Effective Data Architecture Works Hand in Hand With Good Governance

Analytics and artificial intelligence initiatives are prompting organizations to take a fresh look at their data ecosystems, from technical capabilities to governance. Many leaders are recognizing that data architecture plays an essential role in ensuring data is accessible and usable — that is, can the data answer the questions it’s intended to answer? Ideally, business use cases drive architecture decisions: A sales team wants to accomplish X with its data, and the architect works backward from that outcome to develop the solution.

Organizations that have relied heavily on Software as a Service haven’t always had to worry about data architecture; many solutions handle those tasks on the back end. As organizations expand their use of analytics or develop AI models, however, they may need to implement data architectures that can support them. In other cases, architecture solves a specific pain point, such as a team spending hours every quarter consolidating spreadsheets or a modernization initiative that results in a need to adapt existing processes.

Data architecture and governance are separate but tightly coupled disciplines. Governance leads with people and processes — addressing data stewardship, data security and lifecycle management — while architecture focuses on the next step: What does the data mean, and how can it be used downstream?

Architecture is focused on infrastructure, systems integration and resilience. It ensures data processes can happen quickly and at scale. It also addresses data safety — for instance, if a data record drops, does the system automatically try again? These details ladder up to the architect’s primary goal: designing solutions that enable the organization to use data in specific ways. Architects also help organizations do more with their data. A business unit may have access to 1,000 unstructured data records. A data architect might identify that those records could be much more extensive, comprising a rich, multidimensional data set with multiple ways to model it out. 

Proper architecture, aligned to business use cases, has a lower total cost of ownership over time. Workflows are more consistent and have fewer touchpoints. Platforms are more secure and data pipelines are more resilient, so they have fewer points of failure and maintenance is reduced. Increasingly, artificial intelligence is making data architecture, modeling and engineering more efficient, so that TCO will continue to decline.

A common challenge with analytics initiatives is that leaders don’t fully understand their current (or future) state as it relates to data accessibility. A business unit may have a clear use case in mind without realizing that the data isn’t quite there. For example, different stakeholders may define the data differently, or “data heroes” may retain information that isn’t documented or broadly accessible. 

At the same time, solving these challenges doesn’t necessarily require a full-scale infrastructure upgrade. In some situations, resolving a data-sharing problem is a relatively simply fix, achieved by the right technology. What matters most is that business use cases, rather than technology for its own sake, drive the implementation.

Whether an organization wants to reduce friction in data workflows, enhance analytics or move forward with AI, data architecture is a critical piece of the puzzle. When governance and architecture are well designed and working together, organizations can build data practices that ensure quality, accessibility and usability.