A large US financial institution wanted to test categorization accuracy rates from multiple vendors, including MX. To do this, the institution ran 5 million transactions through each vendor’s system and then analyzed the results.

Click here to download a PDF version of this document.

The results showed that MX reached categorization accuracy rates of over 90 percent, which was 30 percentage points higher than the closest competitor. This means the closest competitor’s users have to manually categorize 4 out of every 10 transactions whereas MX users only have to review 1 out of every 10 transactions — many of which are transactions that cannot be categorized such as cash withdrawals and checks.


With our proprietary user-learning algorithms, MX results reach even higher than 91% after deployment. To learn more about the MX categorization process, read “How MX Cleanses, Categorizes, and Classifies Transaction Data.”