Glossary

917 terms defined in plain English. Each one links to the latest coverage in the index, and to the tools and guides that use it.

By topic

A

A/A test: An experiment that splits users into two identical groups with no real difference, used to check that the testing system itself is unbiased.

A/B testing: A controlled experiment that compares two variants to measure which one produces a better outcome.

Absenteeism rate: The share of scheduled work time lost to unplanned employee absence over a period.

Activation rate: The share of new users who complete the specific actions that mark them as having found a product's core value.

Actuarial modeling: Statistical modeling of future insurance losses, used to set premium prices and reserve for future claims.

Ad viewability: A measure of whether an ad impression had the actual opportunity to be seen, based on standardized on-screen thresholds.

Administrative data: Records collected by government agencies for operational purposes, later reused for research, statistics, or evaluation.

Advanced metering infrastructure (AMI): Smart meters and the communication network that lets utilities collect near real-time usage data from them.

Agentic analytics: The use of AI agents that autonomously plan and execute multi-step analytics work, such as investigating a metric change.

Aha moment: The point at which a new user first experiences the core value of a product clearly enough to want to continue.

AI agent: A system that uses a language model to plan and take multi-step actions, such as calling tools or APIs, toward a goal.

AI copilot: An AI tool embedded in a work application that assists a human with suggestions or drafts, rather than working unsupervised.

Air quality index (AQI): A standardized scale that converts measured pollutant concentrations into a single number describing health risk.

AIS vessel tracking: Using shipboard AIS radio broadcasts to track vessel position, speed and identity for maritime and environmental analysis.

Algorithmic bias: Systematic, unfair skew in a model's outputs that disadvantages particular groups, usually traced to skewed training data or design choices.

Algorithmic trading: Using computer programs to automatically place and manage trade orders based on predefined rules.

All-source analysis: Combining intelligence from multiple collection disciplines into one assessment instead of relying on a single stream.

Alpha: The portion of an investment's return not explained by its exposure to a benchmark or risk factors.

Alpha decay: The tendency of a trading signal's predictive power and profitability to weaken over time as more traders exploit it.

Alternative data: Non-traditional data sources, such as satellite imagery or card transactions, used to generate an investing edge.

Altmetrics: Measures of research attention outside traditional citation counts, such as social media mentions, downloads and news coverage.

American Community Survey (ACS): An ongoing U.S. Census Bureau survey providing detailed annual demographic, social, and economic data.

Analytics center of excellence (CoE): A central group that sets standards, tools and training for analytics work across an organization's business units.

Analytics certification: A credential, from a vendor, platform, or professional body, that verifies a specific analytics skill or tool proficiency.

Analytics engineer: A role that builds and maintains the transformed, tested data models analysts and dashboards rely on.

Analytics maturity model: A staged framework describing how an organization's use of data and analytics develops from ad hoc to fully embedded.

Analytics ROI: A measure of the financial return an organization gets from its investment in analytics people, tools and data.

Analytics translator: A role that bridges business stakeholders and technical data teams, turning business problems into analytics questions.

Annual recurring revenue (ARR): The value of a subscription business's recurring revenue, normalized to a one-year run rate.

Anomaly detection: Identifying data points or patterns that deviate significantly from expected or normal behavior.

Anonymization: Altering data so no individual can be re-identified from it, even by combining it with other available information.

Anti-money laundering (AML): The laws, controls, and analytics financial institutions use to detect and prevent the disguising of illegal funds.

Apache Iceberg: An open table format adding ACID transactions, schema evolution, and time travel to files stored in a data lake.

Apache Kafka: An open-source distributed event streaming platform that stores records in ordered, durable, replayable logs called topics.

Apache Parquet: An open-source, columnar file format for storing large analytical datasets efficiently in disk or object storage.

Apache Spark: An open-source distributed computing engine for processing large datasets in parallel across a cluster of machines.

API integration: Connecting two systems by having one call the other's application programming interface to exchange data.

App store optimization (ASO): Improving a mobile app's visibility and conversion rate within app store search and browse listings.

Application performance monitoring (APM): Monitoring that tracks application response time, errors, and throughput, tying degraded performance to its root cause.

ARIMA (autoregressive integrated moving average): A classical time-series forecasting model that combines a series' own past values, differencing, and past forecast errors.

Assay data analysis: Processing and interpreting the results of a biological or chemical test to determine a measurable outcome.

Asset allocation: The decision of how to split an investment portfolio across asset classes such as stocks, bonds, and cash.

Astronomical survey data: The large, systematically collected datasets produced by telescopes that repeatedly scan wide areas of the sky.

Attention metrics: Measures of how much genuine human attention content receives, going beyond impressions to viewability, scroll depth, and time in view.

Attribution modeling: A method for assigning credit for a conversion across the marketing touchpoints a customer encountered beforehand.

Attribution window: The fixed period after an ad click or view during which a resulting conversion is still credited to that ad.

Attrition risk: A predictive score estimating how likely an individual employee or customer is to leave in a given period.

Audience measurement: Quantifying how many people are exposed to media content and who they are, across TV, streaming, print, and digital.

Audience retention curve: A chart showing what percentage of viewers are still watching a video at each point in its runtime.

Audio analytics: The analysis of sound and audio signals, from speech and music to ambient noise, to extract patterns and meaning.

Augmented analytics: Using AI and machine learning to automate parts of data preparation, insight discovery, and explanation in analytics tools.

Authorization rate: The share of attempted card or bank payments that the issuer approves rather than declines.

Autocorrelation: The correlation of a time series with a lagged, delayed copy of itself, used to detect repeating patterns and dependence over time.

Automated valuation model (AVM): A statistical or machine-learning model that estimates a property's market value from comparable sales and property data, without a human appraiser.

Automatic speech recognition (ASR): Technology that converts spoken audio into written text, forming the input layer for most speech and voice analytics.

AutoML (automated machine learning): Tooling that automates parts of building a machine learning model, such as feature selection, algorithm choice and hyperparameter tuning.

Average daily rate (ADR): The average revenue earned per occupied hotel room per day, excluding unsold rooms.

Average deal size: The mean revenue value of closed-won deals over a period, found by dividing total value by deal count.

Average draft position (ADP): The average pick number at which a player is selected across many fantasy sports drafts, a consensus measure of value.

Average handle time (AHT): The average duration of a customer service interaction, including talk time, hold time and after-call work.

Average order value (AOV): The average amount spent each time a customer places an order, revenue divided by number of orders.

Average revenue per daily active user (ARPDAU): The average revenue earned per daily active user, a monetization metric common in mobile games and free-to-play apps.

Average revenue per user (ARPU): The average revenue generated per subscriber or user over a defined period, typically a month or year.

Average session duration: The mean length of time visitors spend in a session, measured differently by different analytics platforms.

Average treatment effect (ATE): The mean difference in outcome between treated and untreated units across a full population, the standard headline result of an experiment.

B

Backfill: Reprocessing historical data through a pipeline, typically after a bug fix, schema change or new metric definition.

Backlink: A hyperlink from one website to a page on another, treated by search engines as a signal of authority or relevance.

Backtesting: Testing a trading or investment strategy against historical data to see how it would have performed.

Batch processing: Processing data in large, scheduled groups rather than continuously as each record arrives.

Batting average on balls in play (BABIP): The batting average on balls put in play, excluding home runs and strikeouts, used to gauge how much luck shaped a batting line.

Batting strike rate: A cricket batter's scoring speed, runs scored per 100 balls faced, used to judge tempo rather than overall output.

Bayesian inference: A statistical approach that updates the probability of a belief as new evidence arrives, starting from a prior assumption.

Benchmark index: A market index used as the reference point for judging whether a portfolio or fund's performance and risk are competitive.

Beta: A measure of how much an asset's returns move relative to a benchmark, most often the broad market.

Bibliometrics: The statistical study of publications, citations and authorship patterns to measure research output and influence.

Bid-ask spread: The gap between the highest price buyers will pay and the lowest price sellers will accept for an asset.

Biodiversity net gain (BNG): A planning requirement that a development leave measurable biodiversity higher than before construction began.

Black-Scholes model: A mathematical model for pricing European-style options from the underlying's price, volatility, and time to expiry.

Blended CAC (blended customer acquisition cost): Total customer acquisition cost calculated across all marketing channels combined, including organic and unpaid efforts.

Bond duration: A measure of a bond's price sensitivity to changes in interest rates, expressed in years.

Bookings, billings, and revenue: Three distinct SaaS figures often confused: signed contract value, amount invoiced, and amount recognized as earned.

Bootstrapping: A technique for estimating the uncertainty of a statistic by repeatedly resampling the observed data with replacement.

Bot traffic: Website activity generated by automated programs rather than human visitors, ranging from crawlers to fraudulent scripts.

Bottleneck analysis: Identifying which step in a process limits the speed of the whole system, so improvement effort targets the right place.

Bounce rate: The share of sessions that show no interaction beyond the entry page, defined differently by GA4 and classic web analytics.

Box plus-minus (BPM): A basketball model estimating a player's point contribution per 100 possessions relative to league average, using only box-score stats.

Brand health: A composite view of how a brand is perceived and performing, combining awareness, sentiment, consideration, and loyalty metrics.

Brand lift: The measurable increase in brand awareness, favorability, or purchase intent caused by an ad campaign.

Break point conversion (rate): The share of break point chances a tennis player turns into a won game on the opponent's serve.

Break-even analysis: Calculating the sales volume or revenue at which total costs are exactly covered, with neither profit nor loss.

Browser fingerprinting: Identifying a device from browser and device characteristics, such as screen size or fonts, without setting a cookie.

Budget categorization: Sorting income and expenses into defined categories to see where money goes and compare it against a plan.

Budget variance (government): The difference between a government agency's budgeted and actual spending or revenue over a reporting period.

Budget vs. actual: A report comparing what was budgeted for a period against what actually happened, line by line.

Bullwhip effect: The tendency for small changes in end-customer demand to be amplified into far larger order swings further up the supply chain.

Burn rate: How fast a company spends cash, usually measured monthly, and used to estimate how long its remaining cash will last.

Business glossary: A controlled list of business terms and their agreed definitions, kept separate from the technical fields that implement them.

Business intelligence (BI): The practice of turning company data into dashboards, reports and metrics that guide business decisions.

C

CAC payback period: How many months it takes for the gross margin a customer generates to repay the cost of acquiring them.

California Consumer Privacy Act (CCPA): A California state law giving residents rights over their personal data and requiring businesses to disclose how it is collected and sold.

Capacity factor: A power plant's actual output over a period, as a share of what it would produce running at full capacity the whole time.

Capitalization rate (cap rate): The ratio of a property's annual net operating income to its current market value, used to compare income-property returns.

Carbon accounting: Measuring, tracking and reporting an organization's greenhouse gas emissions in standardized units of CO2-equivalent.

Carbon credit: A tradable certificate representing one tonne of CO2-equivalent avoided, reduced or removed from the atmosphere.

Carry trade: Borrowing in a low-interest-rate currency to invest in a higher-yielding one, profiting from the rate gap unless the exchange rate moves against it.

Cart abandonment rate: The share of shopping carts that are created but never converted into a completed purchase.

Cash conversion cycle (CCC): The number of days it takes a company to turn spending on inventory and operations back into collected cash.

Catastrophe model (cat model): A model that simulates potential natural disasters and estimates the resulting financial losses across an insured portfolio.

Causal inference: The set of methods used to determine whether one variable actually causes a change in another, not merely correlates with it.

Cell site analytics: Analyzing performance and usage data from individual cell towers to manage network capacity, coverage and quality.

Census tract: A small, relatively stable geographic area defined by a national statistical agency for reporting population data.

Central limit theorem: The principle that the average of many independent samples tends toward a normal distribution, regardless of the data's shape.

Centrality measures: A family of graph metrics that rank how important or well-connected each node is within a network.

Change data capture (CDC): A technique for detecting and streaming only the rows that changed in a source system, instead of reloading everything.

Change failure rate: The percentage of production deployments that result in a degraded service requiring a rollback, hotfix, or incident.

