Analytics governance might not seem exciting, but it can improve innovation and mitigate risks. It's also critical to responsible data management and analytics practices.
Most data and AI teams are about as enthusiastic about governance as they are about filing their tax returns. But as the complexity and scale of data and AI initiatives grow, so do the associated risks -- you need data and analytics governance policies.
IT teams often think of data and analytics governance as rigid rules and processes that hinder agility and innovation. The reality is quite the opposite. Strong governance provides a framework that enables your organization to explore new opportunities confidently and collaborate effectively. When designed and executed well, governance enables innovation, rather than constraining it.
It's like when my granddaughter would visit when she was a toddler. We would make the room child safe: Put breakables away and turn sharp furniture corners to the wall or edge them with bumpers. Then, the little whirlwind was free to explore and play to her heart's content because the environment was safe. Your data analysts may not be toddlers, but governance allows them to experiment, explore and develop with minimum constraint because the data landscape is safe.
Data governance alone is not enough. Analytics governance is equally essential, especially when working with data science, machine learning and AI.
Analytics governance vs. data governance
Analytics governance is a framework that establishes policies, standards and processes to ensure the responsible, ethical and compliant use of data in analytics initiatives. It is the oversight of data for specific purposes. Analytics governance ensures data insights are technically sound, and that they are ethically and legally permissible. It goes beyond the scope of traditional data governance, which focuses on broad policies for managing and securing data assets. Data governance lays the foundation for data quality, integrity and availability, whereas analytics governance specifically addresses how organizations can use and apply that data in analytics projects.
Data governance lays the foundation for data quality, integrity and availability, whereas analytics governance specifically addresses how organizations can use and apply that data in analytics projects.
For example, an online retailer that collects customer data as part of its operations should already have a data governance policy that ensures the secure storage and management of that data. Analytics governance establishes guidelines for how the retailer can use that data in analytics projects. Customers might allow the retailer to use their data to optimize operations, but they might not want their data used for marketing or product development. You have probably made this choice online about how different sites use your internet cookies.
Analytics governance might sound like a constraint, but it is an enabling function when well managed. Analytics governance can mitigate the risks associated with data misuse or unethical practices, and it can help grow trust and confidence among stakeholders. It helps individual analysts navigate the complex landscape of data privacy, consent and compliance while supporting analytics' power to drive innovation and growth.
Analytics governances challenges
Some problems are specific to analytics. For example, it's common to find a proliferation of duplicate reports, dashboards, visualizations and models, all built by different users or teams for similar purposes. One reason for this proliferation is the decentralized nature of analytics initiatives.
Business units or teams often undertake analytics projects using various tools and platforms. This siloed approach can result in multiple versions of similar reports or dashboards, each created independently without awareness of existing reports. A lack of standardization and coordination exacerbates the problem, leading to a fragmented and inefficient analytics landscape in your organization.
Moreover, the rapid pace of technology adoption and the emergence of new analytics tools contribute to the proliferation of artifacts. Teams experimenting with different platforms and tools might create additional versions of reports or models, further compounding the issue. Without proper governance and oversight, uncontrolled growth of analytics artifacts can lead to duplicated efforts, inconsistent results and wasted resources.
When multiple teams create similar analyses, they duplicate work and increase the risk of inconsistent or conflicting insights. This can erode trust in the analytics process and hinder effective decision-making. Furthermore, the proliferation of artifacts strains organizational resources, including data storage, computing power and human capital, leading to inefficiencies and increased costs.
Analytics governance platform capabilities
Most major analytics vendors -- such as Microsoft, Oracle, Qlik, Salesforce and SAP -- support analytics hubs for you to collect and organize reports, dashboards and visualizations. But generally, the hubs support only the vendor's artifacts and have limited opportunities for integration across silos.
Analytics portals have emerged as a more integrative approach for complex environments. Vendors such as Digital Hive, Metric Insights and ZenOptics have developed tools in this market. A good analytics portal enables several features:
A single pane of glass for analytics assets. Analytics portals and catalogs can provide a one-stop shop for organizing and discovering analytics content, even when distributed across your infrastructure. Teams can manage reports from one system and visualizations from another. They can also use a centralized dashboard with a folder structure that reflects their workflow and priorities, regardless of where the original artifacts reside.
Manage access and permissions. Analytics portals and catalogs might include features for managing user access and permissions across the enterprise landscape, ensuring that only authorized users can view or interact with specific analytics assets for specific purposes.
Collaboration. Analytics portals and catalogs facilitate collaboration among teams by providing features for sharing and commenting on analytics content.
Monitor usage and performance. Analytics portals and catalogs often include capabilities to track usage metrics, such as views, downloads and user engagement. These measurements provide insights into the adoption and performance of analytics assets.
Some teams successfully created analytics catalogs by extending the functionality of data catalogs. The approach is less successful at handling the profusion of reports and visualizations that might be built over a single data set, even though the data is well-governed.
What about AI?
As AI technologies become more prevalent and integrated into business and society, the demand for effective AI governance will grow. AI governance will have unique demands, such as tackling algorithmic bias, fairness and explainability. In practice, AI governance should be built upon a solid foundation of analytics governance. A well-established analytics governance framework helps your organization ensure that its data and analytics practices are reliable, trustworthy and compliant. This foundation provides a strong starting point for addressing AI's challenges and risks.
How to implement analytics governance
Analytics governance projects typically start by assembling a virtual team that includes IT, data management, business units and compliance. This cross-functional team oversees the governance process and makes critical decisions related to analytics initiatives, including the choice of tools and platforms.
The next step is to establish clear goals, including regulatory compliance, improved collaboration and deduplication of effort. When setting these goals, it's crucial to consider the organization's overall strategy and priorities. The governance board should also consider security guidelines and expectations from data governance to ensure regulatory compliance for data use, protect sensitive information and maintain stakeholder trust.
The discussion and evaluation of analytics governance tools should happen early, typically after establishing the analytics governance board and defining projects and goals. Aside from features and functions, it's critical to consider integration with all the analytics tools you currently use. Tools that you cannot govern are risky silos, so there should be as few as possible. You might find legacy tools that you must integrate with a modern framework. You might have to govern those tools with explicit and careful policies, and you will likely want to migrate off that platform as soon as possible.
Donald Farmer is principal of TreeHive Strategy and advises software vendors, enterprises and investors on data and advanced analytics strategies. He has worked on some of the leading data technologies in the market and previously led design and innovation teams at Microsoft and Qlik.
