Here are some examples of how 5S applies to data quality and data management:
Sort: During the sort phase, the focus is on separating the required from the unnecessary. In data terms this can be applied to the identification of data items that can be archived for example or metadata that is no longer used in the organisation.
During data migrations it’s common to find up to 40% of historical data in a system that is no longer relevant. There is a vast amount of surplus information held in operational systems and if retained simply adds to the operational and capital burden of the company.
Straighten: This step ensures that everything has an appropriate place. How many times have you asked for a data model for example only to be given 4 or 5 versions from different team members. The same can be applied to data definitions.
During Straighten you’re looking to create definitive locations for all of your data management real estate.
Scrub (or Shine): Historically in manufacturing this referred to the physical cleansing of the workplace but in the world of data it can refer to a variety of different actions. There is the obvious act of cleansing data to improve defective values but in traditional quality circles there is also the importance of looking for ways to keep the data clean so obviously prevention plays a key role here too. In traditional manufacturing, processes are adapted to prevent the workplace becoming dangerous through spills and the same process can be applied to data.
Standardise: The key here is consistency. If one worker from another team should enter the workplace of a co-worker they should be able to navigate the folder system to find the appropriate metadata or data quality rule for example. How often do we work for organisations where there is a total lack of standards around naming conventions or data defect escalation for example. Standardisation is one of the most critical steps of 5S when applied to data.
Sustain: The final step is self-explanatory but nonetheless critical. Processes must be adopted to ensure that 5S is woven into the fabric of everyday working life. A key element of this stage is visual reporting. By highlighting the performance of the team and creating total transparency then workers are more likely to pull together and preserve the standards they’ve worked hard to maintain.
Connection with Visual Management
Another key aspect of 5S is that it enables organisations to apply Visual Management to the workplace and this is key to data quality improvement.
In past interviews on Data Quality Pro I kept discovering that successful data quality leaders repeatedly cited the importance of having visual progress reports and impact alerts at prime locations within the organisation. This allowed stakeholders to instantly observe the impact the data quality initiatives were having. It also helped foster greater community and collaboration too.
In traditional quality initiatives Visual Management goes beyond basic reporting to include tactics such as appropriate labelling and other safety conscious visual cues for workers. Again, I’ve witnessed data quality leaders applying similar initiatives by giving workers access to basic colour coded reports and charts to denote the health of data. In past initiatives of my own I’ve used visual cues to help new recruits navigate the various data management procedures they need to follow.
Modern data profiling tools for example make great use of Visual Management to guide workers to defective data one minute and then help stakeholders observe the financial impact of data quality the next.
Data Quality Management can often be perceived as an alien tactic for organisations new to the discipline so by extending traditional quality tactics such as 5S and Visual Management it makes it easier for organisations to embrace and sustain data improvement initiatives.
A vital aspect of 5S of course is that whilst it creates company-wide benefits it is actually a very local initiative. Teams take control of their data quality and come together to create a better working environment. In my experience this ability to help workers become self-sufficient in data quality has proven immensely beneficial to sustaining data quality over the long-term. You can’t control data quality centrally forever, you have to create a “hub and spoke” framework so that distributed teams solve their own battles but draw on central resources where necessary.
What about your data quality initiatives? Have you had success with quality principles such as 5S and Visual Management when applied to data? What has worked well for you? Why not share your experiences below in the comments section.
This is a guest post from Dylan Jones as part of our X88 Insight Series where we invite industry experts to share their views on Data Quality, Data Migration and Data Governance. Dylan is the founder and editor of Data Quality Pro and Data Migration Pro.