Planning for Data Quality with Limited Resources
Jan 28

Planning for Data Quality with Limited Resources

Whose job description includes “ensure that the data you use and record is of maximum quality”? Not many. However, it’s an absolute requirement of pretty much anybody in the modern workplace. Data is king. Good data can inform business decisions that bring immense success. But, equally, bad data can cripple a business.

So, we know it’s important but we will rarely have time to stop our daily job and work on improving the quality of data we already hold.

We need tweaks to our processes that maximise quality as we go along.

Assign Ownership

Who’s responsible for the data that needs to be cleansed? If it’s “everybody” then nobody will take responsibility. Ensure that for each piece of data your whole team knows who, specifically, is responsible for its quality.

Think what it’s needed for and by whom

If a set of data has only a single use or few users then it’s easy to scope the data quality requirements. If it’s for general use by many users then you need to think harder about accuracy, completeness, period of validity, etc.

Understand limits to quality of existing data

If you’ve inherited a set of really poor data then consider the effort needed to tidy it up versus the possible benefits. Maybe it’s not worth it but you should keep in mind known inaccuracies, omissions when you use it. Qualify any results/decisions derived from poor data.

Periodically review

Make time to quickly review each data set occasionally. What might have been fairly current or valid at one time may no longer be fit for purpose. If it’s not much use now then consider deleting it. Everybody’s uncomfortable about deleting old data but is it really any use and do you have a legal responsibility to delete it? Also don’t forget to consider GDPR legislation

Use appropriate tools

Reviewing data in a system might be difficult in-situ so consider exporting it temporarily for review in a tool such as a spreadsheet. A filtered, sorted, tabulated table is a much easier environment to spot inconsistencies and omissions. Just don’t leave the data in there and, once reviewed, don’t leave the spreadsheet copy lying around in a folder. That’s a nightmare for data security and compliance!

Related Topics

About The Author

Gary Rooksby has over 25 years’ experience implementing and evolving corporate systems including manufacturing and quality systems for a range of major clients such as the MOD. For the last 18 years Gary has specialised in Sample Management Software with emphasis on process optimisation and data management. Gary works in partnership with clients and draws on his wealth of experience to help institutes and their teams to maximise the benefits from new and upgraded systems. Business needs are constantly evolving, and Gary loves the changing challenges. Gary always focuses on delivering value to the users, whether that is financial, scientific or simply easing workloads. He believes that the system is never an end in itself; it is a tool to help the users achieve their goals and that principle is always at the heart of any system or data designs.