July 4, 2024

It’s not (just) about Match Rates: Clarity Through Enrichment

Clarity through enrichment

4 minutes

In a previous article, we discussed the importance of match rates and why so many banking institutions rely on this metric to determine how well their transaction enrichment solutions are working for them and their customers.

As a quick refresher, Match rates refer to a positive result from the transaction enrichment solution that correlates a transaction to a specific merchant. Moreover, as bank transactions are always messy and always require finessing to find the correct merchant information, no transaction enrichment solution matches and enriches 100% of the transactions.

In this article we will consider how enrichment is the better measurement of success when thinking about financial transactions. 

Consider the following example where, for a particular set of data, the match rate (as previously defined) is 75%. This does not mean that a properly functioning transaction enrichment system will simply ignore the remaining 25% of transactions. There are numerous techniques to resolve those unmatched transactions. If the MCC code is present, that is often a backup method for categorising the transaction. A well designed solution will insert an intuitive category icon (i.e., shopping, grocery, travel). The categorisation alone is often the difference between the consumer recognising or not recognising the transaction.

Process of enriching the initial raw data
Process of enriching the initial raw data

An even simpler approach would be to apply basic cleanup processing of the merchant description; e.g., ‘BOBS INSURANCE 311121189’ would be changed to ‘Bob’s Insurance’ which is immediately more recognisable to the consumer. These techniques, among many more, are all helpful and would occur before any advanced AI tools were introduced to clean the transactions. We will delve into the role that Machine Learning plays in enrichment in a future article.

Accuracy Over Match Rates

One final aspect of looking at match rates is that it is important to understand that it isn’t always about logos. Another example to consider is where the consumer would see the following logos next to a transaction in their banking app.

Logo of the local pharmacy
Example of a local pharmacy logo
Standardised Green Cross logo for pharmacies

Example of a standardised pharmacy logo

Savvy banking institutions will realise that this icon does not necessarily provide the accuracy or detail useful for the user to understand that transaction. For a large pharmacy chain, the logo of the pharmacy is useful to display but many of the transactions will be for small, independent pharmacies or pharmacies that are embedded in larger chains. Conversely, some regions will standardise this cross such that consumers do not know the logo of the local pharmacy but just the green cross logo alone. So again, the method of enriching the transaction data is more important than the match rate itself. This is routinely referred to as the accuracy of the merchant matching and is paramount in designing the best consumer experience.

Conclusion

In summary, the goal is to enrich the transaction, not just match the transaction to merchant information. When done correctly this will enable banking institutions to help consumers identify transactions more easily. This means thinking beyond the simple metrics of “match rates” or “how many logos are in the database”, and solving for the real problem of consumer clarity of their transactions.