Complete, accurate, actionable and current data: the dream of every IT manager. The corporate world increasingly embraces the importance of data quality, but the optimization and above all the preservation of data quality requires discipline and planning. Short data quality optimization stints are like bad diets: they do yield short-term results but end up in a yo-yo effect in the long term.
Type2Solutions has helped businesses to optimize their data for many years. “We answer questions from businesses about data quality almost every day. That is a good thing, because it is badly needed, especially in today’s connected business environment where data problems can lead to severe issues, if not hefty fines.” says Jack de Hamer, partner at data management specialist Type2Solutions.
“ The increasing digitization of the business environment and stricter requirements of web giants such as bol.com and alibaba most certainly drive the increased awareness of data quality in the market. While master data often has to be provided only once, the quality and accuracy are frequently insufficient. More often than not there is no proper process in place to monitor and preserve data quality. ”
Where to Start
De Hamer observes that businesses often do not know where to start when it comes to data quality optimization. “A one-time cleansing exercise for a specific purpose is not a solution. There are more benefits to a long-term solution.” And by long-term solution, de Hamer means data quality monitoring software. Those solutions are able to monitor the quality of large data volumes and should be able to notify the owner of the data when exceptions occur.
De Hamer recognizes that data quality optimization is not the core business of most companies. “A manufacturer makes products, and all its processes are geared towards doing that as efficiently as possible. Data quality optimization requires a whole different set of skills.”
Data quality optimization is a continuous process. “A quick one-time cleanse will get you back right where you were in no time. Or maybe in a worse place. Data is like a living organism, it is constantly moving and evolving.
Therefore, constant monitoring and cleansing are necessary. Look at it like maintaining the woodwork on your house. Regular upkeep not only makes it look better, it also increases its life expectancy. But beyond that, it prevents much worse consequential damage like decay and leakage, all due to overdue maintenance.”
It is not easy to quantify the damage caused by ‘overdue’ maintenance of data. It largely depends on the potential consequential damage. “Costs can range from a few euros per customer for mailings that are returned as undeliverable, to thousands or even hundreds of thousands of euros from the consequences of an incorrect invoice, customs declaration or label.”
Return on investment should however not be the main focus for companies, argues de Hamer: “I think it would be better if businesses improved the quality of their data to gain a strategic advantage and as a means to maintain and extend their customer base.”
A real-world example: NedZink
NedZink is a good example of a company that has committed to optimizing the quality of their data.
A number of their larger customers had asked NedZink to exchange data through EDI. It did not take long for NedZink to realize that the quality of their master data did just not cut it for this kind of operation.
Their data was appropriate for internal use, where it was usable despite a few extra handling steps, but an external party would not accept that.
NedZink decided to go all the way and to make data quality optimization a standard step in their process. The recurrent cleansing, in combination with the Data Quality Monitor, have dramatically increased the quality of the data.
The obtained level of quality and the speed at which the project was finalized made that NedZink was ready for electronic data exchange long before their own customers were.
Want to get started?
Optimize your data quality with this simple and practical 5-step guide we put together to help you to get a grip on data quality.
Step 1. Define the kind of data to be collected and the elements within that data which should be optimized
Step 2. Define rules for each data element and automate the control mechanisms
Step 3. Assign the responsibility for the optimization of a data entity to a single person
Step 4. Automate data validation according to their definition, and keep the results
Step 5. Correct exceptions and adjust their definition if necessary