TECHNOLOGY, AI & ROBOTICS
WHY POLICY ALONE IS INSUFFICIENT
Good data management policy is a critical foundation for any new data governance program, but policy on its own is insufficient to change the way an agency manages its data, says Joah Iannotta of ABS Consulting.
Most organisations are driven to revamp or create a new enterprise data governance policy because there has been an abundance of innovation, new solutions, and problem- solving at the ‘State’ level within the organisation. These solutions can create challenges—particularly with respect to data being interoperable or sufficiently documented to withstand audit or regulatory scrutiny. For example, in the absence of an enterprise data governance program, two different grant programs designing solutions to capture applicant information may do so separately in their own silo. These two different solutions may not use the same data standards, data naming conventions, or other aspects of data management that would allow the two programs to integrate easily. As a consequence, analysing grant data from both programs for insights on important factors that could make both programs more effective can become significantly more difficult and time consuming. If in the long term the grant programs were to be merged into a single operational platform, integrating them would be significantly more difficult than if an overarching data governance policy had been implemented from the programs’ inception.
Establishing an enterprise data governance policy is a good first step to bring consistency, interoperability, and other benefits to a data ecosystem, but if the governance program stops with policy and guidance that only establish high-level principles such as “new and existing data systems must adopt standards to ensure data interoperability,” then it is not likely that an actual change would occur in organisational processes or staff behavior. Sticking with the grant example, for a data governance program to be more effective, it would need to take an active role in designing standards that could work for both grant programs, facilitating discussions and brokering an agreement on the standards for specific data elements. It should also include a compliance program with a deadline by which the standards must be adopted to help ensure that those standards were actually being implemented.
Successful data governance policy programs should integrate policy into implementation by using their compliance program as the bridge between policy principles and driving change in data management practice. Further, the accompanying compliance program should paint the picture of what implementation looks by establishing concrete expectations of what will be considered successful compliance with policy. These expectations should be published as part of a compliance policy or procedure so that the lines of business have a yardstick by which to measure their efforts. Ideally, this should focus on the changes the lines of business
34 | TOMORROW’S FM
would need to implement. For example, if a new data governance policy included a principle of strengthening data quality, the compliance document could articulate that lines of business must establish specific parameters for monitoring data quality aligned with the needs of the program they serve. An example of a data quality parameter that the compliance program could evaluate would be that grant recipient payee data will have less than 1% of missing values to help ensure timely and accurate delivery of payments. An effective compliance program should be able to envision a path to evaluate successful compliance. If it cannot do that or needs to rely on a program official’s attestation of compliance, then it is not reasonable to expect the lines of business to be able to comply with a policy directive on their own.
When data governance policy shops end their efforts with the publication of a document, they can leave program and business leads without direction on prioritization, timelines, and the necessary level of effort and resources needed for implementation. In many cases, the absence of this information can cause what was previously a willing coalition of business support for better data governance to get cold feet and hinder the passage of new data policies.
As policy is developed and socialised with an approving executive council, a draft compliance and evaluation plan should also be presented. This plan should always have a phased approach for the simple reason that it will take any agency time to change its processes, but also to help the lines of business prioritise and stage work.
A phased approach typically has several elements:
Allow for some passage of time before a policy goes into effect It is prudent to have a period between the publication of the policy and when that policy comes into effect, otherwise there can be extended periods of time in which the organisation is out of compliance with its own policy. A good rule of thumb is to have the policy become effective no sooner than six months after it is published. During that six-month period, the compliance program should provide training for staff to aid in adopting new data management processes that will allow them to comply with new policy. Training should be provided in a number of venues, including Communities of Practice for systems owners, data scientists, analysts, and (most importantly) data stewards. Training should be complemented with communication strategies, such as stories on an internal website and emails to stakeholders with links to guidance and recordings of training. The goal of the compliance
twitter.com/TomorrowsFM
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60
