Photo: Data Center, by Bob Mical, CC license
Early in my career I couldn’t stand anyone talking to me about master data. It just wasn’t interesting, and seemed to be just a minor detail in how systems run.
I have since learned that (incorrect or inconsistent) master data can stop in its tracks the coolest, most sophisticated and expensive enterprise software anywhere, which means it can stop all kinds of transactions in your business.
Master data is data that should have a consistent definition and format across your organization.
For example, a ‘finished good’ (product ready to sell) in a manufacturing company should have the same spelling, Universal Product Code (UPC) and internal product number whether it is listed in a sales report or a manufacturing report.
A gallon of Sherwin-Williams paint, for example, might have the following pieces of master data:
|DESCRIPTION||‘Sherwin-Williams Fine White Eggshell 1 gal.’|
|UNIT OF MEASURE||pounds|
For applications to exchange data with one another, regardless of method, they have to use the same definitions of data. If the sales application uses “lbs.” as part of the description, and the finance application uses “pounds,” some type of translation (more software code) will have to be built between the two applications.
If your master data is not standard across your enterprise, the extra code necessary to translate incongruous data will add to the complexity and cost of integration. These custom integration algorithms will be specific to your enterprise and therefore make your whole application structure more difficult for a IT new hire or a contractor to understand.
To further control costs and complexity of integration:
- Use a hub structure if multiple applications need the same data. This could be an “exchange” where pieces of data are posted then retrieved when needed by applications plugged into the exchange. This avoids creating spaghetti (your application maps look like spaghetti).
- Document everything clearly and insist that all documentation be kept current. Otherwise, you will end up with a complicated setup that is mission critical to your business but that only one person is familiar with.
- Let your ERP system define your master data standards, and keep those standards consistent in all systems. If you have an ERP system, most likely it required you to define and set up master data. Since ERP systems usually span multiple functions, the master data setup is shared among functions and therefore common to departments like sales, finance and manufacturing. Keep these standards in place and use them to define the data transfer with any new application.