A–ZGlossary

A controlled experiment where two versions of something (a webpage, feature, message, or product) are shown to different user segments simultaneously. The version that performs better on a defined metric wins.

A communication technique that involves fully concentrating on, understanding, and responding to a speaker, rather than passively hearing words. It includes verbal and non-verbal cues, paraphrasing, and asking clarifying questions to confirm understanding.

Average Daily Rate: the average revenue earned per occupied room per day. Calculated as total room revenue divided by the number of rooms sold. One of the three pillars (alongside Occupancy Rate and RevPAR) for understanding revenue performance.

Answer Engine Optimization: the emerging practice of optimizing content to appear directly in AI-powered answer engines (like ChatGPT, Perplexity, or Google's AI Overviews) rather than just traditional search result pages. As users increasingly get answers without clicking links, AEO is becoming as important as SEO.

A collaborative technique for organizing large volumes of unstructured data (insights, ideas, quotes, observations) into meaningful clusters by natural grouping. Originally developed by Japanese anthropologist Jiro Kawakita; also called the KJ Method.

The design and coordination of multiple AI agents working together on complex tasks, where each agent has a specialized role, defined tools, and clear boundaries. Orchestration determines how agents communicate, share context, handle failures, and escalate decisions to humans.

AI systems that can autonomously plan, execute multi-step tasks, use tools, and adapt their approach based on intermediate results, moving beyond single-prompt chat interactions into genuine workflow automation. Agentic AI can browse the web, write and run code, manage files, and orchestrate other AI models to accomplish complex goals.

An iterative, people-centered approach to delivering value. Rather than following a rigid plan, Agile teams work in short cycles, gather feedback, and adapt continuously, embracing change as a competitive advantage.

Safety boundaries, content filters, behavioral constraints, and validation checks applied to AI systems to prevent harmful, off-topic, unreliable, or policy-violating outputs. Guardrails operate at multiple layers: model training (RLHF), system prompts, output filtering, and application-level validation.

The strategic adoption of artificial intelligence into existing workflows and processes. Not about replacing what works, but about finding the specific points where AI creates the most value.

The minimum understanding of AI capabilities, limitations, and effective use that every knowledge worker needs to collaborate with AI systems productively. Includes: knowing what AI can and cannot do, recognizing hallucinations, understanding when to trust and when to verify, and developing the judgment to use AI as a thinking partner rather than an oracle.

A deliberate, organization-wide plan for how artificial intelligence will be adopted, governed, and leveraged to create competitive advantage. Covers use-case prioritization, data readiness, build-vs-buy decisions, capability building, and risk management, not merely a list of tools to deploy.

The tendency of large language models to tell users what they want to hear rather than what is accurate: excessive agreeableness, unwarranted flattery, and conflict avoidance at the expense of truthfulness. Sycophantic AI validates weak ideas, avoids constructive pushback, and reinforces existing biases instead of challenging them.

A communication philosophy where the default mode is asynchronous (written messages, documents, recorded videos) rather than synchronous (meetings, calls, real-time chat). Meetings become the exception, reserved for decisions, alignment, and human connection, not the default for sharing information.

Also known as backlog refinement: the ongoing process of reviewing, clarifying, estimating, and prioritizing product backlog items. Ensures the team always has a ready, well-understood queue of upcoming work.

A structured retrospective conducted after an incident, failure, or significant miss, with the explicit commitment that no individual will be punished or blamed. The goal is systemic learning: what conditions made this failure possible, and what changes will prevent recurrence?

The percentage of visitors who land on a page and leave without any further interaction: no clicks, no scrolls past a threshold, no second page view. In Google Analytics 4, the inverse metric "engagement rate" is preferred: a session counts as engaged if it lasts over 10 seconds, triggers a conversion, or includes 2+ page views.

A facilitation technique where a large group is divided into smaller subgroups to work on specific tasks, questions, or discussions simultaneously, before reconvening to share insights with the full group.

