MD101 Glossary

A/B Testing is a comparative method for decision-making that lets teams validate changes by comparing user outcomes across variants. It’s not just about optimisation but learning what works for users in the real world before scaling.

AI Agents are intelligent and autonomous systems that interact with users to help make accurate and quick decisions to meet the users’ goals by acting on behalf of the users. These are designed with usability in mind to reduce the complexities of tasks and improve user journeys, empowering businesses scale, adapt, and cater to user needs efficiently.

AIOps refers to an approach that applies machine learning to IT operations, with the goal of improving incident detection, root cause analysis, and automation. Viewed through a product thinking lens, it’s a capability designed to deliver continuous value by reducing operational noise, shortening downtime, and enabling intelligent decision-making across IT systems. AIOps refers to an approach that applies machine learning to IT operations, with the goal of improving incident detection, root cause analysis, and automation. Viewed through a product thinking lens, it’s a capability designed to deliver continuous value by reducing operational noise, shortening downtime, and enabling intelligent decision-making across IT systems.

API Gateway refers to the capability that acts as a unified access point for backend services, designed to make APIs secure, reliable, and easy to use. It simplifies how developers interact with distributed systems by handling routing, authentication, rate limiting, and monitoring.

API Rate Limiting is a control mechanism that manages how often clients can access API resources over time. It’s about ensuring consistent performance, protecting systems from overload or abuse, and enabling reliable experiences across different user tiers and product use cases.

Access Control means ensuring the right users have the right level of access at the right time. It impacts user trust, compliance, and operational efficiency.

Alerting is how systems proactively notify users when something goes wrong, be it with data, processes, or performance. It exists to help teams respond quickly, reduce downtime, and maintain trust in data products, without relying on constant manual monitoring.

Audit Logs primarily provide transparency, support compliance, and help diagnose issues by offering a clear trail of user and system actions. They help users understand and trust how systems are used.

AutoML refers to an abstraction layer over machine learning that empowers non-experts to build, deploy, and iterate on ML models by automating complex tasks, delivering faster insights and enabling broader AI adoption across different domains.

Automated Data Pipelines are systems that move data from source to destination reliably, on schedule, and without manual intervention. They exist to make data consistently available where it’s needed while freeing up teams to focus on building features and insights, not fixing broken flows.

Backfilling is the process of filling in missing or outdated data to restore completeness and accuracy. It helps keep data products trustworthy and usable. Especially when addressing gaps, recovering from bugs, or onboarding new sources.

Batch Processing is the method of handling large volumes of data at scheduled intervals instead of in real time. It allows systems to process data efficiently and cost-effectively, while still making it available when users need it. Batch is ideal for powering dependable data products where immediacy isn’t critical so teams can deliver value without overwhelming systems.

Business Glossary is a shared catalog of key terms and metrics used across an organisation, written in clear, accessible language. It eliminates confusion, aligns teams on what data actually means, and helps everyone speak the same language when making decisions.

Business Intelligence (BI) is about turning raw data into trusted, actionable insights. These are designed to serve diverse business users by optimising clarity, speed, and contextual relevance in decision-making. This translates to having dashboards, reports, and tools at the time of need to understand what’s happening and what to do next.

Build vs. Buy is the strategic decision of whether to create a data solution in-house or adopt a readdy-to-use external product. It's not just about cost, it’s about aligning with user needs, speed to value, long-term ownership, and the ability to differentiate through data capabilities.

A Caching Strategy ensures faster user experiences, optimised resource usage, and scalable system performance. Choosing the right approach directly impacts the reliability, responsiveness, and cost, making it a key design decision for delivering consistent performance under load.

CDC keeps data products, analytics, and services in sync with the source-of-truth systems without needing full reloads. It enables fresher insights, supports event-driven features, and reduces resource overhead.

A Cloud Data Platform is a foundational solution that simplifies how users store, access, and work with data at scale. It's strength is in abstracting infrastructure complexity, accelerate data-driven products, and enable teams to focus on delivering insights and innovation. Some examples of such platforms would include Snowflake, Databricks, DataOS, and dbt.

Composability is the ability to build flexible, modular data solutions by assembling independent parts. It enables teams to move faster, adapt to change, and deliver experiences tailored to specific and dynamic user needs: treating data capabilities as building blocks rather than fixed systems.Composability is the ability to build flexible, modular data solutions by assembling independent parts. It enables teams to move faster, adapt to change, and deliver experiences tailored to specific and dynamic user needs: treating data capabilities as building blocks rather than fixed systems.

Compute costs directly shape how scalable and responsive a system can be, influencing trade-offs between performance, complexity, and budget. Managing these costs well means designing smarter workflows to deliver value without waste.

Consumer-Grade UX means delivering a user experience for data tools that matches the simplicity, speed, and intuitiveness users expect from everyday consumer apps. It's about removing friction, making complex tasks feel easy, and driving adoption through thoughtful design.

Contract-Driven Development is the practice of defining clear, upfront agreements (contracts) between data producers and consumers. It ensures teams can work independently, reduces integration risks, and treats data interfaces as stable, reliable products.