As I work on CARD.com we are considering how to leverage Merchant Classification Codes (MCC) data for a variety of purposes. It will be helpful to better understand our customers, to give our customers more information about their transactions so they can know who is using their card (for example, a few weeks ago I got a transaction from "TSI Inc." but by digging into the MCC I was able to see it was for gas and that reminded me of which gas station it was).
These codes are used on every single credit and debit card transaction to identify the nature of the business. This is helpful for accounting, tax reporting, and can even allow for some pretty interesting analysis like the Brighter Planet mastercard app that assigned a carbon footprint to debit card transactions based, in part, on MCC values.
Unfortunately, MCC values are mostly available in formats that are incomplete and hard to parse.
Sources of MCC Data
- The Wikipedia Merchant Category Code page has some information and links to other sites
- The IRS maintains a list but it lacks specific codes for many values in the 3000 range (this can be seen as a benefit in some ways, but for a database lookup table that's a problem).
- Visa maintains a pdf but it's formatting makes it a real pain to import anywhere. It appears to be the same as the IRS data.
- The USDA procurement division has a document that they use
A new source: Merchant Category codes on Github
So, I created a github repository of mcc values. The data is available as
- ODS - The Open Document Spreadsheet readable by LibreOffice, OpenOffice.org or Google Docs.
- XLS download- Usable by basically all versions of Microsoft Excel since 1997 and things like Apple's Numbers software.
- CSV - Usable by many, many other systems and easily parsed by code.
My hope is that other people will use this data, improve it (fork it), add relevant information, and make pull requests - that is, send their fixes back into the main version - so that we can all benefit by having an easily read and well maintained bit of data.
Funny mistakes in MCC codes
Some of my favorite mistakes in the underlying data:
- 4215 - Fright Forwarders, which obviously should be freight forwarders, but a typo makes it about shipping scary things.
- There are several "Planation" entries like "Seaines Planation" and "Amelia Island Planation" which seem like they should be Plantation. In the case of Seaines it also seems likely that the first word is mis-spelled and is either Caines Plantation or Sea Pines Plantation. I guess that having one code per hotel got pretty annoying to maintain after a while and this shows just one detail of how/why.