Change-point detection: Identifying the points in a time series where its underlying statistical behavior shifts abruptly rather than drifting gradually.

Chargeback rate: The share of a merchant's transactions that a cardholder disputes and the issuer reverses.

Chi-square test: A statistical test that checks whether observed frequencies in categorical data differ from what chance would produce.

Chief data officer (CDO): The senior executive responsible for an organization's data strategy, governance and analytics capability.

Choropleth map: A map where regions are shaded or colored to represent the value of a statistic, such as population density.

Churn rate: The share of customers or revenue a business loses over a given period.

Citation analysis: Studying the pattern of citations between publications to measure influence, map research fields or evaluate researchers.

Civic tech: Technology built to improve how citizens interact with government and participate in public decisions.

Claims analytics: Applying analytics to insurance claims data to speed processing, control cost, and detect fraud.

Classification: A supervised learning task that assigns each input to one of a fixed set of discrete categories or labels.

Clearance rate: The share of reported crimes that end in an arrest or are otherwise resolved by law enforcement within a period.

Click ID: A unique parameter appended to a URL when an ad is clicked, used to match that click to a later conversion.

Click-through rate: The percentage of people who click a link, ad, or search result out of everyone who saw it.

Clickstream: The raw, timestamped sequence of pages, clicks, and interactions a single visitor generates while browsing a site or app.

Climate model downscaling: Converting coarse global climate model output into finer-resolution projections usable for a specific region or site.

Climate scenario analysis: Modeling how different future climate and policy pathways would affect an organization's assets, costs or strategy.

Climate value-at-risk (climate VaR): An estimate of how much climate change and the transition to low carbon could reduce the value of a portfolio.

Clinical decision support (CDS): Software that gives clinicians patient-specific alerts or recommendations at the point of care.

Clinical trial analytics: Analytics applied to designing, running, and analyzing clinical trials, from site selection through final results.

Closing line value (CLV): How favorable a bettor's odds were when they placed a bet compared with the market's final odds before the event started.

Cloud unit economics: Cloud infrastructure cost measured per unit of business value, such as cost per customer, transaction, or API call.

Clustering: An unsupervised technique that groups data points so items within a group are more similar to each other than to items outside it.

Coastal flood modeling: Simulating where and how deeply coastal areas will flood from storm surge, high tides and sea-level rise.

Cohort analysis: Grouping customers by a shared starting point, such as signup month, and tracking how their behavior diverges over time.

Columnar storage: A file and database layout that stores each column's values together, instead of storing each row together.

Combined ratio: An insurer's total claims and expenses as a share of premium earned, the standard measure of underwriting profitability.

Commitment-based discounts: Reduced cloud pricing offered in exchange for committing to a minimum level of spend or usage over a fixed term.

Common table expression: A named, temporary result set defined at the start of a SQL query and referenced later in that query.

Community detection: Identifying groups of nodes in a network that are more densely connected to each other than to the rest of the graph.

Competitive intelligence: The systematic gathering and analysis of public information about competitors to inform strategy and decisions.

Completion percentage over expected (CPOE): An American football stat comparing a quarterback's actual completion rate to the rate a model expects given each throw's difficulty.

Composable CDP: A customer data platform assembled from separate best-of-breed tools on top of an organization's own warehouse, rather than one bundled product.

Computational fluid dynamics (CFD): Using numerical simulation to predict how gases and liquids flow and interact with surfaces.

Computer vision: A field of AI that trains models to interpret and extract information from images and video, such as detecting objects or faces.

Concept drift: A change over time in the underlying relationship between a model's inputs and the outcome it predicts.

Confidence interval: A range calculated from sample data that, under repeated sampling, would contain the true value a stated percentage of the time.

Confusion matrix: A table comparing a classifier's predicted labels against actual labels, breaking results into true and false positives and negatives.

Connector: A prebuilt, reusable component that handles the specifics of extracting data from or loading data into a system.

Consent management platform (CMP): Software that collects, records, and enforces a person's consent choices for cookies and data collection across a site or app.

Consent mode: A Google framework that adjusts how its tags behave based on a visitor's cookie consent choices.

Consumer price index (CPI): A measure of the average change in prices paid by consumers for a fixed basket of goods and services.

Contango and backwardation: The two shapes a futures curve can take, contango slopes upward with time to delivery, backwardation slopes downward.

Content engagement: The volume and rate of likes, comments, shares, and other interactions a piece of content receives relative to its reach.

Content velocity: The rate at which an organization or creator publishes new content over time, tracked to sustain audience growth and visibility.

Context window: The maximum amount of text, measured in tokens, a language model can consider at once when generating a response.

Contract analytics: Using software to extract, review, and analyze clauses, terms, and obligations across many contracts at scale.

Contribution margin: Revenue minus all variable costs, showing how much each sale contributes toward covering fixed costs and profit.

Conversation intelligence: Software that records, transcribes and analyzes sales or support calls to surface talk patterns, objections and coaching cues.

Conversion API (CAPI): A method of sending conversion and event data directly from a business's own servers to an ad platform, bypassing the browser.

Conversion lift study: A randomized test comparing a holdout who saw no ads against an exposed group, to measure advertising's true incremental effect on conversions.

Conversion rate: The proportion of sessions, users, or clicks that complete a defined goal, such as a purchase or signup.

Cookie: A small piece of data a website stores in a visitor's browser to recognize them on return visits or across pages.

Cookieless tracking: Measuring visitor behavior without relying on persistent third-party cookies, using first-party, server-side or modeled data.

Core Web Vitals: Google's page-experience metrics for loading speed, interactivity and visual stability, used as a minor ranking signal.

Correlation: A statistic between -1 and 1 that measures the strength and direction of a linear relationship between two variables.

Corsi: A hockey puck-possession proxy counting all shot attempts, on goal, missed or blocked, for and against while a player is on the ice.

Cost allocation: Assigning shared or indirect costs, such as overhead, to the departments, products, or customers that use them.

Cost of living index (COLI): An index comparing the relative cost of a fixed basket of goods and services across different places.

Cost per acquisition (CPA): The average amount spent on advertising to generate one conversion, such as a sale, lead, or signup.

Cost per click (CPC): The amount an advertiser pays, on average, each time a user clicks on a paid ad.

Cost per mille (CPM): The price an advertiser pays for one thousand ad impressions, regardless of clicks or conversions.

Cost-benefit analysis (CBA): Comparing a policy or project's monetized benefits against its costs to judge whether it is worth doing.

Counterfactual: What would have happened to a unit or group without a treatment, the unobservable baseline every causal claim is measured against.

Course completion rate: The share of students who enroll in a course and finish it, a core measure of engagement in online and adult learning.

Crack spread: The price gap between crude oil futures and refined product futures like gasoline, used to gauge refining profit margins.

Crash rate: The share of app sessions, or the rate per active user, in which the app crashed or force-closed unexpectedly.

Crawl budget: The number of pages a search engine's crawler will fetch from a site within a given period before moving on.

Creator economy: The ecosystem of independent content creators who monetize an audience directly through platforms, sponsorships, and subscriptions.

Credit card panel data: Anonymized, aggregated card transaction data used to estimate company revenue and consumer spending before official reports.

Credit scoring: Ranking a borrower's creditworthiness with a numeric score derived from their credit history and other data.

Credit spread: The extra yield a bond offers over a comparable risk-free benchmark, compensating for credit risk.

Crime mapping: Plotting reported crime incidents geographically to reveal spatial patterns and support resource deployment decisions.

Crime rate: The number of reported crimes in an area relative to its population, usually expressed per 100,000 residents.

Crop modeling: Using mathematical models of plant growth, weather and soil to simulate and forecast crop development and yield.

Cross-validation: A method for estimating how well a model generalizes by repeatedly training and testing it on different splits of the same data.

Cultural analytics: Applying computational and statistical methods to large collections of cultural material to study patterns over time.

CUPED (controlled-experiment using pre-experiment data): A variance-reduction technique that uses each user's pre-experiment behavior to sharpen an experiment's estimate without adding more traffic.

Custom dimension: An attribute a site owner defines and sends to their analytics tool to capture data not tracked by default.

Customer 360: A single, unified profile of a customer that consolidates data from every touchpoint and system into one view.

Customer acquisition cost: The average total sales and marketing spend required to acquire one new paying customer.

Customer data platform (CDP): Software that unifies customer data from multiple sources into persistent profiles usable by marketing teams.

Customer effort score (CES): A survey metric measuring how much effort a customer had to spend to resolve an issue or complete a task.

Customer health score: A composite score combining usage, engagement, support and billing signals to estimate a customer's renewal or churn risk.

Customer journey analytics: Analyzing a customer's touchpoints across channels and time to understand and improve the full path to purchase and beyond.

Customer journey mapping: The practice of visualizing every touchpoint a customer has with a brand, from first awareness through purchase and beyond.

Customer lifetime value (CLV): The total net profit a business expects to earn from a customer over the entire relationship.

Customer satisfaction score (CSAT): A survey metric that asks customers to rate how satisfied they were with a specific interaction, product or purchase.

Customer segmentation: Dividing a customer base into groups that share meaningful traits so each can be targeted or served differently.

Cycle time: The total elapsed time to complete one unit of work, from when active processing starts to when it finishes.

D

Daily active users (DAU): The count of unique users who take a qualifying action in a product within a single calendar day.

Daily fantasy sports (DFS): Paid fantasy contests lasting a single day or slate, where entrants build a roster under a salary cap for cash prizes.

Damage assessment: Systematically estimating the scope and severity of destruction after a disaster to guide response and recovery funding.

Dark social: Content sharing that happens through channels analytics tools cannot attribute, such as private messaging apps and copied links.

Dashboard: A visual display of the metrics a team needs to monitor, usually updated on a schedule or in real time.

Data analyst: A practitioner who explores and reports on data, typically with SQL and BI tools, to answer specific business questions.

Data catalog: A searchable inventory of an organization's data assets, with descriptions, ownership and usage attached.

Data classification: Categorizing data by sensitivity or type so that handling, access, and protection rules can be applied consistently and automatically.

Data clean room: A controlled environment where two or more parties analyze combined data together without either side seeing the other's raw records.

Data contract: A formal, versioned agreement on a dataset's schema and semantics between the team producing it and its consumers.

Data deduplication: Identifying and removing records that represent the same real-world entity or event more than once.

Data dictionary: A structured reference listing every table, column, and field in a database along with its type, meaning, and constraints.

Data drift: A change in the statistical properties of input data over time that can silently degrade a deployed model's accuracy.

Data engineer: A practitioner who builds and maintains the pipelines and infrastructure that make data reliably available for analysis.

Data ethics: The study and practice of using data responsibly, addressing fairness, consent, transparency, and harm beyond what regulation strictly requires.

Data fabric: An architecture that uses metadata and automation to integrate and provide unified access to data across many sources.

Data freshness: How current a dataset is relative to real time, measured as the lag between an event happening and it being queryable.

Data governance: The policies, roles and processes that determine how data is defined, owned, accessed and kept compliant.

Data ingestion: Moving data from its source systems into a storage or processing system where it can be transformed and analyzed.

Data lake: A storage repository that holds raw data of any format cheaply, with structure applied only when the data is read.

Data lakehouse: A storage architecture that adds warehouse-style structure and transactions directly on top of low-cost data lake storage.

Data layer: A structured object on a page that exposes site and event data for a tag management system to read.

Data lineage: A traceable record of where data came from, how it was transformed, and where it is used downstream.

Data literacy: The ability to read, interpret, question, and communicate with data accurately enough to make sound decisions.

Data mart: A smaller, focused subset of a data warehouse built around the needs of one business function or team.

Data masking: Replacing sensitive values with realistic but fake substitutes so data can be used for testing or analysis without exposure.

Data mesh: A decentralized approach where domain teams own and publish their own data as products, instead of one central team.

Data modeling: The practice of organizing data into entities, relationships and schemas that make it accurate and efficient to query.

