked server connection, and do the conversion during the insert. Then use the temp table as your base recordset.

Also keep the yyyy-mm-dd standard. If you switch to mm-dd-yyyy it can cause problems depending on your default language. (In Europe they interpret as dd-mm-yyyy).