A check digit algorithm calculates a check digit based on an original character string, such as an account number. The receiver recalculates the check digit to verify data entry accuracy. If the recalculated character string contains the correct check digit, the data is error-free and may be used. However, a character string that does not include the correct check digit indicates a transfer error, which signals that data must be re-entered and/or reverified.

When a check digit system is used, error detection and data implementation complexities and compromises are inevitable. Simple check digit systems easily understood by humans cannot detect errors with complete accuracy, unlike complex systems that use more complicated error detection algorithms.

A recommended check digit characteristic is left-padding with zeros, which never alters the original check digit and allows for the application of digits with varied and dynamic length.