Data observability: Continuous automated monitoring of pipelines and datasets to catch freshness, volume and quality problems early.

Data pipeline: The automated sequence of steps that moves data from source systems to a destination, transforming it along the way.

Data portfolio: A curated collection of a practitioner's own projects and analyses, used to demonstrate skill to employers or clients.

Data product: A dataset or data service built, owned and maintained with product-level rigor for a defined set of consumers.

Data profiling: Automatically scanning a dataset to summarize its structure, value distributions, and quality before it is trusted or modeled.

Data provenance: The documented history of where a dataset came from and every transformation applied to it before its current state.

Data quality: The degree to which data is accurate, complete, consistent and fit for the purpose it is being used for.

Data residency: The requirement or practice of storing and processing data within a specific country or region's physical or legal boundaries.

Data retention policy: A documented rule for how long different categories of data are kept before they are deleted or archived.

Data scientist: A practitioner who combines statistics, programming and domain knowledge to build models and extract insight from data.

Data SLA (service-level agreement): A formal commitment about a dataset's freshness, availability or quality, with defined targets and consequences.

Data sovereignty: The principle that data is subject to the laws of the country in which it is collected or stored, regardless of who owns the system.

Data standards: Agreed formats, fields, and definitions that let datasets from different agencies or systems be combined reliably.

Data stewardship: The assignment of named people who are accountable for the quality, definition, and appropriate use of specific data domains.

Data storytelling: Combining data, visuals and narrative so an audience understands an analysis and knows what to do about it.

Data strategy: An organization's documented plan for how data will be collected, governed and used to support its business goals.

Data validation: Checking that data meets explicit, predefined rules for type, range, format, or consistency before it is used or trusted.

Data vault: A modeling methodology that splits data into hubs, links, and satellites to make a warehouse auditable and adaptable to change.

Data virtualization: Querying data across multiple, separate systems in place through a single interface, without first copying it.

Data visualization: Representing data graphically so that patterns, trends and outliers are easier to see than in a table of numbers.

Data visualization literacy: The ability to accurately read, interpret, and critically evaluate charts and graphs, and to recognize when one is misleading.

Data warehouse: A centralized, structured database optimized for fast historical reporting rather than day-to-day transactions.

Data-driven attribution (DDA): An attribution model that uses statistical analysis of historical conversion paths to assign fractional credit to each touchpoint.

Data-driven decision-making (DDDM): Grounding organizational decisions primarily in data and analysis rather than relying mainly on intuition or precedent.

DataFrame: A two-dimensional, table-like data structure with named, typed columns, used to manipulate data in code.

DataOps: Applying DevOps practices, such as automated testing, version control and continuous delivery, to data pipelines.

Days of supply: How many days current inventory would last at the recent or expected rate of usage before running out.

dbt (data build tool): An open-source tool from dbt Labs that lets analytics teams write data transformations as version-controlled, tested SQL.

Debt-to-income ratio (DTI): The share of a person's gross monthly income that goes toward debt payments, used to assess borrowing capacity.

Decentralized exchange (DEX): A crypto trading platform that executes trades directly between users' wallets via smart contracts, without a central custodian.

Decision intelligence: A discipline combining data, analytics, and decision modeling to design, support, and improve how organizations make decisions.

Decision tree: A supervised model that predicts an outcome by following a sequence of if-then splits on feature values, shaped like a tree.

Deep learning: Machine learning using neural networks with many layers, able to learn complex patterns directly from raw data such as images or text.

Deep linking: A link that opens an app directly to a specific piece of content or screen, instead of the app's default landing page.

Deferred revenue: A balance-sheet liability for cash received from a customer before the related goods or services have been delivered.

Deforestation monitoring: Tracking loss of tree cover over time, usually from satellite imagery, to detect and quantify forest clearing.

Degree days: A measure of how much and for how long temperature departs from a reference point, used to estimate heating and cooling demand.

Demand forecasting: Estimating future customer demand for a product so inventory, production and staffing can be planned ahead.

Denormalization: Deliberately introducing redundancy into a data model to reduce joins and speed up read queries.

Density functional theory (DFT): A quantum mechanical method that predicts a material's properties from electron density rather than tracking every electron.

Dependency ratio: The ratio of dependents, children and older adults, to the working-age population, per 100 people.

Deployment frequency: How often an organization successfully releases code to production, one of the four DORA metrics.

Descriptive analytics: Analytics that summarizes what has already happened in the data, without explaining why or predicting what comes next.

Design of experiments (DOE): A structured, statistical approach to planning experiments so the effect of each factor can be estimated efficiently.

Diagnostic analytics: Analytics focused on explaining why something happened by drilling into patterns, correlations, and root causes behind results.

Difference-in-differences (DiD): A causal method that compares the before-and-after change in a treated group against the before-and-after change in an untreated group.

Differential expression analysis: Statistically identifying which genes are expressed at significantly different levels between two or more conditions.

Differential privacy: A mathematical technique for publishing aggregate statistics while guaranteeing no individual record can be inferred from the results.

Digital twin: A live virtual model of a physical asset or process, kept synchronized with real sensor data, used to simulate and predict behavior.

Dimension table: A table of descriptive attributes, such as customer or product details, that a fact table references for context.

Dimensionality reduction: Techniques that compress a dataset with many variables into fewer dimensions while preserving as much meaningful information as possible.

Direct traffic: Website visits an analytics tool cannot attribute to a referring source, such as a typed URL or a bookmark.

Directed acyclic graph (DAG): A graph of one-way dependency edges between tasks, with no cycles, used to define the run order of a pipeline.

Discounted cash flow (DCF): A valuation method estimating an asset's worth as the present value of its expected future cash flows.

Displacement risk: A modeled estimate of how likely a neighborhood or household is to experience housing-driven displacement.

Distant reading: Analyzing patterns across a large collection of texts computationally instead of closely reading each individual work.

Distributed computing: Splitting a computation across multiple machines that coordinate over a network to solve it faster or at larger scale.

Distributed tracing: Following a single request across multiple services in a distributed system to see where time was spent or an error occurred.

Dollar-cost averaging (DCA): Investing a fixed amount at regular intervals regardless of price, buying more shares when prices are low and fewer when high.

Domain authority: A third-party score predicting how well a domain is likely to rank, based largely on its backlink profile — not a Google metric.

Donor lifetime value (DLV): The total giving an organization expects from a donor over the full length of their relationship with it.

Donor retention rate: The share of donors from one period who give again in the next period, a core measure of nonprofit fundraising health.

DORA metrics: Four measures of software delivery and operational performance covering release speed, stability, and recovery.

Double materiality: Assessing both how sustainability issues affect a company's finances and how the company affects the world.

Draft analytics: The use of statistical models to evaluate amateur prospects and to value draft picks themselves.

Driver-based planning: Building a financial plan from operational drivers, such as units sold or headcount, rather than top-down growth assumptions.

Driving accuracy: The percentage of tee shots on par-4 and par-5 holes that land in the fairway, a measure of off-the-tee control.

Duckworth-Lewis-Stern method (DLS): The formula used to set revised targets in limited-overs cricket matches shortened by rain or other interruptions.

Dynamic pricing: Automatically adjusting prices in near real time based on demand, inventory, timing, and competitor behavior.

E

E-discovery (electronic discovery): Identifying, collecting, and reviewing electronically stored information as evidence in litigation or investigations.

Early warning indicators (education): Data signals such as attendance and grades that flag students at risk of falling behind or dropping out.

Early warning system (EWS): A monitoring and alerting system that detects a developing hazard and issues warnings before it strikes a population.

Earned Media Value (EMV): An estimate of what organic social mentions, shares, and coverage would have cost if purchased as paid advertising.

EBITDA (earnings before interest, taxes, depreciation, and amortization): A profitability measure that adds back interest, taxes, depreciation, and amortization to earnings.

Economy rate: A cricket bowler's runs conceded per over, the standard measure of how tight or expensive their bowling is.

Effect size: A standardized measure of the magnitude of a difference or relationship, independent of sample size.

Effective field goal percentage (eFG%): A basketball shooting percentage that counts a made three-pointer as worth 1.5 field goals, reflecting its true point value.

Electronic health record (EHR): The digital, longitudinal record of a patient's health data, maintained across their encounters with a provider.

Electronic lab notebook (ELN): Software that replaces the paper lab notebook, letting researchers record experiments, protocols and observations digitally.

Elo rating: A rating system that updates a competitor's skill estimate after each result, weighted by the strength of the opponent faced.

ELT (extract, load, transform): A data integration pattern that loads raw data first and transforms it afterward using the warehouse's own compute.

Email deliverability: The rate at which sent emails actually reach a recipient's inbox rather than being blocked, filtered, or bounced.

Email open rate: The share of delivered emails recorded as opened, a metric increasingly distorted by privacy features that pre-fetch images.

Embedded analytics: Charts, reports or dashboards built directly into another application instead of a separate BI tool.

Embeddings: Numeric vector representations of text, images, or other data, positioned so similar items sit close together in the vector space.

Emergency resource allocation: Deciding where to deploy limited personnel, equipment, and supplies during a disaster response, based on need and access.

Emissions factor: A standard value converting a unit of activity, such as fuel burned, into the greenhouse gases it produces.

Employee engagement: The degree to which employees feel committed to, motivated by and emotionally invested in their work and organization.

Employee net promoter score (eNPS): A single-question survey metric, adapted from NPS, measuring how likely employees are to recommend their employer.

Employee turnover rate: The share of employees who leave an organization, voluntarily or involuntarily, over a given period.

Endpoint detection and response (EDR): Software that continuously records endpoint activity, applies detection logic, and lets responders investigate and contain threats.

Engaged time: The time a user actively interacts with content, such as scrolling or clicking, as distinct from a page simply staying open.

Engagement rate: The share of sessions that meet a minimum bar for meaningful interaction, GA4's replacement for bounce rate.

Ensemble forecasting: Running many slightly varied forecast simulations to express weather or climate uncertainty as a range instead of one number.

Environmental DNA (eDNA): Genetic material shed by organisms into water, soil or air, sampled and sequenced to detect which species are present.

Error budget: The amount of unreliability a service is allowed over a period, equal to 100% minus its service level objective.

ESG reporting: Disclosing an organization's environmental, social and governance performance to investors, regulators and the public.

ETL (extract, transform, load): A data integration pattern that transforms data before loading it, unlike ELT which transforms after loading.

EU AI Act: The European Union's regulation classifying AI systems by risk level and imposing obligations that scale with that risk.

Evacuation modeling: Simulating how a population will move out of a hazard area to plan routes, timing, and capacity before a disaster hits.

Event data model: The structured format — event name, timestamp, user ID, and properties — that a single tracked user action is recorded in.

Event streaming: Continuously publishing and consuming records of things that happened, in order, through a durable, replayable log.

Event taxonomy: The standardized naming and structure for the user actions a product tracks, so events mean the same thing everywhere.

Event tracking: Capturing specific user actions, such as clicks or purchases, as discrete, named pieces of data.

Event-driven architecture (EDA): A system design where components communicate by producing and reacting to events, rather than calling each other directly.

Exactly-once processing: A processing guarantee that each event affects the result exactly once, even after retries or redelivery.

Exit poll: A survey of voters conducted immediately after they leave the polling place, used to estimate results and behavior.

Exit rate: The share of sessions that ended on a given page, out of all sessions in which that page appeared.

Exit velocity: The speed of a baseball as it leaves the bat immediately after contact, measured in miles per hour by tracking systems.

Expansion revenue: Additional recurring revenue from existing customers, through upsells, cross-sells, or usage growth.

Expected assists (xA): A soccer metric estimating the probability that a completed pass will directly lead to a goal, based on similar historic passes.

Expected goals (xG): A shot-quality metric estimating the probability that a given attempt results in a goal, based on similar historic shots.

Expected points added (EPA): An American football stat measuring how much a single play changes a team's expected point value on the scoreboard.

Expected shortfall (ES): The average loss in the worst-case scenarios that exceed a value-at-risk threshold.