A prompting technique where the model is asked to reason through a problem step by step before producing a final answer. By making intermediate reasoning explicit, chain-of-thought significantly improves accuracy on tasks requiring logic, math, multi-step analysis, or nuanced judgment. First demonstrated by Wei et al. (2022) at Google.

Continuous Integration and Continuous Delivery: the practice of frequently merging code changes into a shared repository (CI) and automatically preparing them for release (CD). The goal is to make deployment boring, routine, and low-risk.

Short, focused coding exercises designed for deliberate practice. Like a martial arts kata, the point is not the solution itself but the repeated practice of problem-solving patterns and techniques.

A day-long intensive practice event focused on the fundamentals of software development, not on shipping features. Participants practice the same problem (Conway's Game of Life) multiple times with different constraints and pair partners, deleting all code at the end of each session.

The presence of different thinking styles, problem-solving approaches, and cognitive frameworks within a team, distinct from demographic diversity, though often correlated with it. Cognitive diversity is what makes a team collectively smarter than its smartest individual member.

The total amount of mental effort being used in working memory at any given moment. Originally developed by John Sweller for educational design, the concept applies broadly: every interface, process, or decision adds load. When load exceeds capacity, performance collapses: people make errors, give up, or default to heuristics.

An educational approach where individuals work together to solve problems, complete tasks, or create knowledge, moving beyond passive information transfer to active, social co-construction of understanding.

The strategic practice of transforming a single piece of content into multiple formats and channels to maximize reach and return on creative investment. A conference talk becomes a blog post, a carousel, a newsletter issue, and a video clip, each adapted to the native format and audience of its platform.

The practice of deliberately designing what context an AI model receives: system prompts, retrieved documents, conversation history, tool outputs, and structured metadata. Context engineering treats the model's input as a first-class engineering artifact, not an afterthought. The quality of AI output is bounded by the quality of the context you provide.

The gradual degradation of an AI system's effectiveness as the context it relies on becomes outdated, irrelevant, or contradictory. In long-running agent sessions, accumulated context can drift from the original intent, leading to increasingly poor decisions. In RAG systems, stale or poorly maintained knowledge bases produce confident but outdated answers.

The maximum amount of text (measured in tokens) that a large language model can process in a single interaction. Everything the model "sees" during a conversation, including the system prompt, conversation history, retrieved documents, and its own responses, must fit within this window. Once exceeded, earlier content is lost.

An ongoing effort to improve products, services, or processes over time, through incremental improvements or breakthrough changes. Rooted in the Japanese concept of Kaizen: "change for the better."

A culture and mindset of ongoing growth through experience, feedback, and structured practice. Not just attending courses, but building learning into daily work, making it a habit rather than an event.

A set of three user-centric performance metrics defined by Google that measure real-world page experience: Largest Contentful Paint (loading), Interaction to Next Paint (responsiveness), and Cumulative Layout Shift (visual stability). Since 2021, Core Web Vitals are a confirmed Google ranking signal.

The deliberate use of limitations (time, resources, format, rules) as a catalyst for creative thinking. Counterintuitively, constraints often produce more original and useful ideas than unlimited freedom, by forcing the mind to explore less obvious territory.

Customer Relationship Management: a system and strategy for managing all interactions with current and potential customers. Tracks communication history, preferences, and touchpoints to build stronger, more personalized relationships.

CLS: the Core Web Vital that measures visual stability. It quantifies how much visible content unexpectedly shifts during the page's lifetime. A good CLS score is 0.1 or less, meaning elements stay where you expect them as the page loads and renders.

The practice of designing and building systems that collect, store, process, and transform raw data into a form that can be reliably used for analysis and decision-making. It's the infrastructure behind every data-informed product, even if nobody sees it.

Using data as a key input to decisions, alongside experience, judgment, and context, rather than relying solely on intuition or authority. Note: "data-informed" (not "data-driven") preserves the human role in interpreting and acting on evidence.

