Data Engineering

AI risk management is the coordinated process of identifying, analysing, evaluating, treating, monitoring, and communicating risks arising from the design, development, deployment, and use of AI systems.

Algorithmic bias is a systematic pattern in an algorithmic system that can produce unfair, inaccurate, or disproportionately harmful outcomes for particular people or groups.

Google Cloud’s fully managed, serverless analytical data warehouse for querying and processing large datasets.

Composable architecture is a modular approach in which systems are assembled from independent, interchangeable capabilities connected through defined interfaces.

Consent Mode is a Google tag-platform feature that adjusts tag and app-SDK behaviour according to a user’s consent choices.

Conversion lift is the incremental increase in conversions caused by an intervention, estimated by comparing outcomes for an exposed treatment group with a comparable unexposed control group.

Cost of goods sold is the expense recognised for inventory or goods sold during a reporting period, based on the costs assigned to those goods.

Cross-device tracking links activity from two or more devices to the same person, household or profile.

Cross-selling is the practice of offering a customer a complementary product or service alongside an existing or intended purchase.

A customer data platform (CDP) is packaged software that creates persistent, unified customer profiles and makes them available to other systems.

A customer journey is the sequence of interactions, decisions and experiences a person has while pursuing a goal with or around an organisation.

Customer relationship management (CRM) is the strategy, processes and technology used to manage interactions and data across prospect and customer relationships.

Data exfiltration is the unauthorised transfer, disclosure, or removal of data from a system, network, application, or controlled environment.

Data governance is the system of decision rights, accountability, policies, standards, and controls used to manage data as an organisational asset.

A repository designed to store large volumes of raw or lightly processed structured, semi-structured and unstructured data.

A data architecture that combines lake-style flexible storage with warehouse-style management, reliability and analytical capabilities.

Data lineage is a traceable record of where data originated, how it moved and changed, and where it is used across systems and processes.

The process of discovering patterns, relationships or anomalies in datasets using statistical, computational and machine-learning methods.

Data observability is the ability to understand and diagnose the health, behaviour, and reliability of data across pipelines and platforms using metadata and telemetry.

A data pipeline is a sequence of processes that collects, moves, validates, transforms, and delivers data from one or more sources to target systems.

Data quality is the degree to which data is accurate, complete, consistent, timely, valid, unique, and fit for its intended use.

A data warehouse is a central analytical data store that integrates historical data from multiple sources for reporting, analysis, and decision support.

Encryption at rest protects stored data by transforming it into ciphertext that requires authorised cryptographic keys to read.

Entity SEO is the practice of making identifiable people, organisations, products, places, and concepts—and their relationships—clear and consistent across content and machine-readable signals.

Event-driven architecture is a design approach in which components produce, detect, and respond to events representing changes or occurrences in a system.

Explainable AI comprises methods and system properties that help relevant people understand the basis, behaviour, or limitations of an AI system and its outputs.

Extract, Load, Transform is a data-integration pattern in which source data is loaded into the target platform before transformations are executed there.

Extract, Transform, Load is a data-integration pattern in which data is retrieved from sources, transformed into the required structure, and then loaded into a target system.

Data collected directly by an organisation through its own interactions with customers, users, prospects or audiences.

Generative Engine Optimization is the practice of improving how accurately and visibly an entity or its content is retrieved, represented, attributed, and cited in responses produced by generative search and answer systems.

Idempotency is the property of an operation that produces the same intended state whether it is performed once or repeated with identical inputs.

Incrementality is the additional outcome caused by an intervention compared with the outcome that would have occurred without it.

Integration Platform as a Service is a cloud-delivered platform for building, operating, and governing integrations among applications, data sources, APIs, and processes.

JavaScript Object Notation is a lightweight, text-based, language-independent format for exchanging structured data.

A knowledge graph is a structured representation of entities, their attributes and the relationships between them, organised for machine querying and reuse.

A large language model (LLM) is a language model trained on very large datasets and parameterised to process or generate language and related sequences.

Lead generation is the process of attracting prospective buyers and obtaining sufficient permission or information to begin a measurable commercial relationship.

Lead nurturing is the coordinated use of relevant communications and experiences to help a prospect progress towards an appropriate next decision over time.

Lead scoring is a method for assigning values to prospect attributes and behaviours to support prioritisation or routing against an agreed outcome.

A lifecycle stage is a governed classification representing where a person, account or opportunity currently sits in an organisation's defined commercial relationship.

LLM tokenisation is the conversion of text or other input into model-specific token identifiers that a large language model can process.

Marketing attribution is the assignment of credit for a conversion or other outcome to marketing touchpoints along a measured customer path.

A marketing qualified lead (MQL) is a lead that meets agreed marketing criteria indicating stronger fit or engagement than other leads, but is not necessarily sales-ready.

A model card is a structured document that describes an AI or machine-learning model’s intended uses, performance, evaluation conditions, limitations, and relevant governance information.

Monthly recurring revenue (MRR) is the normalised recurring subscription revenue a business expects from active customers for one month.

The NIST AI Risk Management Framework is a voluntary, rights-preserving and use-case-agnostic framework for helping organisations manage risks to individuals, organisations, and society across the AI lifecycle.

Observability is the ability to understand a system’s internal state and behaviour from the telemetry and contextual information it produces.

The use of statistical or machine-learning models to estimate future or unknown outcomes from available data.

An AI architecture that combines large language model generation with real-time retrieval from external knowledge sources to produce grounded, verifiable responses.

A strategic function that aligns sales, marketing, and customer success operations under a unified data model, process framework, and technology stack to drive predictable, efficient revenue growth.

The collection and forwarding of measurement events through organisation-controlled server infrastructure rather than solely from a user’s browser or device.

Structured data is machine-readable information added to content using a defined vocabulary and format to describe entities, properties, and relationships explicitly.

System integration is the practice of connecting applications, data stores, infrastructure, and business processes so they exchange information and operate as a coordinated system.

Technical debt is the future cost and constraint created when a technical solution is expedient now but requires additional correction, maintenance, or redesign later.

A Urchin Tracking Module parameter is a standardised utm_-prefixed query parameter added to a destination URL so analytics tools can classify campaign traffic.

Vendor lock-in is the condition in which switching from a provider, platform, or product becomes disproportionately difficult or costly because of technical, contractual, data, or operational dependencies.

A view-through conversion is a conversion attributed to an ad impression when the person did not click or otherwise interact with that ad before converting within a defined window.

Website cookies are small name–value records that a web server asks a user agent to store and return with eligible HTTP requests to maintain state.

Zero-party data is information that a customer intentionally and proactively shares with an organisation, such as preferences, intentions, context, or desired recognition.