Expected value (EV): The probability-weighted average outcome of a bet or decision, used to judge whether it is favorable over the long run.

Expense ratio: The annual percentage of fund assets deducted to cover a fund's operating costs, charged regardless of performance.

Explainable AI (XAI): Techniques and models that make a machine learning system's predictions understandable to humans, rather than a black box.

Exploratory data analysis: An open-ended first pass through a dataset to understand its structure and spot patterns before formal analysis.

Exponential smoothing: A forecasting technique that weights recent observations more heavily than older ones, with the weight decaying exponentially with age.

Exposure data: The detailed inventory of assets, locations, values and characteristics at risk that a catastrophe or risk model runs against.

Extended detection and response (XDR): A platform that correlates telemetry across endpoints, network, identity, and cloud into a single detection and investigation view.

F

F1 score: The harmonic mean of precision and recall, giving a single score that balances both when neither error type is clearly more costly.

Fact table: The table in a dimensional model that stores measurable, numeric events, each row linked to the dimensions describing it.

Factor investing: Building portfolios around specific, historically rewarded drivers of return, such as value or momentum.

FAIR data principles: Guidelines stating that research data should be Findable, Accessible, Interoperable and Reusable.

False discovery rate (FDR): The expected share of "significant" results that are actually false positives, when many statistical tests are run and checked at once.

Fantasy projection: A statistical forecast of the fantasy points a player is expected to score in an upcoming game or season.

Farm management analytics: Software and data practices that track a farm's field operations, inputs and finances to support planting and business decisions.

Feature adoption: The share of eligible users who take up and continue using a specific feature after its release.

Feature engineering: Creating, transforming or selecting the input variables a model is trained on, to make the underlying pattern easier for it to learn.

Feature flag: A configuration switch that turns a piece of functionality on or off, or for specific users, without deploying new code.

Feature store: A centralized system that stores, versions and serves machine learning features so training and live prediction use identical definitions.

Fenwick: A hockey shot-attempt metric like Corsi that excludes blocked shots, counting only shots on goal and missed shots.

Fertility rate (total fertility rate, TFR): The average number of children a woman would have in her lifetime at current age-specific birth rates.

Fielding independent pitching (FIP): A pitching metric on the ERA scale built only from strikeouts, walks, hit-by-pitches and home runs, a pitcher's most direct outcomes.

Fill rate: The share of customer demand fulfilled immediately from available stock, without delay, backorder, or substitution.

Financial close: The recurring process of finalizing a company's accounting records and producing financial statements for a period.

Fine-tuning: Further training a pretrained model on a smaller, targeted dataset to adapt its behavior to a specific task or domain.

Finite element analysis (FEA): A numerical technique that divides a complex structure into small elements to simulate stress, heat or other physical behavior.

FinOps: An operating model for managing cloud spending collaboratively across engineering, finance, and business teams.

First contact resolution (FCR): The share of customer support issues resolved in a single interaction, with no follow-up or escalation needed.

First-click attribution: An attribution model that gives all credit for a conversion to the very first marketing touchpoint that introduced the customer.

First-party data: Data a company collects directly from its own customers and properties, rather than acquiring it from outside sources.

First-pass yield (FPY): The percentage of units that come out of a process correctly on the first attempt, with no rework or scrap.

Fiscal transparency: The degree to which a government publicly discloses how it raises, spends, and manages public money.

Fleet telematics: Collecting real-time vehicle data, location, speed, fuel use, and diagnostics, from a fleet's onboard systems and sensors.

Focus time: Time spent working on a task without interruption, tracked as a measure of capacity for concentrated work.

Follower growth rate: The percentage change in an account's follower count over a period, used to track audience-building momentum on social platforms.

Footfall analytics: Measuring and analyzing how many people enter, move through, and dwell in a physical space, most often a retail store.

Forward points: The difference between a currency's spot and forward exchange rate, driven by the interest rate gap between the two currencies.

FP&A (financial planning and analysis): The finance function responsible for budgeting, forecasting and analyzing company performance against plan.

Fraud detection: Identifying transactions, claims, or behavior that are likely fraudulent, using rules, statistics, or machine learning.

Frequency capping: Limiting the number of times a single user is shown the same ad within a defined time period.

Fundamental analysis: Valuing a security from the underlying business's financial health and economic conditions, not its price history.

Fundraising ROI (return on investment): The ratio of funds raised to the cost of raising them, used to compare efficiency across fundraising campaigns.

Funnel analysis: A method for tracking how users move through an ordered sequence of steps toward a goal, and where they drop off.

Futures curve: The set of prices for futures contracts on the same underlying asset across different expiration dates, plotted together.

G

Game economy balancing: Tuning a game's currencies, rewards and costs so progression, spending and player behavior stay in intended proportion.

Gas fees: The transaction fees paid to a blockchain network to process and validate an operation, priced by network demand.

General Data Protection Regulation (GDPR): The European Union's comprehensive data protection law governing how personal data of people in the EU is collected and used.

Generative AI: AI systems that create new content, text, images, audio, or code, rather than only classifying or predicting a label.

Generative engine optimization (GEO): Optimizing content to be cited or summarized by AI-generated answers, rather than only to rank in a list of links.

Genomics analysis: Computational analysis of DNA sequence data to identify genes, variants and patterns linked to traits or disease.

Geo experiment: An experiment that randomizes treatment by geographic region rather than by individual user, often to measure marketing's incremental impact.

Geocoding: Converting a text address or place name into geographic coordinates, or the reverse, for mapping and analysis.

Geofencing: Drawing a virtual boundary around a location so entry or exit can trigger an action or be logged as an event.

Geographic information system (GIS): Software for capturing, storing, analyzing and displaying data tied to specific locations on the Earth's surface.

Geospatial intelligence (GEOINT): Intelligence built from imagery, mapping, and geographic data to describe and assess activity tied to a place.

Get-out-the-vote (GOTV) analytics: Analytics applied to voter mobilization programs, measuring and optimizing contact efforts that increase turnout.

Gini coefficient: A single number, from 0 to 1, measuring income or wealth inequality within a population.

Goal tracking: Configuring an analytics tool to record when a visitor completes an action the business treats as valuable.

Gradient boosting: An ensemble method that builds trees one at a time, each one correcting the errors made by the trees before it.

Graph analytics: Analyzing data modeled as nodes and edges to study connections, structure, and influence rather than isolated records.

Greens in regulation (GIR): The percentage of holes where a golfer reaches the green in par minus two strokes, leaving a putt for par or better.

Grid analytics: Analyzing sensor and meter data from the electricity grid to balance load, detect faults, and plan investment.

Gross margin: The share of revenue left after subtracting the direct cost of producing or delivering what was sold.

Gross merchandise value (GMV): The total value of goods sold through a platform over a period, before fees, discounts, and returns are deducted.

Gross Rating Point (GRP): A measure of total advertising exposure, calculated as reach multiplied by frequency, expressed against a target population.

Gross revenue retention (GRR): The percentage of recurring revenue kept from an existing customer base after churn and downgrades, excluding any expansion.

Guardrail metric: A secondary metric watched during an experiment to catch harm the primary metric would not show, such as load time or churn.

Guardrails: Rules, filters, or checks placed around a language model to keep its inputs and outputs within acceptable, safe bounds.

H

H-index: A researcher's largest number h such that h of their papers have each been cited at least h times.

Habitat fragmentation: The breaking up of a continuous habitat into smaller, more isolated patches, typically by roads, farmland or development.

Hallucination: When a language model generates confident, fluent output that is factually wrong or unsupported by its source material.

Headcount planning: Forecasting how many people a company will hire, lose, and pay for across a plan period, by team and role.

Health economics and outcomes research (HEOR): Research that measures the clinical, economic, and patient-reported value of a treatment or health intervention.

Health Insurance Portability and Accountability Act (HIPAA): The U.S. law that sets privacy and security rules for protected health information handled by healthcare providers and insurers.

Heart rate variability (HRV): The variation in time between consecutive heartbeats, used as an indicator of recovery and nervous-system stress.

Heatmap: A visual overlay of a webpage showing aggregated visitor clicks, cursor movement or scrolling using color intensity.

Heterogeneous treatment effects (HTE): Differences in a treatment's true effect across subgroups, showing an intervention can help some segments and harm or not affect others.

High-performance computing (HPC): Using clusters of powerful, tightly linked computers to run computations far larger than a single machine can handle.

Historical GIS: Applying geographic information systems to historical data to map how places, boundaries and populations changed over time.

Holdout group: A segment of users deliberately kept out of a treatment, campaign or feature, so its true incremental impact can be measured against them.

Hospital readmission rate: The share of discharged patients who are readmitted to a hospital within a set window, usually 30 days.

Hotspot analysis: Statistical method for finding geographic clusters where an event, such as crime or disease, occurs more than chance predicts.

House price index (HPI): An index tracking how home prices change over time using repeat sales of the same properties.

Housing affordability index: An index measuring whether a typical household can afford a typical home in a given housing market.

Hub-and-spoke analytics model: An operating model that pairs a central standards-setting analytics team with analysts embedded in business units.

Human Development Index (HDI): A composite index combining life expectancy, education, and income to rank and compare countries' overall development.

Hyperparameter tuning: The process of finding the settings that control how a model learns, such as learning rate or tree depth, rather than what it learns.

Hypothesis testing: A statistical framework for deciding whether observed data provides enough evidence to reject a default assumption.

I

Idempotency: The property where running an operation multiple times leaves the same result as running it once.

Identity resolution: Matching data from different sources and devices to a single underlying person or household despite no shared identifier.

Identity stitching: Linking the different identifiers a single person generates across devices, sessions, and platforms into one profile.

Impact evaluation: Measuring a program's causal effect on outcomes by comparing what happened to a credible counterfactual.

Implied probability: The probability of an outcome derived by converting betting odds, decimal, fractional or American, into a percentage.

Implied volatility (IV): The volatility level that, plugged into an options pricing model, reproduces the option's current market price.

Impression: One instance of an ad being served and rendered on a page or app, whether or not anyone noticed or clicked it.

In-app purchase (IAP): A purchase of digital content, currency, or a subscription made from within a mobile app.

In-memory database: A database that keeps its primary copy of data in RAM instead of on disk, trading persistence for low latency.

Incremental model: A transformation that updates only new or changed rows on each run instead of rebuilding the entire table.

Incrementality testing: An experiment measuring the true causal lift of a marketing activity by comparing it against a held-out control.

Indicator of compromise (IOC): A piece of forensic evidence, such as a file hash or IP address, suggesting a system has been breached or contacted by an attacker.

Influencer marketing: Paying or partnering with individuals who have an engaged social media following to promote a product to their audience.

Install attribution: Determining which marketing channel, campaign, or ad is responsible for a given app install.

Instrumental variables (IV): A causal-inference technique that uses a variable affecting the outcome only through treatment, to work around unmeasured confounding.

Internal rate of return (IRR): The discount rate at which an investment's projected cash flows have a net present value of zero.

Interquartile range: The range covering the middle 50% of a dataset, from the 25th to the 75th percentile, used to measure spread.

Invalid traffic (IVT): Ad impressions, clicks, or conversions that do not represent genuine human interest, from bots, click farms, or tracking errors.

Inventory turnover: How many times a company sells and replaces its inventory over a given period, usually a year.

J

Journal impact factor (JIF): A measure of how often a journal's articles are cited on average, used as a proxy for the journal's prestige.

K

K-means clustering: A clustering algorithm that partitions data into k groups by repeatedly assigning points to the nearest of k cluster centers.

Keyword research: Identifying and evaluating the search terms an audience uses, to guide what content to create and how to target it.

Know your customer (KYC): Verifying a customer's identity and assessing their risk before and throughout a business relationship.

Knowledge graph: A network of entities and the labeled relationships between them, used to represent and query structured knowledge.

KPI (key performance indicator): A measurable value tracked over time to show whether a team or process is on track toward a goal.

L

Labor force participation rate (LFPR): The share of the working-age population that is either employed or actively looking for work.