The deteriorating quality of decisions made after a long session of decision-making. As mental energy depletes, people default to the path of least resistance: choosing the status quo, deferring decisions, or making impulsive choices. First studied by social psychologist Roy Baumeister, the concept explains why willpower and judgment are finite daily resources.

K. Anders Ericsson's framework for purposeful, structured practice. It requires four elements: a clearly defined goal, full concentration, immediate feedback, and repetition with reflection. It is not simply doing more of the same work.

A human-centered innovation framework that emphasizes empathy, ideation, and rapid prototyping. It starts with deeply understanding people and their needs before jumping to solutions.

The coordinated use of online channels (SEO, social media, email marketing, content creation, and paid advertising) to attract, convert, and retain guests. In hospitality, it's the bridge between an exceptional property and the guests who would love it.

The percentage of total bookings that come through the property's own website or direct channels rather than through OTAs (Online Travel Agencies) like Booking.com or Airbnb. Higher direct bookings mean higher margins and a direct relationship with the guest.

The structured upfront period of a consulting engagement dedicated to understanding the client's context deeply before proposing or building anything. Includes stakeholder interviews, process observation, data review, and assumption mapping. The goal: define the right problem before solving it.

Google's quality evaluator framework: Experience, Expertise, Authoritativeness, and Trustworthiness. Added 'Experience' in 2022 to reward first-hand, lived expertise over aggregated or AI-generated content. High E-E-A-T signals strongly influence both search rankings and AI citation likelihood.

A structured process to welcome new team members and integrate them into the organization. Goes beyond paperwork and tool setup to include cultural integration, clear learning paths, and early relationship building.

A numerical representation of text (or images, audio, code) as a dense vector in high-dimensional space, where semantically similar content is positioned closer together. Embeddings capture meaning, not just keywords: "happy" and "joyful" land near each other, while "bank" (financial) and "bank" (river) land far apart based on context.

A framework for continuously monitoring and classifying technologies by their maturity and relevance to your organization. Popularized by ThoughtWorks, it plots technologies across four rings (Adopt, Trial, Assess, Hold), giving teams a shared vocabulary for technology decisions.

The structure that defines how an independent consultant or advisory firm works with a client: including scope, duration, deliverables, decision rights, communication cadence, and commercial terms. Common models include project-based, retainer, embedded, and fractional arrangements.

Greg McKeown's discipline of pursuing less but better. Not about doing more things efficiently, but about doing only the right things: the disciplined pursuit of less that creates the highest possible contribution.

A process where the outputs of a system are used as inputs to that same system, creating a self-correcting or self-reinforcing cycle. In team contexts, feedback loops are the mechanisms by which teams learn from their actions and adjust course.

The process of further training a pre-existing foundation model on domain-specific data to specialize its behavior, tone, or knowledge. Unlike prompting (which guides behavior at inference time) or RAG (which retrieves external knowledge), fine-tuning permanently alters the model's weights, baking new patterns into the model itself.

A disciplined approach to exploring multiple plausible futures rather than predicting a single one. Uses tools like scenario planning, trend analysis, and weak-signal detection to help organizations make more resilient strategic decisions under uncertainty.

A class of artificial intelligence systems capable of producing original content (text, images, code, audio, and more) by learning patterns from vast training datasets. Unlike traditional AI that classifies or predicts, generative AI creates.

Gross Operating Profit Per Available Room: the profit generated per available room after deducting operating expenses. Unlike RevPAR, which only measures revenue, GOPPAR reveals actual profitability by accounting for costs.

A data-driven, experimentation-heavy approach to finding scalable strategies for rapid user acquisition, activation, and retention. Originated in startups as a way to grow fast with limited resources by testing creative, often unconventional, tactics.

Carol Dweck's concept that abilities and intelligence can be developed through dedication, hard work, and learning from failure. The opposite of a fixed mindset, which sees talent as innate and unchangeable.

A holistic approach to mapping and optimizing every touchpoint of the guest journey, from first impression to post-stay follow-up, using a blend of data-driven insights and human-centered design.

