A recent State of Data Governance Report shows that 98 percent of surveyed organizations consider data governance important, and ~50 percent have already started their implementation. They did so either through their own employees or via an external company providing data governance consulting services.
Effective Data Governance usually spans across the whole organization, including Human Resources, Processes, and IT systems. When it comes to enterprise data governance, smaller businesses and start-ups often back off, intimidated by the size of the project: “At my level, the benefits would hardly overcome the investments needed”.
In reality, this is a false stereotype. Data Governance is more about a way of treating and thinking of your data rather than a long and time-consuming endeavor to be completed all at once. It does not require as much effort on a smaller scale, while still delivering tangible benefits. Let’s take a closer look at some key practices and their impact on your IT ecosystem.
What are the key data governance practices?
Data Governance assumes a number of practices to be applied to the way you handle data across the whole organization – both from the process and IT landscape perspectives.
From an IT point of view, data governance implies that you need to design, document, and execute the following strategies and models across your entire IT ecosystem:
- The Backbone
- Data Model
This is an integrated model defining what types of data your organization produces and consumes. This model is usually organization-wide and system-agnostic. It is further mapped to your IT landscape to define which of the systems operate each type of data. Having such a model in place can help you minimize data redundancy and safeguard data consistency and quality.
Data management practices
These are the major rules regulating how data is handled. They include strategies for data storage, organization, back-up, and preservation. It is important that the practices you apply also include reprocessing, modification/mutation, lineage, and retention policies as they are essential for preserving data quality and help to achieve all the benefits mentioned in the previous section.
Essential practices and strategies
Data access strategy
This strategy defines what types of information your organization works with and the way it is collected, stored, and retained. It also focuses on how data consumers (both users and software systems) access each type of data (either through API or direct database access). This is critical for regulatory compliance as it centralizes each user’s data, as well as ensuring system interoperability and maintainability.
Data import/export model
Data governance practices require that every integration point is thoroughly documented including its data model and nature, whether it is “push” or “pull”, what protocol and schedule it uses, and what interface it has.
This model reduces maintenance efforts and helps to safeguard data integrity, system interoperability, and substantially.
Identity Access Management (IAM)
When it comes to security and access management, IAM does the job. It defines the roles of different users (including development & maintenance teams, data stewards, and end-users) within a network and the circumstances in which they can be granted additional privileges for handling confidential data. This model also contains information on the way access audit journals and historical data access are managed.
Implementing this strategy is essential for regulatory compliance. Additionally, it has a significant impact on system maintainability.
Data Security and Privacy Model
This crucial model is used to separate classified and non-classified data - both current and historical – across all your IT systems. It is mapped onto IAM as they work together to safeguard compliance with GDPR and similar data protection regulations.
Metadata management strategy
Metadata management provides context to your information and helps both users and automated systems “read” it the right way. A metadata management strategy defines how metadata is managed, documented, controlled, and indexed, whether data cataloging is used and how your master data is governed. Having this strategy in place improves the way your IT systems work together and makes it easier to maintain them.
Data metrics control
KPIs are essential for every business process as they enable you to keep things under control and understand trends and interdependencies. Same with the data you handle – it’s important to define essential metrics and their management processes. This will help you make sure your SLAs are met, identify trends, and take corrective measures to avoid bigger issues in the future. Moreover, it safeguards data availability and system scalability.
It’s imperative to accurately document all technical decisions about data platforms and models. To promote business productivity, you should have data platform maintenance and onboarding guides in place. This substantially reduces further maintenance efforts.
What is the best way to start?
Of course, data governance implementation is an ambitious project requiring substantial effort and investments. The length and delivery time of such projects has been deemed the major bottleneck for the application of data governance principles in 2020 according to the State of Data Governance Report. We can therefore recommend that it takes place on a step-by-step basis. Start immersing the practices with the backbone and extend the procedure with the processes most essential for your business goals.
It’s extremely likely that you are already employing some of the data governance practices without realizing it. You might not necessarily call it data governance, but you probably have unspoken rules and regulations that you follow to organize and manage information. So why not go further? By implementing proper data management solutions, you can instantly improve your business processes, save time and money, and make room for valuable opportunities.
Boris Trofimov,Software Architect, Sigma Software