Companies and organisations are becoming increasingly dependent on data to operate its business, and therefore we need to talk more about the practical challenges in how to set up a well-functioning data governance.
There are two primary drivers for data governance, regulatory compliance and monetising on data. Regulatory compliance has been a prevalent driver for data governance, which explains why highly regulated industry sectors, e.g. financial and pharma companies, have come farthest in their commitment to and establishment of data governance.
The increasing complexity of the regulatory landscape will push today less regulated industry sectors to think more seriously about data governance. However, companies with strategic imperatives to monetise on data also need to ensure to guard their data assets from a commercial and operation standpoint to assure customer satisfaction, business performance and to manage business risk.
All companies and organisations have some type of enterprise governance structure and model in place, with the objective to oversee and properly manage the strategic, tactical, and operational risk of the enterprise and its tangible assets. Data governance is not much different, even if data is an intangible asset. It is about making informed decisions about how to manage the company’s data assets and ensure efficient utilisation of trusted data for the intended purpose of use.
Data governance is an organisation wide effort
There is common fallacy that the IT unit is best suited to drive and take care of data governance, which may be grounded in the reality that many digital transformations start with migration to the cloud followed by investment in sophisticated data platforms, master data systems, data catalogues etc.
Technology and tools are very important enablers for good data governance, but it is only one element of data governance. For example, good data governance may contribute to minimise data management costs such as data storage, data processing and operational cost, a common cost burden on IT.
Another fallacy about data governance is that it is about exercising authority and control based on defined policies and processes. That may be one reason why some companies perceive that they have failed to establish data governance. Demand and control is not the recipe for effective data governance. Data governance is company holistic continuous activity that shall effectively address and remove operational, tactical and strategic obstacles to derive the expected value from data and/or assure compliance. That requires all functions and levels of the company to take ownership based on clear roles and responsibilities in the governance structure, which is very well explained by Gartner.
Let me share an example from my own experience to show that data governance is an organisation wide effort.
Data governance – use case
The global company had an operational model with a global sourcing unit managing and autonomous procurement offices in the local markets. Corporate sourcing signed frame agreements with globally present vendors to achieve economies of scale in pricing for the whole company, with the aim to support the local markets to keep their spending down. Limited information was available to follow up prices in the local market vs the global frameworks. My unit was asked to analyse procurement data to look for further opportunities to support the local market to benefit from the global frame agreements. We felt confident that this analysis would be easy, considering that we had all the company-wide procurement data available in the central data warehouse. What we found was interesting from several aspects.
We started small, with analysing one product group that we knew was procured in high volumes in all markets. The outcome showed a large span of difference in prices between the markets, both well over as well under the price spans in frame agreement. This showed a potential for reduction in spend by better utilisation of the frame agreement but also a potential to lower the negotiated price in the frame agreement. We gained great interest from the category manager, who clarified that the product needed to be broken down into sub-categories to be fully comparable on the negotiated price level. Even then it was easier said than done to deepen our analysis, since the level of details data needed was not available in the procurement data.
The root cause pointed at the procurement process which was not designed to capture these details. The resolution to this challenge would entail a redesign of the procurement process, revised data/information models, updated master data, configuration of source system and change in work instruction for the people working in the local procurement offices. The benefits from these required changes had to be weighed against the additional cost for changes to the procurement process. This case shows the complexity of data governance, where a data quality and availability issue may become a balancing act between multiple stakeholders, their goals, and budgets.
Design a data governance models that fits the company and organisation
Most companies and organisations undergoing a data transformation journey understand the concept and importance of data governance, at least at some levels of the company, but often struggle on the practical level to scale and establish a well functional data governance in place.
There is no lack of comprehensive and well defined data governance frameworks out there (e.g DAMA, CMMI, ARMA). However, going from concept to reality has never been easy particularly when it requires cross functional collaboration and alignment, which is as concluded a prerequisite for a well working data governance.
As was shown in the use case good data governance in practice, usually comes down to the noteworthy balance among conflicting needs and goals, which arise in various areas for many reasons.
Despite all detailed data governance framework and concepts, unfortunately there is no one-sise fit all for data governance model.
Depending on the company’s degree of dependency to data to operate its business, data governance will become a smaller or larger element of the enterprise governance. A well-functioning data governance will become a well integrated part of the enterprise governance model.
When introducing a data governance structure and model it is imperative to understand the current structural capital, e.g. operational model, financial model, organisational structure, core processes, culture, and how these will be impacted by treating data as an asset. Does the company have centralised, decentralised or a hybrid operational model and how will this impact the design of the data governance model? Is it realistic to enforce a centralised data governance model in a decentralised company culture and vice versa? What type of data does the company work with? What data is of common value (e.g. master data) for the whole enterprise and what data is more functional specific data, and what it means for the operational model and data governance model? Shall new roles be introduced or can the data stewardship responsibilities become an additional dimension of an existing role, do we need to set up new forums or can we effectively add stewardship items on the agenda in existing ones? And so on.
The industry sector, company size and legacy will make the challenge look different for every company and organisation. Large organisation = larger complexity. It all comes down to the management and control of what data an organisation has, where it’s being stored, what information is confidential or trade secrets and how it’s being used.
In response to above mentioned challenges there are concepts like data mesh and non-invasive data governance that addresses the question from an operational and technical perspective.
Let’s continue the conversation about how to design a model that fits your company and how to successfully go from concept to reality.
