Structural analytics: The overlooked mode of analytics that dominates banking
The banking sector is undergoing a profound transformation as organisations try to modernise their data and analytics capabilities. Significant strides have been made in digitisation, cloud adoption, and the deployment of AI and machine learning. Yet modernising structural analytics — the analytical backbone of Risk and Finance functions — remains a formidable and often overlooked challenge.
What is structural analytics?
Analytics is often categorised by business domain or methodological approach. But a more helpful lens is to consider three foundational modes :
Exploratory analytics is focused on uncovering patterns, trends, and relationships in data. It’s open-ended, iterative, and human-led. Analysts need flexibility to experiment, pivot, and test assumptions — which calls for tools optimised for rapid prototyping, data manipulation, and visualisation.
Generative analytics uses machine learning and AI to produce novel outputs from data inputs. These workflows prioritise scalability, automation, and easy deployment, often working with large and unstructured data sets.
Structural analytics applies predefined, deterministic transformations to generate outputs with fixed meaning— such as regulatory capital, liquidity metrics, credit decisions, and stress testing results. In these contexts, control, transparency, and auditability are the top priorities.
Why structural analytics is different — and difficult
Banking is an outlier among industries: structural analytics is central to how institutions run and report, largely due to regulation. The problem is that most enterprise platforms are built for exploratory and generative analytics – i.e. “data science”.. These environments optimise for flexibility, speed, and scalability — not repeatability, traceability, or model control. As a result, they often fall short in heavily governed domains.
Lacking proven alternatives, many CROs and CFOs stuck between:
🧩 Retrofitting data science platforms to support structural analytics
🏗️ Lifting legacy applications into the cloud with minimal transformation
Unfortunately, neither path consistently delivers the control or efficiency seen in other parts of the bank.