When an AI model generates information that is fluent and confident-sounding but factually incorrect, fabricated, or unsupported by its training data. Hallucinations can range from subtle (wrong dates, invented citations) to egregious (entirely fictional events presented as fact). The term is debated (some prefer "confabulation") but it is the industry standard.

A reframing technique that converts problem statements into open-ended innovation questions. "We struggle with user onboarding" becomes "How might we make onboarding feel like a discovery?" That small shift from problem to possibility changes the energy in the room: it implies the problem is solvable and invites creative exploration.

A partnership approach where AI amplifies human creativity, judgment, and decision-making rather than replacing them. The human brings context, taste, and strategic intent. The AI brings speed, breadth, and tireless iteration.

A system design pattern where humans review, approve, correct, or override AI outputs at critical decision points. Rather than full automation or full manual control, HITL finds the productive middle ground: AI handles the heavy lifting, humans handle the judgment calls.

The deliberate process of generating, evaluating, and implementing new ideas within an organization. It bridges creativity and execution, ensuring promising ideas don't die in brainstorming sessions but become real, tested products and processes.

INP: the Core Web Vital that measures responsiveness. It captures the latency of every click, tap, and keyboard interaction throughout a page's lifecycle, then reports a value near the worst-case. A good INP is 200 milliseconds or less, meaning the page visually responds to user input almost instantly.

A learning technique where someone observes an experienced colleague performing their work to understand workflows, decision-making processes, and unspoken organizational practices.

A visual representation of the steps a person takes to accomplish a goal, whether a customer buying a product, a new employee onboarding, or a patient navigating healthcare. Maps typically capture actions, thoughts, emotions, and pain points across each stage.

The systematic approach to capturing, organizing, sharing, and leveraging organizational knowledge. A well-designed system acts as a "second brain" for the organization, making critical information accessible to everyone who needs it.

LCP: the Core Web Vital that measures loading performance. It marks the time from when a page starts loading to when the largest visible content element (typically a hero image, heading block, or video poster) is fully rendered in the viewport. A good LCP is 2.5 seconds or less.

A philosophy originating from manufacturing (Toyota) focused on maximizing value while minimizing waste. Every activity is evaluated through one question: does this create value for the customer?

A methodology for building businesses and products that advocates for shorter development cycles, rapid experimentation, and learning what customers actually want. Developed by Eric Ries, it popularized concepts like the Build-Measure-Learn loop and the Pivot.

A set of 33 simple facilitation methods that replace conventional meeting formats. They distribute participation more equally, ensuring everyone's voice is heard, not just the loudest or most senior.

A fast, structured workshop format (typically 60–90 minutes) designed to generate solutions, make decisions, and produce an actionable task list, without the dysfunction of unstructured meetings. Developed by AJ&Smart as an accessible alternative to the full Design Sprint.

A branch of artificial intelligence where systems learn to make predictions or decisions by identifying patterns in data, without being explicitly programmed for each task. It powers most modern AI applications.

Model Context Protocol: an open standard developed by Anthropic that provides a universal interface for connecting AI models to external data sources, tools, and services. MCP standardizes how AI applications discover and interact with capabilities (file systems, databases, APIs, code repositories), replacing fragile, custom integrations with a consistent protocol.

The practice of present-moment awareness and intentional focus. Not about emptying the mind, but about training attention: noticing where it goes and gently redirecting it to what matters.

The smallest, fastest test that can generate reliable evidence for or against a specific assumption. Where an MVP validates a product concept with real users, an MVE validates a single hypothesis, often before any product exists. It is the unit of learning, not the unit of delivery.

Minimum Viable Product: the simplest version of a product that can validate core assumptions with real users. Not a half-baked product, but a focused experiment designed to learn the most with the least effort.

Marshall Rosenberg's communication framework based on empathic listening and honest expression. It separates observation from evaluation, identifies feelings and needs, and makes clear, actionable requests.

