Chaired by technology journalist Luke Dormehl, our guest speakers included strategic data advisor, Andy Youell, and Rob Clark, customer account and sales director for UNIT-e. Putting the questions to the panel, Luke wanted to know: what is data governance and why is it essential at every institution?
The impact of silos
Andy outlined the problems that arise from the absence of data governance and the rapid increase in the volume and complexity of data, which is a result of Moore’s Law, which sees processor speeds and storage capacity doubling every 18 months. This rapidly changing environment creates huge challenges for organisations, which are greatly exacerbated by the siloed nature of systems that can’t share data and departments which often have different approaches to data management. These silos erode trust in data, especially that of other departments in an organisation, and cause duplication of data in multiple places and duplication of work; all of which gears cost into routine operations. Worst of all, organisations can feel stuck. This lack of efficiency and trust makes change difficult, costly and risky.
Rob spoke of his first-hand experience with the impact of these silos and “issues of ownership” in the education sector, and how they impact the aspirations of education organisations, the service they deliver and how they communicate with students. These impacts also highlight the importance of starting with business needs first, rather than data governance for its own sake.
Data and the pace of change
Our speakers also highlighted the rapid changes they have seen in this area in the last two decades; in our expectations of what data can do for us, mainly driven by interaction with big tech consumer platforms and their seemingly effortless deployment of data; and in the rapid professionalisation of the field and increase in data skills.
In the edtech field specifically, there has also been an increased commercialisation of the higher education and further education sectors, which has seen new data use cases and needs emerge. As Rob explained, institutions now want to be able to address “dropped baskets during an online enrolment” in much the same way as Amazon or eBay might handle an interrupted sale – with an email that attempts to complete the enrolment. Institutions are also taking advantage of historical data, for example success rates of alumni, for advertising; and the need to rank on data-driven “league tables”. There are also increasing statutory demands for data management, including to comply with the General Data Protection Regulation (GDPR).
Road to data governance
So, what is data governance? As Andy pointed out: “data governance is still evolving and can’t be pinned down to one thing. But it can be seen as including security management, compliance with regulations and quality assurance.” A more elaborate 10-point model has been fleshed out by the Data Management Association.
For his definition of data governance, Rob pointed to Section 2.2 of Higher Education Code of Governance, which calls for effective arrangements to be put in place for the management and quality assurance of data. Andy liked this emphasis on corporate governance as a model for data governance because it focused on defining points of accountability and authority. In addition, addressing data governance in this way avoided treating it as a transformation project, rather than a core operational principle.
Both panellists emphasised the scope of data that FE and HE organisations must work with. Rob explained how data can “originate in as many as 22 different systems including finance, HR, payroll, student records, estates, library, timetabling, attendance, learning analytics, unified communications, and student evaluation of teaching. All these systems have external and internal stakeholders and, in some cases, regulators.” For example, finance data must comply with financial authority regulation, HR employment laws must be followed, payroll must account to the HM Revenue and Customs, and student records must meet the requirements of the Office for Students and the Education and Skills Funding Agency.
With so many systems, there will always be duplication of data and the potential for lack of coherence and data be out of date. As Andy pointed out, user expectation is growing to derive actionable analytics and insights by “knitting together” data from these disparate systems, perhaps also with external data sources. This depends on proper oversight and quality data, and so data governance is essential to enable the creation of valuable insight from raw data held in multiple sources.
So, what are the key steps to implementing good data governance? Rob emphasised that data governance was not primarily a technology issue. The more pertinent issues were the need for clearly defined roles and responsibilities; accountability; authority to make decisions; and measurements with targets and objectives. “Soft” issues such as culture and attitudes to data were the most important areas to be addressed; and more awareness was needed of the consequences of poor data governance for students, staff and the college itself. As Andy said, organisations need to “adopt a culture of being custodians, curators, conservators of data, rather than administrators. There needs to be careful creation, like a museum or library curator, to make the most of what you’ve got.”
It’s time to get strategic about data
Andy emphasised the need to step back from the operational tasks around data, as well as that the need to approach governance is at a higher, strategic level: “We need to be more strategic and create points of accountability, and take granular, technical decisions in the context of broader strategic aims around data, which in turn support the broader strategic aims of the organisation.”
A common mistake the panellists observed in the implementation of data governance was seeing it as a project, rather than something embedded in business as usual. “If you build the processes around a project, data governance becomes the project. When the project ends all that falls away and the risk is nothing happens. Doing it properly will involve difficult decisions. There is a challenge in doing things properly, but it’s worthwhile. It requires resolve at a senior level,” said Andy.
Finally, Andy and Rob looked to the future for data governance: they saw increasing professionalisation of the field; data eventually being treated “with the same respect” as money, and with similar levels of governance; and data governance approaches, embedded in business as usual, successfully making up some of the ground where large digital transformation projects have struggled.
Download your free white paper
If you want to find out more, Andy Youell explores the topic in more detail in his new white paper, where he sets out the case for data governance and considers what a coherent and pragmatic approach looks like for FE and HE institutions.