Laboratory information management system (LIMS): Software that tracks samples, tests and results through a laboratory's workflow, from intake to reporting.

Land-use modeling: Simulating how a region's land will be developed or change use over time, based on zoning, transport, and economic factors.

Landing page: The first page a visitor sees in a session, used in analytics as an entry-point performance dimension.

Lapsed donor: A donor who gave in the past but has not made a gift within an organization's defined recency window.

Large language model (LLM): A neural network trained on vast amounts of text to predict and generate language, the technology behind most modern AI assistants.

Last-click attribution: An attribution model that gives all credit for a conversion to the final marketing touchpoint before it happened.

Late-arriving data: Records that reach a pipeline after the time window they logically belong to has already been processed.

Launch angle: The vertical angle, in degrees, at which a batted baseball leaves the bat relative to the ground.

Lead scoring: A method of ranking leads by assigning points for behaviors and attributes that correlate with a likelihood to buy.

Lead time: The total time between triggering an order or task and having it completed and available for use.

Lead time for changes: The time from when code is committed to when it is running successfully in production, one of the four DORA metrics.

Lean analytics: Applying lean management's focus on flow and waste reduction to choosing and acting on the one metric that matters most right now.

Learning analytics: Collecting and analyzing data on learners and their interactions to understand and improve learning and the environments it happens in.

Life cycle assessment (LCA): A method for quantifying a product's environmental impact across every stage from raw materials to disposal.

Life expectancy: The average number of years a person could expect to live given current age-specific mortality rates.

Light curve: A graph of an astronomical object's brightness measured repeatedly over time, used to detect transits, flares or pulsation.

Likely voter model: A statistical filter pollsters apply to a survey sample to estimate results among the people who will actually vote.

Likert scale: An ordinal rating scale, typically five or seven points, running between two opposite attitudes or opinions.

Line movement: How a betting line or odds change over time between when a market opens and when the event begins.

Linear programming (LP): An optimization method that finds the best outcome for a problem whose objective and constraints are all linear relationships.

Link prediction: Estimating the likelihood that a connection exists, or will form, between two nodes in a graph that are not currently linked.

Litigation analytics: Using data on past case outcomes, judges, and courts to inform legal strategy, such as predicting how a case might proceed.

Live operations (live ops): The ongoing management of a live game after launch through events, content updates, and tuning based on player data.

LLM evaluation: Systematically testing a language model's outputs for quality, accuracy, safety, and consistency before and after deployment.

LLM observability: Monitoring a language model's live behavior in production, tracking its inputs, outputs, latency, cost, and failure patterns.

LLM-as-a-judge: Using a language model to score or compare the outputs of another model, as a scalable substitute for human review.

Load forecasting: Predicting future electricity demand so utilities can plan generation, purchasing, and grid operations.

Load management: The practice of proactively resting or limiting an athlete's workload, guided by data, to reduce the risk of injury.

Log aggregation: Collecting log data from many separate systems into one centralized, searchable store instead of checking each system individually.

Logic model: A diagram mapping how a program's resources and activities are expected to lead to short- and long-term outcomes.

Logistic regression: A statistical model that estimates the probability of a binary outcome as a function of input variables.

Lookalike audience: A group of new prospects an ad platform selects because their profile statistically resembles an advertiser's existing customers.

Loss given default (LGD): The share of an exposure a lender expects to lose if a borrower actually defaults, after recoveries.

Loss ratio: The share of insurance premium collected that is paid out in claims, a core measure of underwriting profitability.

LTV:CAC ratio (customer lifetime value to customer acquisition cost ratio): The ratio of what a customer is worth over their lifetime to what it costs to acquire them.

M

Machine learning: Building models that learn patterns from data to make predictions or decisions, rather than following hand-coded rules.

Margin of error: The range added to and subtracted from a survey estimate to express the uncertainty from sampling only part of a population.

Markdown optimization: Using data and models to decide when, how much, and on which products to cut price so excess inventory clears at the best return.

Market basket analysis: Finding which products are frequently bought together, to inform bundling, placement, and recommendations.

Market share: The percentage of total sales in a market or category captured by a specific company, product, or brand.

Marketing efficiency ratio (MER): Total revenue divided by total marketing spend, a single blended measure of overall marketing performance.

Marketing mix modeling: A statistical technique estimating how much each marketing channel and external factor contributes to results.

Marketing qualified lead (MQL): A lead that marketing activity has judged likely enough to convert that it gets handed to sales for follow-up.

Massively parallel processing: An architecture that splits a query across many independent compute nodes, each processing its own data slice.

Master data management (MDM): The discipline of maintaining one authoritative, consistent record for core entities like customers or products.

Materialized view: A query result stored physically and refreshed on a schedule or trigger, rather than recomputed on every read.

Maximum drawdown (MDD): The largest peak-to-trough decline in an investment's value before it reaches a new high.

Mean time between failures (MTBF): The average time a repairable system operates between one failure and the next, used to measure reliability.

Mean time to detect (MTTD): The average time between when an incident begins and when a team becomes aware of it.

Mean time to resolve (MTTR): The average time from when an incident is detected to when it is fully resolved and normal service is restored.

Media rights valuation: Estimating what a sports league or event's broadcast and streaming rights are worth to networks and platforms.

Median: The middle value of a dataset when sorted from lowest to highest, splitting it into two equal halves.

Median household income: The income level at which half of households earn more and half earn less, in a given year and place.

Message broker: Middleware that routes messages between producing and consuming systems so they don't need to connect to each other directly.

Metadata management: The discipline of capturing, organizing, and maintaining data about data so it can be found, trusted, and used consistently.

Metric cardinality: The number of unique time series a metric produces, driven by every distinct combination of its labels or tags.

Micro-batching: Processing data in very small, frequent batches to approximate streaming without a dedicated stream processor.

Microtargeting: Using granular voter or consumer data to reach narrowly defined audience segments with tailored messages.

Migration flows: Measured movement of people into and out of a geographic area, used to calculate net migration.

Minimum detectable effect (MDE): The smallest true effect an experiment is designed to reliably distinguish from no effect, given its sample size and power.

MITRE ATT&CK framework: A public knowledge base of adversary tactics and techniques used as a shared vocabulary for describing attacker behavior.

MLOps (machine learning operations): The set of practices for deploying, monitoring and maintaining machine learning models reliably in production, extending DevOps to ML systems.

Mobile measurement partner (MMP): A third-party platform that attributes app installs and in-app events to the marketing source that drove them.

Model monitoring: Ongoing tracking of a deployed model's inputs, predictions and outcomes to catch accuracy loss, drift or failures after launch.

Model risk management (MRM): The governance practice of validating, monitoring, and controlling the risk that a model is wrong or misused.

Molecular dynamics (MD): Simulating the physical movement of atoms and molecules over time to study how a molecular system behaves.

Monetization funnel: The sequence of steps from installing a game to becoming a paying player, used to find where potential spenders drop off.

Money-weighted return (MWR): A return measure that accounts for the size and timing of an investor's own deposits and withdrawals, like an internal rate of return.

Monitoring and evaluation (M&E): The ongoing tracking of a program's activities plus periodic assessment of whether it achieved its intended outcomes.

Monte Carlo simulation: A technique that models uncertainty by running a scenario many times with randomly sampled inputs to see the range of outcomes.

Monthly recurring revenue (MRR): The value of a subscription business's recurring revenue, normalized to a one-month run rate.

Motorsport telemetry: Real-time sensor data streamed from a race car to the pit wall, covering speed, tyres, brakes, and engine performance.

Moving average: A calculated series that averages a fixed window of recent data points, smoothing out short-term noise to reveal a trend.

Multi-armed bandit (MAB): An adaptive experiment that shifts traffic toward better-performing variants in real time, instead of splitting it evenly until a fixed end date.

Multi-touch attribution: An attribution method that splits conversion credit across every touchpoint in a customer's path, not just one.

Multilevel regression and poststratification (MRP): A method for estimating opinion in small geographic or demographic groups from a national survey by modeling and reweighting.

Multiple comparisons problem: The increased risk of a false positive that comes from running many statistical tests and highlighting any one significant result.

Multiple on invested capital (MOIC): The total value a private investment has returned, divided by the total capital invested, without adjusting for the time it took.

Multivariate testing (MVT): An experiment that varies several page or product elements at once, to measure each element's individual effect and how elements interact.

MVRV ratio (market value to realized value): A crypto valuation metric comparing an asset's current market value to the value at which coins last moved on-chain.

N

Named Entity Recognition (NER): An NLP technique that locates and classifies names of people, organizations, places, and other entities within text.

Natural language query: Asking a data question in plain language and getting an answer without writing code or a query language.

Net interest margin (NIM): A bank's net interest income as a share of its interest-earning assets, a core measure of lending profitability.

Net Promoter Score (NPS): A single-question survey score that gauges customer loyalty by asking how likely someone is to recommend a company.

Net revenue retention (NRR): The percentage of recurring revenue kept from an existing customer base, including expansion and after subtracting churn.

Net sentiment: The share of positive mentions minus the share of negative mentions in a set of analyzed text, summarized as one score.

Net worth tracking: Periodically totaling personal assets minus liabilities to see how one's overall financial position changes over time.

NetFlow analysis: Examining summarized records of network traffic exported by routers and switches to understand traffic patterns.

Network churn prediction: Using network performance and usage data to predict which telecom subscribers are likely to cancel their service.

Network latency: The time it takes data to travel from a source to a destination across a network, usually measured as round-trip time.

Network utilization: The share of a network's total available capacity that is actually being used over a given period.

Neural network: A model made of layers of connected nodes that learns to map inputs to outputs by adjusting connection weights during training.

Next-best-action (NBA): A system or process that recommends the single most valuable action to take for a specific customer at a specific moment.

Nonresponse bias: Distortion in survey results that occurs when people who decline to participate differ systematically from those who respond.

Normal distribution: A symmetric, bell-shaped probability distribution defined entirely by its mean and standard deviation.

Normalization: Structuring a database so each fact is stored in exactly one place, reducing redundancy and inconsistency.

Normalized difference vegetation index (NDVI): A satellite-derived index measuring live vegetation density and health from how plants reflect red and infrared light.

North star metric: The single metric a team chooses as the best proxy for the core value a product delivers to customers.

Novelty effect: A temporary spike in engagement with a new feature or design that fades once users stop reacting to its newness and it becomes routine.

Nowcasting: Estimating the current or very near-term value of an economic indicator before official data is released.

Null hypothesis: The default assumption in a statistical test, typically that there is no effect or no difference, until evidence says otherwise.

Numerical weather prediction (NWP): Forecasting weather by solving the physical equations governing the atmosphere on a computer, using current observations as a start.

O

Occupancy rate: The share of available capacity, such as hotel rooms or rental units, that is filled or in use over a given period.

Ocean color remote sensing: Measuring the color of reflected sunlight from the ocean surface by satellite to estimate chlorophyll, sediment and water quality.

OKR (objectives and key results): A goal-setting framework that pairs one qualitative objective with a small set of measurable key results.

OLAP (online analytical processing): A category of technology optimized for fast, multi-dimensional aggregation and slicing of data for analysis.

On-base plus slugging (OPS): A baseball hitting stat adding on-base percentage and slugging percentage into a single number summarizing overall production.

On-chain analytics: Analyzing transaction and wallet data recorded directly on a public blockchain to study crypto asset activity.

On-time in full (OTIF): The share of customer orders delivered both on the promised date and complete, with nothing missing or short.

On-time performance (OTP): The share of trips, flights, deliveries, or services completed within a defined acceptable window of their scheduled time.

One big table: A denormalized modeling approach that joins facts and dimensions into a single wide table instead of many.

Open access: Making research publications freely available to read online, without a subscription or paywall.

Open data: Government or institutional data published in reusable formats, free for anyone to access, use, and redistribute.

Open data portal: A public website where a government publishes machine-readable datasets for citizens, researchers, and developers to reuse.

Open table format: A specification that adds transactions, schema management, and versioning to files stored in a data lake.

