Embed from Getty Images
Good set of data standards lowers data quality issues for your company. Keep these characteristics in mind when you establishing or revising a set of data standards.
Recently one of my client agrees to revamp their master data as a preparation to implement an ERP system. It is a newly found, about a year old, small trading company with only about 100 SKU of products. As I reviews their current master data and proposing a new set of data standards, I found that they suffer the similar challenges as large corporation in relation with data standards, only with different scale.
Sadly, data standards, especially master data, are always an afterthought topic when starting a company. They mainly being forced by technical requirements of the system to support their main operation. Establishing data standards always come to play only when there is a need for a new or upgrade system.
Why do we need to care about data standards? Well, because it directly impacts the quality of data generated for and from your business operations. It is obvious that data quality issues can cause a company lots of unneccessary expenses, missing opportunities, or at a minimum, productivity losses due to rework. A set of proper data standards being follows by the users can significantly reduced, if not completely eliminated, the chance of data quality issue.
Good Data Standards(?)
So how do we define “good” data standards? This is actually subjective. I would like to refer to it as ‘proper’ data standards as it depends on the business context. One indication of a good implementation of data standards is users use it as part of their normal business operation. Not just once every year when they do data clean-up. Of course, there are always some complaints about some standards, or some exceptions here and there, but in general it is being followed and the data quality issue is in-control.
There are 3 characteristics that would help driving good data standards. Being practical, clear linkage to business, and prepare to evolve.
There is no use if you develop a set of data standards that so detail and so perfect thus nobody can follow it during their normal operation. Users will abandon the data standard very soon, they will cut the corner, tricks the system just to be able to create a new customer right now, or change a product description quickly. Keep the strict rules of your data standards as many as needed but as few as possible. In the mean time, there is no hurt to have a set of guidelines or best practices for the users to follow, as long as it doesn’t slow down the business operation flow.
The content detail of the data standards is not as important as the consistency that it is being applied.
Clear linkage to Business
Don’t just set the data standards for the sake of having standards. Every standards must support some kind of the business vision, mission, or strategy. It is always a good idea to plan for the future, if you understand where the business is heading to.
But don’t overdo it, either. You don’t need to force an 8 digits product code for a company with only hundreds of SKUs to prepare for the future growth. Most of the time those 4-5 leading zeros will causes the users to curse you and it will be eliminated when loaded to Excel anyway.
Establishing a good business case when provide training of the standards to users will also help the adoption. Most people are willing to follow the standards if it clear to them how it is going to help the business. Knowing that dummy values in product dimension will cause problem in logistic department will help the data encoding person to try harder to fill the real values.
Prepare to Evolve
Good data standards has to be able to evolve along when business changed. Exception handling processes should be explicitely documented so that users have some way out if they stuck on the process. Transition strategy from legacy set of data standards should also be thought through and executing accordingly. Setting periodic revision cycle in advance provides venue to revise and make data standards up-to-date. This could be one every one, two or three years based on your business pace.
Data standards development is usually an underappreciated work. But it is important nonetheless. The benefit of having good set of good data standards is huge on releasing organization capacity on dealing with data quality issues downstream and become a strong foundation for further business growth.