Scenario 4 - Fuzzy Matching

"Incoming_leads_dirty.csv" contains leads. Despite this, we need to know if any of them are existing customers.

"crm_accounts_clean.csv" contains all customers from our CRM system.

If we perform a "Merge another table" action (like the previous scenario) joining on the company names, we get a small number of matches. The problem is the company names were manually typed and might contain errors and typos. 

If we add a "Match" action in "Fuzzy" mode before the Merge another table action, to map the company names first, then the merge will match many more and we can filter for just the ones that aren't existing customers.

Setting the match distance in the Match action is important. Too low and we'll miss matches. Too high and we'll get unintentional matches. For this data, a match distance of 5 is about right and leaves us 75 records which aren't existing customers.