Open-core model: A business model where a product's core is open-source and free, while advanced features are sold as a proprietary, paid layer on top.

Open-source intelligence (OSINT): Intelligence produced from publicly available information — social media, news, public records — rather than classified sources.

Optimization modeling: Using mathematical models to find the best possible decision under a defined set of constraints and an objective to optimize.

Option Greeks: A set of measures showing how an option's price responds to changes in the underlying, volatility, time, and rates.

Order book: A real-time list of all outstanding buy and sell orders for an asset, ranked by price.

Organic traffic: Visits arriving from unpaid listings on a search results page, as distinct from paid search, social or direct visits.

Organizational network analysis (ONA): Mapping and analyzing the informal networks of communication and collaboration inside an organization.

Origin-destination (O-D) data: Records of where trips start and end, used to understand travel demand across a transportation network.

Outlier: An observation that differs markedly from the rest of the dataset, either from genuine variation or from an error.

Overall equipment effectiveness (OEE): A composite manufacturing metric combining availability, performance, and quality into one score of how well equipment is used.

Overall evaluation criterion (OEC): The single metric or composite formula a team agrees in advance will determine whether an experiment's treatment is judged a success.

Overfitting: When a model learns the noise and specific quirks of its training data so closely that it performs poorly on new data.

Overround: The built-in margin a bookmaker adds so a market's implied probabilities add up to more than 100%.

P

P-value: The probability of observing data at least as extreme as the result, if the null hypothesis were actually true.

Packet loss: The percentage of data packets sent across a network that never arrive at their destination.

Page load time: How long a webpage takes to become usable to a visitor, measured at several different possible points.

PageRank: An algorithm that scores a node's importance by the number and importance of other nodes linking to it.

Pages per session: The average number of pages a visitor views during a single session on a site.

Pageview: One load or reload of a web page recorded by an analytics tool, the most basic unit of web traffic measurement.

Pareto analysis: A technique that ranks causes or categories by their contribution to a problem, to focus effort on the small number responsible for most of it.

Participatory budgeting (PB): A process where community members directly decide how to spend a defined portion of a public budget.

Particulate matter (PM2.5 and PM10): Airborne solid and liquid particles small enough to be inhaled, measured by size and used as a core air-quality indicator.

Partitioning: Physically splitting a table's data into segments, usually by a column like date, so queries can skip irrelevant ones.

Passes allowed per defensive action (PPDA): A soccer pressing-intensity stat measuring how many passes the opposition completes before a team makes a tackle or interception.

Path analysis: Examining the sequences of actions users take through a product to find common, high-value, or drop-off routes.

Pattern-of-life analysis: Studying the routine movements and behavior of a person, vehicle, or location over time to detect meaningful deviations.

Pay equity analysis: A statistical review of compensation data to detect unexplained pay gaps by gender, race or other protected characteristics.

Payment success rate: The share of attempted payments that complete successfully, from checkout initiation through to settlement.

Paywall conversion rate: The share of visitors who hit a paywall and go on to subscribe or pay for access, a core metric for publishers.

PDO: A hockey stat adding on-ice shooting percentage and save percentage, read as a rough indicator of how much puck luck a team is riding.

People analytics: Applying data analysis to workforce data such as hiring, performance and attrition to guide HR decisions.

Percentile: The value below which a given percentage of observations in a dataset fall.

Performance budgeting: Allocating public funds based on measured program outcomes and performance targets rather than prior spending levels.

Performance management in government: Setting measurable goals for public agencies and tracking results against them to guide budgeting and accountability.

Personal KPI (key performance indicator): A metric an individual defines and tracks over time to monitor progress toward their own goals.

Personally identifiable information (PII): Any data that can identify a specific individual, alone or combined with other information, and is subject to privacy regulation.

Persuasion modeling: Predicting which voters are movable on a candidate or issue, as distinct from how likely they are to vote.

Pharmacovigilance: Detecting, assessing, and monitoring adverse effects of a drug throughout its use, especially after approval.

Photometry: Measuring the brightness or intensity of light from an object, most commonly a star or other astronomical source.

Phylogenetic analysis: Inferring the evolutionary relationships between organisms or gene sequences and representing them as a branching tree.

Physical climate risk: The risk to assets, operations or people from climate-related physical hazards, both sudden events and long-term shifts.

Pipeline coverage: The ratio of open sales pipeline value to a revenue target, used to judge whether enough deals exist to hit it.

Planogram: A diagram specifying exactly where each product goes on a shelf or fixture, used to plan and audit retail layouts.

Player efficiency rating (PER): A per-minute basketball rating that condenses a player's box-score production into one pace-adjusted number.

Player retention: The share of players who return to a game after a given number of days since their first session, tracked at fixed intervals like day 1, 7, and 30.

Player segmentation: Grouping players by behavior, spend, or engagement to target content, offers, and communication differently for each group.

Player tracking data: High-frequency positional data recording where players and the ball move throughout play, not just discrete events.

Player valuation model: A quantitative model estimating a player's on-field or market value by combining performance data with context.

Plus-minus (+/-): The net difference in points or goals scored by a team versus its opponent while one specific player is on the field, court or ice.

Podcast downloads: The number of times a podcast episode's audio file has been requested from a hosting server, the industry's standard audience metric.

Point-in-time data: Data reconstructed exactly as it appeared on a given historical date, including figures that were later revised.

Poll aggregation: Combining multiple individual polls, weighted by quality and recency, into a single averaged estimate of opinion.

Population health management (PHM): Using aggregated patient data to improve health outcomes and control cost across a defined group of people, not just individuals.

Population pyramid: A chart showing a population's age and sex distribution, whose shape reveals growth, aging, or decline.

Portfolio rebalancing: Buying and selling holdings to bring a portfolio's asset weights back to its target allocation after they drift.

Possession value model: A framework that assigns a value to every on-ball action in a possession based on how much it changes scoring probability.

Poverty headcount ratio: The share of a population living below a defined poverty line, most often expressed as a percentage.

Precision agriculture: Using field-level data, from sensors, satellites, and GPS-guided equipment, to tailor farming inputs to specific conditions rather than whole fields.

Precision and recall: Two complementary classifier metrics; precision is how many predicted positives were correct, recall is how many actual positives were found.

Predictive analytics: Analytics that uses historical data and statistical models to estimate what is likely to happen next.

Predictive maintenance: Using sensor data and models to predict when equipment is likely to fail so it can be serviced before it breaks down.

Predictive policing: Using historical crime data and statistical models to forecast where or by whom future crime is likely, to guide patrols.

Preregistration: Publicly documenting a study's hypotheses, methods and analysis plan before collecting or examining the data.

Prescriptive analytics: Analytics that recommends specific actions by combining predictions with optimization or decision rules.

Price elasticity: How much the quantity demanded of a product changes in response to a change in its price.

Price-to-earnings ratio (P/E ratio): A valuation ratio comparing a company's share price to its earnings per share.

Probability of default (PD): The estimated likelihood that a borrower fails to make required payments on a debt within a given time horizon.

Probable maximum loss (PML): The largest loss an insurer or asset owner expects from a single severe event, at a specified probability threshold.

Process mining: Reconstructing how a business process actually runs from system event logs rather than from how it is documented.

Procurement analytics (government): Analyzing public purchasing data to control spend, detect irregularities, and evaluate vendor and contract performance.

Product affinity: A measure of how strongly two or more products relate to each other in customer interest or purchase behavior.

Product-led growth (PLG): A go-to-market approach where the product itself, often via free trials or freemium use, drives acquisition and expansion.

Product-qualified lead (PQL): A free or trial user whose in-product behavior signals they are ready to be approached about a paid plan.

Program evaluation: Systematic assessment of whether a public or nonprofit program achieves its intended outcomes, and why.

Programmatic advertising: The automated, auction-based buying and selling of digital ad inventory using software instead of manual insertion orders.

Progressive passes: Completed soccer passes that advance the ball a meaningful distance toward the opponent's goal, used to gauge ball progression.

Prompt engineering: Designing and refining the instructions given to a language model to reliably get the output you want.

Proof of concept (PoC): A small-scale, time-boxed test of whether a proposed tool or approach actually works for a specific use case before wider investment.

Propensity model: A predictive model that scores how likely an individual is to take a specific action, such as buying, churning, or clicking.

Propensity score matching (PSM): A method that pairs treated and untreated units with similar odds of having received treatment, to approximate a randomized comparison.

Proteomics: The large-scale study of the proteins present in a cell, tissue or organism, including their identity and abundance.

Pseudonymization: Replacing identifying data with a reversible token, keeping the mapping back to real identities separate and restricted.

Public market equivalent (PME): A method comparing a private fund's returns to what the same cash flows would have earned invested in a public market index.

Purchasing managers' index (PMI): A survey-based index summarizing whether business activity in manufacturing or services is expanding.

Purchasing power parity (PPP): A theory that exchange rates should adjust so identical goods cost the same amount across countries once converted to a common currency.

Pythagorean expectation: A formula estimating a team's expected win percentage from the points or runs it scores and allows, rather than its actual record.

Q

Quality of experience (QoE): A subjective measure of how satisfactory a network or service feels to the user, distinct from raw technical performance.

Quality of hire: A composite measure of how well new hires perform and stay, used to judge recruiting effectiveness beyond speed and cost.

Quantified self: The practice of using wearables, apps and personal logs to collect data about oneself in order to understand one's own habits.

Quantitative trading: Trading strategies built on mathematical models and statistical analysis rather than discretionary judgment.

Query engine: The software component that parses, plans, and executes queries against one or more data sources.

Query optimization: The process, largely automated by the database, of choosing the fastest execution plan for a given query.

Questionnaire design: Writing and structuring survey questions and response options to minimize bias and measure reliably.

Quota attainment: The percentage of an assigned sales quota that a rep, team or organization actually achieved in a period.

R

Race pace analysis: Comparing drivers' lap times across a stint, adjusted for fuel load and tyre age, to estimate true race pace.

Random forest: An ensemble model that trains many decision trees on random subsets of data and features, then averages their predictions.

Randomized controlled trial (RCT): An experiment that randomly assigns subjects to treatment and control groups to isolate a causal effect.

Reach and frequency: Reach is how many unique people saw an ad; frequency is how many times, on average, each of them saw it.

Real effective exchange rate (REER): A currency's exchange rate against a trade-weighted basket of other currencies, adjusted for relative inflation.

Real-time OLAP: An analytical database that ingests streaming data continuously and answers aggregation queries with sub-second latency.

Real-world evidence (RWE): Clinical evidence about how a treatment performs, generated from real-world data instead of a controlled trial.

Reanalysis data: A consistent, gridded historical weather record built by rerunning a modern forecast model over past observations.

Recidivism rate: The share of people released from incarceration or supervision who reoffend or are rearrested within a set period.

Recommendation system: A system that predicts and ranks the items a specific user is most likely to want, such as products, content, or actions.

Record linkage: Matching records that refer to the same person or entity across datasets that lack a shared unique identifier.

Referral traffic: Visits that arrive by clicking a link on another website, other than a search engine, paid ad or social network.

Regression analysis: A statistical method for estimating the relationship between one outcome variable and one or more predictor variables.

Regression discontinuity design (RDD): A causal method that compares outcomes just above and below a cutoff, such as a score threshold, that determines who receives a treatment.

Regression to the mean (in sports): The statistical tendency for unusually good or bad performances to move back toward a player's or team's underlying average.

Reinforcement learning: A learning approach where an agent learns a strategy by taking actions in an environment and receiving rewards or penalties over time.

Remote sensing: Collecting data about the Earth's surface or atmosphere from satellites, aircraft or drones without physical contact.

Rent burden: The share of a renter household's income spent on rent, commonly flagged as burdened above 30 percent.

Repeat purchase rate: The share of customers who buy more than once within a given period, a core signal of loyalty.

Reproducibility: The ability of an independent researcher to obtain the same results using the original data, code and methods.

Request for proposal (RFP): A formal document an organization sends to vendors, describing requirements and asking for a structured proposal, price, and timeline.

