Transaction Manual Matching Millions of Rows Leftover After Rules are Run

Summary:

Is there a streamlined way to match off millions of rows in Transaction Matching if they have to be "manually matched"?

Content (please ensure you mask any confidential information):

This is almost certainly more of a process question than a technical question, but we have a scenario where we have a transaction matching process set up with matching rules. Each month, we load about 10 million records to this process using a flat file (if there's a better way please let me know) and we run our matching logic after the load. The good news is that it matches about 80% of the transactions, so 8 million records are matched off. The bad news is that about 2 million records are unmatched as they do not