The percentage of available rooms that are occupied over a given period. Calculated as rooms sold divided by rooms available. The most intuitive hospitality metric, but dangerous when optimized in isolation from ADR and RevPAR.

Objectives and Key Results: a goal-setting framework that aligns teams around ambitious objectives (qualitative) measured by specific key results (quantitative). Popularized by Andy Grove at Intel and later by Google.

The totality of a property's digital footprint: its website, OTA listings, social media profiles, review platforms, and search engine visibility. For unique hospitality properties, the online presence must convey the experience, not just the amenities.

A model coined by Henry Chesbrough arguing that organizations should use both internal and external ideas, and internal and external paths to market, to advance their products and technology. The boundary between a firm and its environment is permeable rather than protective.

A self-organizing meeting format where participants create and manage their own agenda around a central theme. Based on the principle that the people who show up are the right people, and whatever happens is the only thing that could have.

Online Travel Agency: a platform (like Booking.com, Airbnb, or Expedia) that acts as an intermediary between travelers and accommodation providers, taking a commission on each booking in exchange for visibility and distribution.

Two developers working together at one workstation: one writes code (the "driver") while the other reviews each line in real time (the "navigator"). Roles are swapped regularly. It's a conversation about code, not just writing it.

A visual display of the most important metrics for tracking progress toward goals. A good dashboard makes the current state of a system immediately understandable, helping teams focus on what matters and act quickly on signals.

Property Management System: the central software that manages a hospitality property's operations: reservations, check-in/check-out, room assignments, billing, guest profiles, and housekeeping. It runs the day-to-day operations of any property.

A planning technique where a team imagines that a project has already failed, then works backward to identify what likely caused it. Developed by psychologist Gary Klein, it surfaces hidden risks and blind spots before they become real problems.

A design pattern that sequences information and options across time, showing only what is relevant at each stage and revealing complexity as the user needs it. Rather than overwhelming with everything upfront, progressive disclosure respects cognitive limits and guides attention: simple first, detailed on demand.

The practice of crafting effective instructions for large language models to produce useful, accurate, and well-structured outputs. Goes far beyond writing clever sentences. It encompasses system prompt design, few-shot examples, chain-of-thought reasoning, output formatting, and understanding how context windows and token limits shape model behavior.

Amy Edmondson's concept describing an environment where team members feel safe to take interpersonal risks (speaking up, asking questions, admitting mistakes, and challenging ideas) without fear of punishment or humiliation.

Retrieval-Augmented Generation: an architecture pattern that grounds LLM responses in real, retrieved documents rather than relying solely on the model's training data. A RAG system first searches a knowledge base (documents, databases, APIs) for relevant context, then feeds that context to the model alongside the user's question, which sharply reduces hallucinations and enables domain-specific accuracy.

The practice of quickly building a simplified, testable version of a product, feature, or idea to validate assumptions and gather feedback before committing to full development.

Improving the internal structure of existing code without changing its external behavior. It's not rewriting. It's systematically making code cleaner, simpler, and easier to understand, one small transformation at a time.

The practice of effective teamwork across distances and time zones. It goes beyond video calls, requiring intentional communication design, async-first workflows, and tools that make distributed work seamless.

Revenue Per Available Room: the headline metric in hospitality. Calculated as ADR multiplied by Occupancy Rate (or total room revenue divided by total available rooms). It captures both pricing power and demand in one number.

The practice of thinking beyond the immediate, obvious consequence of a decision to ask: "And then what?" Second-order effects are often where the real risks and opportunities , and where shallow analysis consistently fails.

Search Engine Optimization: the practice of improving a website's visibility in organic (non-paid) search engine results. Encompasses technical health, content quality, keyword relevance, page experience, and link authority.

Search Engine Results Page: the page displayed by a search engine in response to a query. A SERP typically contains organic results (ranked by relevance), paid ads, and rich result features like featured snippets, knowledge panels, and "People also ask" boxes.