Research data management (RDM): The practices and policies for organizing, storing, documenting and preserving data throughout a research project.

Research software engineering: Applying professional software engineering practice to code written for scientific research, not just quick one-off scripts.

Response rate: The share of people selected for a survey who actually complete it, a key indicator of potential nonresponse bias.

Responsible AI: A set of practices for developing and deploying AI systems that are fair, transparent, safe, and accountable to the people they affect.

Resting heart rate (RHR): The number of times the heart beats per minute while a person is at complete rest, a common fitness baseline.

Retail media network (RMN): An advertising platform run by a retailer that sells ad placements on its own site, app, or in-store screens using its shopper data.

Retargeting: Showing ads to people who have already interacted with a brand's website, app, or content, based on tracked prior behavior.

Retention analysis: The study of how many users keep returning to and engaging with a product after their first use.

Retrieval-augmented generation (RAG): An architecture that retrieves relevant documents at query time and feeds them to a language model so its answers are grounded in sources.

Return on ad spend: A ratio of revenue earned for every unit of currency spent on advertising.

Return period: The average time expected between two events of a given severity, such as a 100-year flood, based on historical frequency.

Revenue management: The discipline of selling the right unit of fixed, perishable capacity to the right customer at the right price and time.

Revenue operations (RevOps): The function that aligns sales, marketing and customer success operations, data and systems around one revenue process.

Revenue per available room (RevPAR): A hotel performance metric combining occupancy and rate into a single figure of revenue earned per available room.

Revenue per employee: A productivity ratio, total revenue divided by headcount, used to benchmark organizational efficiency.

Revenue recognition: The accounting rules for when revenue counts as earned and can be recorded, which is often not when cash is received.

Reverse ETL: The practice of syncing modeled data from a warehouse back out into operational tools like a CRM or ad platform.

RFM analysis: A segmentation technique that scores customers on how recently, how often and how much they buy.

Risk adjustment: Statistically normalizing outcome or cost comparisons to account for how sick, old, or otherwise high-risk a population is.

Risk assessment instrument (RAI): A statistical tool that scores a person's likelihood of reoffending or failing to appear, used in bail, sentencing, or parole.

Risk profiling: Assessing an investor's ability and willingness to tolerate investment losses, used to guide asset allocation decisions.

Risk scoring: Assigning a numeric score to a customer, transaction, or entity that estimates the likelihood of a specific negative outcome.

ROC AUC (area under the ROC curve): A single number summarizing how well a classifier separates positive from negative cases across every possible decision threshold.

Role-based access control (RBAC): Granting data or system permissions based on a user's assigned role rather than configuring access for each person individually.

Rolling forecast: A financial forecast that is continuously extended, adding a new future period as each period closes.

Root cause analysis (RCA): A structured process for tracing a problem back through its contributing factors to the underlying cause, not just its symptoms.

Route optimization: Computing the most efficient sequence of stops or paths for vehicles, usually to minimize distance, time, or cost.

Row-level security (RLS): Restricting which rows of a table a user can query, rather than granting or denying access to the whole table.

Rule of 40: A SaaS benchmark stating that growth rate plus profit margin should add up to 40% or more.

S

Sabermetrics: The empirical, statistics-driven study of baseball performance, valuing measurable production over traditional scouting judgment.

Safety stock: Extra inventory held beyond expected demand to absorb variability in demand or supply lead time without running out.

Salary cap analytics: Using data and modeling to build a competitive roster within a league's fixed limit on total player salaries.

Sales and operations planning (S&OP): A recurring cross-functional process that aligns sales forecasts, production plans, and financial targets into one plan.

Sales cycle length: The average time from an opportunity being created to a deal closing, either won or lost.

Sales forecasting: Predicting future sales volume or revenue from pipeline data, historical trends or statistical models.

Sales velocity: A single measure of how fast a sales team generates revenue, combining pipeline volume, deal size, win rate and cycle time.

Same-store sales (comparable-store sales, comps): Revenue growth measured only at locations that were open throughout both the current and prior comparison periods.

Sample ratio mismatch (SRM): A significant gap between an experiment's actual traffic split and its intended split, signaling a bug rather than a real treatment effect.

Sample size: The number of observations in a study or test, which determines how precisely and reliably an effect can be measured.

Sampling bias: A systematic distortion that arises when the method used to select a sample favors some members of the population over others.

Sampling frame: The list or source, such as a voter file or address database, from which a survey's respondents are actually drawn.

Sampling in analytics: Computing a report from a subset of the underlying data rather than the complete data set, to return results faster.

Satellite imagery analysis: Extracting information such as land cover, change or object detection from images captured by orbiting satellites.

Savings rate: The share of income that is saved or invested rather than spent, over a given period.

Scenario modeling: Building several named versions of a financial plan, such as best, base, and worst case, under different assumptions.

Scenario planning: A planning method that builds and stress-tests a small number of distinct, plausible futures rather than a single forecast.

Schema evolution: The ability of a table to accommodate structural changes, like an added or renamed column, without breaking data.

Schema-on-read: Storing data in its raw form and applying structure only at query time, instead of when it is first loaded.

Scientific workflow: An automated, documented sequence of computational steps that takes research data from raw input to final result.

Scope 1, 2 and 3 emissions: The GHG Protocol's three-tier system for classifying a company's direct and indirect greenhouse gas emissions.

Scroll depth: How far down a page a visitor scrolls, tracked as a percentage or discrete thresholds to measure content exposure.

Sea surface temperature (SST): The temperature of ocean water at or near the surface, a core input to weather, climate and marine-ecosystem models.

Search engine results page (SERP): The page a search engine returns after a query, containing organic listings alongside ads, snippets and other features.

Search intent: The underlying goal behind a search query, commonly classified as informational, navigational, commercial or transactional.

Seasonality: A regular, predictable pattern in data that repeats at a fixed calendar interval, such as daily, weekly, or annually.

Secondary ticket market: The resale marketplace where tickets already sold by a team or venue trade again, priced by supply and demand.

Security operations center (SOC): The team responsible for continuously monitoring an organization's systems, triaging alerts, and coordinating incident response.

Selection bias: A distortion that occurs when the sample analyzed is not representative of the population it is meant to describe.

Self-service BI: BI tooling that lets business users build their own reports and explore data without waiting on IT or analytics teams.

Sell-through rate: The share of stocked inventory that actually sells within a given period, usually expressed as a percentage.

Semantic layer: A layer that defines business metrics and dimensions once, centrally, so every tool and report uses the same numbers.

Semantic search: Search that matches results to a query's meaning rather than its exact keywords, using vector similarity between embeddings.

Sensor calibration: Adjusting a sensor's readings against a known reference so its measurements are accurate, not just precise.

Sentiment analysis: Using NLP to classify text as positive, negative, or neutral, so large volumes of comments or reviews can be scored automatically.

Sequence alignment: Arranging DNA, RNA or protein sequences to identify regions of similarity that may reflect shared function or ancestry.

Sequential testing: An analysis method that lets experiments be checked continuously without inflating the false-positive rate from repeated peeking.

Serve points won: The share of points a tennis player wins when serving, the primary measure of how dominant their serve is.

Server-side tagging: Sending tracking events through a first-party server rather than firing vendor tags directly from the browser.

Service level indicator (SLI): The specific, measured metric, such as success rate or latency, used to judge a service's performance against a target.

Service level objective (SLO): An internally set reliability or performance target, measured over a time window, set above any contractual SLA.

Session: A grouped window of a user's activity on a site or app, bounded by an inactivity timeout or similar cutoff rule.

Session length: How long a single continuous visit or usage period lasts, from a user's first action until the session ends.

Session replay: A reconstructed playback of one visitor's clicks, cursor movement and scrolling on a page, captured via browser events.

Sessionization: Grouping a visitor's raw, timestamped events into discrete sessions, usually by closing a session after a period of inactivity.

SHAP values (SHapley Additive exPlanations): A method for explaining an individual prediction by fairly attributing it to the contribution of each input feature.

Share of search: A brand's share of total category search volume, used as a fast-moving proxy for market share and brand awareness.

Share of voice (SOV): A brand's advertising presence, such as spend or impressions, expressed as a percentage of the total for its category.

Share of wallet: The percentage of a customer's total spending in a category that goes to one company rather than its competitors.

Shared socioeconomic pathways (SSPs): A set of standardized scenarios describing possible future global development paths, used to frame climate projections.

Sharp money: Wagers from historically accurate, well-informed bettors, which sportsbooks weight more heavily than casual public bets.

Sharpe ratio: A measure of risk-adjusted return: excess return earned per unit of total volatility taken on.

Shot quality: A general term for models that grade how likely a given scoring attempt is to succeed, based on location, type and pressure.

Showback and chargeback: Two ways of attributing shared cloud costs to teams, one for visibility only, the other billing the cost to their budget.

SIEM (security information and event management): A platform that centralizes and correlates log and event data across an IT environment to detect and investigate security threats.

Signal processing: Techniques for analyzing, filtering, and transforming time-varying signals, such as sensor readings or audio, to extract or clean up information.

Simpson's paradox: A pattern where a trend appears in several groups of data but reverses or disappears when the groups are combined.

Single-cell RNA sequencing (scRNA-seq): Measuring gene expression in individual cells separately, rather than as an average across a tissue sample.

Situational awareness: Real-time understanding of what is happening in an operational environment, built from current data for fast decisions.

Six Sigma: A structured methodology for reducing process defects and variation using statistical measurement and the DMAIC cycle.

SKAdNetwork (SKAN): Apple's privacy-preserving framework for attributing iOS app installs without identifying individual users or devices.

Sleep staging: Classifying sleep into stages such as light, deep and REM sleep based on physiological signals.

Slowly changing dimension: The techniques for recording how a dimension table's attributes change over time without breaking historical reporting.

Small area estimation (SAE): Statistical techniques for producing reliable estimates for areas or subgroups too small to sample directly.

Smart city sensor data: Real-time readings from networked devices, such as traffic counters and air-quality monitors, across urban infrastructure.

Snowflake schema: A dimensional model where dimension tables are normalized into multiple related tables instead of one flat table.

SOAR (security orchestration, automation and response): A platform that automates and orchestrates the response steps a security team takes after an alert is raised.

Social listening: Continuously tracking social platforms and public online conversations for mentions of a brand, topic, or competitor, then analyzing themes.

Sortino ratio: A risk-adjusted return measure like the Sharpe ratio, but penalizing only downside volatility.

SPACE framework: A model for measuring developer productivity across satisfaction, performance, activity, communication, and efficiency.

Span of control: The number of direct reports a manager oversees, used to gauge organizational structure and management capacity.

Spatial analysis: Analytical techniques that use the location, distance or arrangement of data to find patterns tabular data hides.

Speaker diarization: Splitting an audio recording into segments and labeling which speaker was talking in each, without necessarily identifying who they are.

Species distribution model (SDM): A statistical model that predicts where a species can occur by relating known sightings to environmental variables.

Speech analytics: Software that transcribes and analyzes spoken audio, such as call center recordings, to extract sentiment, keywords, and compliance signals.

Spend analysis: Analyzing an organization's purchasing data to understand what is bought, from whom, at what price, and where to save.

Split URL testing: An A/B test that sends visitors to entirely separate page URLs for each variant, instead of changing content dynamically on one shared URL.

Sponsorship valuation: Estimating the fair market worth of a sports sponsorship deal, based on exposure, reach, and comparable agreements.

Spurious correlation: A statistically strong relationship between two variables that has no causal or meaningful connection.

SQL (structured query language): The standard declarative language for defining, querying, and manipulating data in a relational database.

Stablecoin: A cryptocurrency designed to hold a steady value, usually pegged one-to-one to a fiat currency such as the US dollar.

Standard deviation: A measure of how spread out a set of values is around its mean, expressed in the same units as the data.

Standard error: The estimated variability of a sample statistic, such as the mean, if the study were repeated across new samples.

Star schema: A dimensional modeling pattern with a central fact table connected to descriptive dimension tables, shaped like a star.

