DQ Jumpstart Case Study

Dan Myers and Mitchell Walter

DQMatters’ Jumpstart Class Provides Actionable Tools for Services Australia

In June 2019, DQMatters held our DQ Jumpstart in-person class in Brisbane, Australia. At that time, six of the Australian Government’s Services Australia staff attended our class to find better ways to document and communicate data quality. The following is an interview with Mitchell Walter (LinkedIn Profile) who is the Director Data Quality Management in the Chief Data Officer Division of Services Australia.

DQMatters: Mitchell, can you tell us what motivated your team of six people to attend initially, and then decide to send more people this February?

Mitchell Walter: Services Australia has over 25,000 employees and provides billions of AU$ per year in payments to Australian citizens. As you can imagine we have a lot of data, so one reason we attended was to identify a refined and standardized data quality language to support our efforts including new data management projects we were starting in mid-2019. The initial team that came through the training in June 2019 were preparing to pilot our Data Quality Framework. We know improving data literacy is important to success and we wanted to get everyone on the same page for the start of the pilot. That training was very positively received and we were keen to get our newer team members along to the course.

DQMatters: Which parts of the DQ Jumpstart were most helpful and how were they used by your agency?

Mitchell Walter: The detailed explanation of each definition of the Conformed Dimensions of Data Quality with Underlying Concepts provided us with language to explain existing DQ issues. Since attending the class, we have  started piloting data profiling tools to benchmark data quality and worked towards improving enterprise-wide understanding of the agency’s DQ issues. The detailed definitions of the CDDQ are supporting us to develop and roll out these capabilities. It helps to make DQ more actionable because it leads you to more quickly identify how you go about measuring DQ for different types of data.

DQMatters: How would you describe your organizational journey towards DQ improvement?

Mitchell Walter: Like many organizations, we have had lots of business rules built into our core systems to aid the acquisition of quality data. We have also had teams responsible for monitoring and remediating specific types of issues, but the approach was not consistent. DQ profiling tools to investigate Completeness and Consistency dimensions have been applied in the past. However, this was more commonly done in context of a data migration project rather than monitoring DQ for its own sake. Use of the techniques outlined in the DQ Jumpstart has provided us with a menu of options that can be applied to proactively measure DQ beyond these useful, but somewhat simplistic dimensions. It is assisting us to work towards a portfolio of reusable DQ metrics based on business rules as we implement Validity and Timeliness related controls. Our goal is to achieve the data quality Holy Grail of Accuracy (compared with other sources of data and correct representation of real-world concepts). Going forward, we plan to use the CDDQ Example Metrics and customize them for our internal customers.

DQMatters: What are your organizational DQ priorities for the next 6 to 12 months?

Mitchell Walter: I’ve run DQ projects before, and this time around we’ve used the definitions of the CDDQ to articulate DQ expectations in a sustainable way, making DQ “Lifestyle,” or journey, not just a once and done project. While implementing data profiling tools to monitor and manage DQ is an important part of our DQ maturity over the coming year, our overall success will also be dependent on increasing data literacy.  An engaged network of data stewards using a common language relating to DQ will help underpin consistent enterprise-wide measurement of data quality.

DQMatters: Thank you for your time Mitchell. In closing, do you have any advice for other organizations beginning their DQ journey or considering DQ training like the DQMatters DQ Jumpstart?

Mitchell Walter: I certainly encourage people to look broadly across industries to identify successful practices. DQ Jumpstart is a modest investment in terms of time and cost that will help you look beyond your current organization. Most of us are facing the same sorts of challenges managing data. When you participate in this type of training not only will you better understand how to measure DQ, and apply quality practices, you will likely meet others that are on the same journey as you. In a large organization this work should progress as a part of an overarching strategy ensuring other core capabilities such as Data Governance and Metadata Management are effective in supporting your efforts. Even if they are not, you can still achieve value, as the old saying goes- 'you don’t need to try and boil the ocean'. Start small, focusing on high value data, or areas you already know need improvement, and build out from there.

Find out more about what DQMatters covers in the DQJumpstart and register here.