Kent Beck's four rules: the code (1) passes all tests, (2) reveals intention, (3) contains no duplication, and (4) has the fewest elements possible. Simplicity is not about doing less. It's about achieving clarity and removing everything that doesn't serve it.

A goal-setting framework where objectives are Specific, Measurable, Achievable, Relevant, and Time-bound. Provides a clear structure for defining goals that can be tracked and evaluated objectively.

A Scrum ceremony at the start of each sprint where the team selects items from the product backlog and commits to delivering them within the sprint. It aligns the team on goals, effort, and shared understanding before work begins.

Site Reliability Engineering: a discipline pioneered by Google that applies software engineering principles to infrastructure and operations problems. SRE teams treat reliability as a software problem, codify operational toil into automation, define Service Level Objectives (SLOs), and use error budgets to balance innovation speed with stability.

The practice of systematically identifying all individuals and groups who have an interest in or influence over a project or change initiative. Maps typically plot stakeholders on axes of interest and influence to prioritize engagement strategies.

Machine-readable metadata embedded in a webpage using formats like JSON-LD (preferred), Microdata, or RDFa. Structured data speaks Schema.org, a shared vocabulary that tells search engines and AI models exactly what your content is: an Article, a Person, a FAQ, an Event, a Product, and so on.

Understanding organizations as complex, interconnected systems rather than collections of isolated parts. It focuses on relationships, feedback loops, and unintended consequences. Seeing both the forest and the trees.

A profile describing someone with deep expertise in one discipline (the vertical bar of the T) combined with broad knowledge across adjacent fields (the horizontal bar). The T-shape enables both specialist credibility and cross-functional collaboration, a combination that is rarer than most people think.

Test-Driven Development: a software development practice where you write a failing test before writing the code to make it pass. The cycle is: Red (write failing test), Green (make it pass), Refactor (clean up).

A structured reflection meeting held at the end of each iteration where the team inspects how they worked together and identifies improvements. It answers: what went well, what didn't, and what to change.

The accumulated cost of shortcuts, quick fixes, and deferred quality work in a codebase or system. Like financial debt, it compounds: a shortcut taken today becomes a constraint that slows down every future change. Coined by Ward Cunningham, the metaphor deliberately frames engineering quality as an economic decision, not a moral one.

A decision filter from Essentialism: when evaluating an opportunity, if it doesn't score at least 90% on your most important criterion, treat it as a 0, a clear no. Eliminates the agonizing middle ground between "kind of good" and "great."

The fundamental unit of text that large language models process. A token is roughly 3/4 of a word in English: "hamburger" becomes ["ham", "burger"], while short words like "the" are single tokens. Tokens determine context window limits, API costs, and processing speed. Understanding tokenization helps explain why models sometimes struggle with spelling, counting, and non-English languages.

A sequenced, milestone-driven plan for moving an organization from its current state to a defined future state. Unlike a project plan, a transformation roadmap accounts for the human, cultural, and strategic dimensions of change, not just tasks and timelines.

Total Revenue Per Available Room: an evolution of RevPAR that includes all revenue streams, not just room revenue. Captures income from dining, spa, activities, merchandise, and ancillary services per available room.

A source control branching model where all developers work on a single branch (the "trunk" or "main" branch), integrating their changes frequently, at least daily. Long-lived feature branches are avoided.

The deepest kind of client-consultant relationship: one where you are sought out for honest perspective and sound judgment, not just technical expertise. A trusted advisor has earned the right to challenge the client's thinking, say difficult things, and care about their success beyond any single transaction.

User Experience Design: the practice of creating products and services that are useful, usable, and delightful from the user's perspective. It encompasses research, information architecture, interaction design, and usability testing.

A development approach popularized in the AI era where programmers describe what they want in natural language and let AI tools generate the implementation, focusing on intent and outcome rather than line-by-line code authorship.

The idea, coined by Alison Coward, that great collaboration is a practice, not a one-off event. Workshop Culture is about embedding the mindset and methods of effective facilitation into how a team works every day, not just reserving them for special occasions.