Stationarity: The property of a time series whose statistical characteristics, such as mean and variance, stay constant over time.

Statistical power: The probability that a hypothesis test correctly detects a real effect, given that one truly exists.

Statistical process control (SPC): Using control charts and statistical limits to tell normal process variation apart from a genuine problem in real time.

Statistical significance: A threshold judgment that an observed effect is unlikely to be due to chance alone, given the data and test used.

Stickiness: The ratio of daily to monthly active users, showing how often the average monthly user returns in a day.

Stream processing: Processing data continuously as individual events arrive, rather than waiting to accumulate a batch.

Strokes gained: A golf metric comparing a player's strokes from a given position to how many strokes a benchmark player needs from there.

Student retention rate: The share of students who continue enrollment from one academic period to the next, most often first year to second year.

Stylometry: Using quantitative features of writing style, such as word frequency and sentence length, to identify or compare authors.

Subscriber churn: The rate at which paying subscribers cancel a service over a given period, the mirror image of subscriber retention.

Success rate (football): The share of American football plays that gain enough yardage, relative to down and distance, to count as a positive outcome.

Supervised learning: A machine learning approach where a model learns from examples that already include the correct answer, to predict that answer on new data.

Supply chain visibility: The ability to track goods, orders, and inventory across every stage of a supply chain, often close to real time.

Surrogate key: An artificial, meaningless identifier generated by the warehouse to uniquely identify a dimension or fact row.

Survey mode effects: Systematic differences in how people answer identical questions depending on the survey's mode of delivery.

Survey weighting: Adjusting survey responses so the sample's demographic makeup matches the target population's known composition.

Survivorship bias: A distortion that occurs when analysis focuses only on entities that "survived" a filter, ignoring those that failed or dropped out.

Sustainable Development Goal (SDG) indicators: The official metrics the United Nations uses to track country-level progress toward the 17 Sustainable Development Goals.

Switchback experiment: An experiment design that alternates a whole system between treatment and control over time, for cases where individual users can't be split.

SWOT analysis (strengths, weaknesses, opportunities, threats): A framework for assessing a company's internal strengths and weaknesses alongside external opportunities and threats.

Synthetic control: A method that builds a weighted composite of untreated units as the counterfactual for a single treated unit, such as one market or region.

T

T-test: A statistical test that compares the means of one or two groups to judge whether an observed difference is likely real.

Tag management: A system for adding and configuring marketing and analytics tracking tags without editing a site's code.

Take-home assignment: A practical exercise given to a job candidate to complete independently, testing real analytical work rather than interview answers.

Tax-loss harvesting: Selling an investment at a loss to offset taxable gains elsewhere, then reinvesting in a similar but not identical asset.

Technical analysis: Forecasting future price movement from historical price and volume patterns rather than company fundamentals.

Tennis Elo rating: An Elo rating system adapted to tennis that estimates win probability between any two players from match results.

Territory planning: Dividing a market into sales territories, by geography, account size or industry, to balance opportunity and workload.

Text analytics: The techniques used to extract structured information and insight from unstructured text, from keyword counts to deep NLP.

Text Encoding Initiative (TEI): An XML standard for encoding literary and historical texts so their structure and scholarly annotations are machine-readable.

Text-to-SQL: Technology that translates a plain-language question directly into a runnable SQL query against a database.

Third-party cookies: Cookies set by a domain other than the one being visited, historically used for tracking visitors across sites.

Threat intelligence: Analyzed information about attacker groups, tools, and techniques used to anticipate and defend against threats, not just react to them.

Throughput: The rate at which a process completes units of work over a period of time, such as orders shipped per hour.

Ticket yield: The average revenue a team or venue earns per ticket sold, after accounting for discounts and dynamic pricing.

Time series forecasting: Predicting future values of a metric using patterns found in its own historical sequence over time.

Time to hire: The number of days between a candidate entering the pipeline and accepting a job offer.

Time to value (TTV): How long it takes a new user, from signup, to first experience a product's core value.

Time tracking: Recording how time is spent across tasks or activities, used to analyze productivity, billing, and where hours go.

Time-series database: A database purpose-built for data points indexed by time, optimized for fast writes and time-range queries.

Time-weighted return (TWR): A return measure that removes the effect of cash flows into or out of a portfolio, isolating the manager's investment performance.

Token: The basic unit of text, roughly a word piece, that a language model reads, processes, and is priced and measured by.

Topic modeling: An unsupervised machine learning technique that discovers the latent themes running through a collection of documents.

Total addressable market (TAM): The total revenue opportunity for a product or service if it captured 100% of its potential market.

Total cost of ownership (TCO): The full cost of using a system over its lifetime, including licensing, infrastructure, implementation, and staff time, not just list price.

Total value locked (TVL): The total value of crypto assets currently deposited in a decentralized finance protocol.

Total value to paid-in (TVPI): A private fund performance ratio, total value returned plus unrealized value, divided by capital investors have actually paid in.

Touchpoint: Any single point of contact between a customer and a brand, on or off the path to purchase, across any channel.

Tracking error: The standard deviation of the difference between a portfolio's returns and its benchmark's returns.

Tracking plan: The document specifying every analytics event a product should fire, its properties, and who owns it.

Traffic flow analysis: Measuring vehicle volume, speed, and density on a road network to identify congestion and evaluate interventions.

Train-test split: Dividing a dataset into separate portions for training a model and for evaluating it, so performance is measured on unseen data.

Training data: The dataset with known outcomes that a machine learning model learns patterns from before being used to predict new cases.

Training load: A quantified measure of how much physical stress a workout or training period placed on the body.

Transaction cost analysis (TCA): Measuring the real cost of executing a trade, including spread, market impact, and timing, versus a benchmark.

Transaction monitoring: Ongoing, largely automated screening of financial transactions to flag activity that may indicate fraud or laundering.

Transition risk: The financial risk to a business from the policy, market and technology shifts of moving to a lower-carbon economy.

True shooting percentage (TS%): A basketball scoring-efficiency rate combining two-point, three-point and free-throw shooting into one number weighted by real point value.

Turnout modeling: Predicting the probability that an individual registered voter will cast a ballot in a given election.

Type I and Type II errors: The two ways a hypothesis test can be wrong, rejecting a true null, or failing to reject a false one.

Tyre degradation: The loss of grip and lap-time performance a motorsport tyre suffers over a stint as it wears and overheats.

U

Underwriting analytics: Data analysis used to decide whether to accept a risk and at what price, in insurance and lending alike.

Unemployment rate: The share of the labor force that is jobless and actively looking for work, excluding people not seeking employment.

Unique visitors: The count of distinct individuals, approximated by a client identifier, who visited a site within a given period.

Unit economics: The direct revenue and costs attributable to a single unit, such as one customer or one transaction.

Unsupervised learning: A machine learning approach that finds structure or patterns in data that has no labeled correct answer to learn from.

Uplift modeling: Predictive modeling that estimates how much a treatment changes an individual's outcome, to target only the people it will actually influence.

Urban heat island (UHI): The tendency of built-up urban areas to be significantly warmer than surrounding rural land, mainly from surfaces and structures.

Usage rate (USG%): The estimated share of a basketball team's possessions a player ends via a shot, turnover or free-throw trip while on the court.

Usage-based insurance (UBI): Insurance priced from how, how much, or how safely a policyholder actually behaves, rather than fixed demographic factors alone.

User and entity behavior analytics (UEBA): Analytics that baseline normal user and device behavior and flag deviations that could signal a compromise, even without a known IOC.

User journey analytics: Analyzing the ordered sequence of touchpoints and actions a user takes across sessions and channels toward a goal.

User properties: Attributes describing a user, such as plan tier or signup date, attached to their profile rather than to a single event.

UTM parameters: Tags added to a URL that record which source, medium, and campaign sent a visitor to a site.

V

Vacancy rate: The share of housing units that are unoccupied and available, a key gauge of how tight a rental market is.

Value at risk (VaR): An estimate of the maximum loss a portfolio is expected to face over a given period at a set confidence level.

Value over replacement player (VORP): How much more value a player provides than a readily available, replacement-level player at the same position.

Value-added model (VAM): A statistical method estimating a teacher or school's contribution to student test-score growth, adjusted for prior scores.

Variable-rate application (VRA): Applying seed, fertilizer or crop inputs at a rate that changes across a field instead of one uniform rate.

Variance: The average of the squared differences between each value and the mean, quantifying spread before returning to original units.

Variance analysis: Comparing actual financial results to budget or forecast and breaking the gap into the specific causes behind it.

Variant calling: Identifying positions where a sequenced sample's DNA differs from a reference genome.

Vector database: A database optimized to store numeric embeddings and quickly find the ones most similar to a given query vector.

Vendor lock-in: The difficulty and cost of switching away from a vendor once an organization is deeply dependent on that vendor's proprietary systems.

Video analytics: Extracting quantitative insight from video, whether audience behavior around published video or automated analysis of video frames.

View-through conversion (VTC): A conversion credited to an ad a user saw but did not click, within a defined attribution window.

Viral coefficient: The average number of new users each existing user brings in, used to measure whether growth is self-sustaining.

VO2 max (maximal oxygen uptake): The maximum rate at which the body can use oxygen during intense exercise, a standard measure of aerobic fitness.

Voice of the customer (VoC): The structured practice of capturing and analyzing customer feedback across surveys, support, reviews and social channels.

Volatility: The degree to which an asset's returns fluctuate over time, the standard proxy for investment risk.

Voter file: A state-maintained list of registered voters, appended with modeled data, used as the base file for campaign analytics.

W

Wallet clustering: Grouping blockchain addresses believed to be controlled by the same person or entity, using transaction pattern heuristics.

Watch time: The total minutes viewers spend watching a piece of video content, a key ranking and monetization signal on video platforms.

Water quality index (WQI): A single composite score summarizing several water-quality measurements into one number for easy comparison over time or place.

Watermarking: A stream processing technique estimating how far event time has progressed, to decide when a window can be finalized.

Wealth screening: Using public and third-party data to estimate a donor's capacity and likelihood to make a major gift.

Weather normalization: Adjusting energy usage or other weather-sensitive data to remove the effect of unusually hot, cold or wet conditions.

Webhook: An automated HTTP callback a system sends to a URL the moment a specific event happens, instead of waiting to be asked.

Weighted average: An average in which some values contribute more to the result than others, based on an assigned weight.

What-if analysis: Testing how a model's output changes when one or more input assumptions are changed, to explore specific alternative outcomes.

Wildfire risk modeling: Estimating the likelihood and potential severity of wildfire at a given location by combining fuel, weather and ignition data.

Win probability model: A statistical model estimating a team's real-time chance of winning a game given the current score, time remaining and situation.

Win rate: The share of sales opportunities that close as won, out of all opportunities that reached a final decision.

Window function: A SQL calculation performed across a set of related rows without collapsing them into a single output row.

Wins above replacement (WAR): A single number estimating how many more wins a player adds over a season than a readily available replacement-level substitute.

Word error rate (WER): The standard accuracy metric for speech recognition, measuring the edit distance between a transcript and the correct reference text.

Workflow orchestration: Scheduling, sequencing and monitoring interdependent data tasks so each step runs only after the ones it depends on succeed.

Workforce planning: Forecasting an organization's future staffing needs and aligning hiring, skills and budget to meet them.

Working capital: The difference between a company's current assets and current liabilities, a measure of short-term operating liquidity.

Y

Yield curve: A plot of bond yields across different maturities, most often for government debt of a single issuer.

Yield mapping: Recording crop yield by GPS location during harvest to build a spatial map of how output varies across a field.

Z

Z-score: The number of standard deviations a value lies above or below the mean of its distribution.

Zero-based budgeting (ZBB): Building each budget cycle from a zero base, requiring every expense to be justified again rather than carried forward.

Zero-copy data sharing: Granting another party live access to a dataset in place, without exporting, copying, or moving the files.

Zero-day vulnerability: A software flaw unknown to the vendor, or unpatched, at the time it is discovered or actively exploited.