When a database is set to be compatible with SQL Server 2011, the behavior of the string-length and substring functions changes ...

When a database is set to be compatible with SQL Server 2011, the behavior of the string-length and substring functions changes when dealing with Unicode supplementary characters. Each Unicode supplementary character, which is defined by having a code point larger than U+FFFF, is counted as one character by these functions rather than two, as it had been in previous versions.