The growth marketing glossary.

A/B testing is a controlled experiment that splits traffic between two versions of a page, ad, email, or flow to measure which drives more of a target action. The result is only trustworthy once it reaches statistical significance — typically a 95% confidence level on a pre-calculated sample size — so the lift is attributable to the change rather than to chance. Calling a winner early is the most common mistake; by one analysis, most winning tests would not have held up if run to their proper sample size. It is the core mechanic of disciplined conversion work, not a one-off.

ABM (Account-Based Marketing) is a B2B strategy that treats individual high-value accounts as markets of one, coordinating marketing and sales around a defined list of target companies. Instead of casting a wide net for leads, ABM concentrates personalized campaigns, content, and outreach on the accounts most likely to become large customers. It depends on tight sales-and-marketing alignment and shared account-level data to work. Done well, it raises deal size and win rates by focusing effort where revenue actually lives.

AEO (Answer Engine Optimization) is the practice of structuring content so it gets selected and surfaced as the answer inside AI-powered search features like Google AI Overviews, Google AI Mode, and voice assistants. It optimizes for the answer-retrieval layer — being the source an engine picks when a user asks a specific question — rather than for a ranked list of blue links. Strong AEO depends on clean, answer-first content, structured data, and clear topical authority. As AI search intercepts more queries before a click ever happens, AEO is how brands stay visible at the moment of the answer.

AI Overviews are Google's AI-generated answer summaries that appear at the top of the search results page, synthesizing information from multiple web pages and linking to them as sources. Launched in the US in May 2024 and now powered by Gemini, they appear on a growing share of queries and typically reduce click-through to traditional results. As of 2026, Google places inline source links next to the specific text they support and lets users nominate Preferred Sources to influence which publishers are surfaced.

AI search refers to search experiences that return a synthesized, often cited answer instead of a list of links — including Google AI Overviews and AI Mode, ChatGPT, Perplexity, and Copilot. These answer engines retrieve passages from many sources, rank the evidence, and generate a single response, frequently using a query fan-out technique that runs several related searches at once. The practical shift for marketers is the zero-click answer: users get what they need without visiting a site, so being the cited source matters more than ranking #1. Winning here is the combined job of AEO and GEO.

Ad fatigue is the decline in an ad's performance — falling click-through and rising cost — that sets in when an audience sees the same creative too many times. On Meta, fatigue commonly appears as average frequency climbs past roughly 3 for cold audiences and 5 to 7 for warm retargeting, with click-through often dropping 20 to 40% from peak within the first week of heavy delivery. The fix is a steady creative refresh cadence, usually every 7 to 21 days depending on audience size and spend, rather than simply raising budget. It is a creative problem before it is a media problem.

Annual Recurring Revenue (ARR) is the normalized, predictable subscription revenue a SaaS business expects over a 12-month period from active contracts, excluding one-time fees, services, and usage overages. It is typically calculated as MRR multiplied by 12, and serves as the headline growth metric for SaaS companies with annual contracts. ARR is a snapshot of contracted run-rate revenue at a point in time, not a measure of revenue already booked under GAAP.

AOV (Average Order Value) is the average amount a customer spends per order, calculated as total revenue divided by the number of orders over a period. It is a direct lever on unit economics: raising AOV through bundling, upsells, or thresholds lifts revenue without paying again to acquire the customer. Because it sits alongside conversion rate and purchase frequency in the revenue equation, a higher AOV widens the gap between LTV and CAC. It is one of the cheapest growth levers a brand can pull.

A backlink is a hyperlink from one website to another, used by search engines as a signal of credibility and authority. Backlinks remain among Google's most important ranking signals in 2026, but quality, relevance, and the authority of the linking domain matter far more than volume, and manipulative link schemes carry real penalty risk. Authoritative links also reinforce the trust and authoritativeness assessed under E-E-A-T.

CAC (Customer Acquisition Cost) is the total cost of winning a new customer, calculated by dividing all sales and marketing spend over a period by the number of customers acquired in that period. It is the benchmark for whether growth is sustainable: healthy businesses keep CAC well below the lifetime value of the customers they acquire. The standard guardrail is the LTV:CAC ratio, where roughly 3:1 or higher signals efficient growth. Rising CAC is often the first sign a channel is saturating.

CPA (cost per acquisition, also called cost per action) is the average amount you pay for one completed conversion — a purchase, lead form, signup, call, or install — calculated as total cost divided by conversions. In Google Ads it underpins Target CPA bidding, where the system optimizes toward a CPA you set. CPA measures business outcomes rather than clicks, which makes it more decision-useful than CPC, though it stops short of profit; pair it with POAS or LTV for a full read.

CPC (cost per click) is the amount an advertiser pays each time someone clicks an ad; impressions cost nothing. In CPC bidding you set a maximum CPC — the most you're willing to pay for a click — and the actual cost is often lower. It is the foundational cost metric of search and most auction-based paid media. Note that some Google Ads formats are shifting billing models: Demand Gen campaigns optimizing for view-through conversions on Discover move from CPC to CPM billing effective July 15, 2026.

CPL (Cost Per Lead) is the average cost to acquire one lead, calculated as total campaign spend divided by the number of leads generated. It is an early-funnel efficiency metric that tells you how affordably you are filling the top of the pipeline. CPL only tells half the story on its own — a cheap lead that never converts is expensive in disguise — so it is most useful alongside lead quality, conversion rate, and CAC.

CPM (cost per mille, or cost per thousand impressions) is the amount an advertiser pays for every 1,000 times an ad is shown, whether or not anyone clicks. It is the default pricing model for awareness, reach, and most programmatic display and video buying, where the goal is exposure rather than immediate clicks. CPM is increasingly relevant on Google as more formats adopt impression-based billing — for example, Demand Gen view-through optimization on Discover moves to CPM as of July 15, 2026.

CRO (Conversion Rate Optimization) is the systematic practice of increasing the share of visitors who take a desired action — buying, signing up, booking — through testing and refining the experience. It works through hypotheses, experiments like A/B tests, and analysis of how users actually behave, not guesswork. Because CRO lifts results from traffic you already have, it compounds the return on every other channel. A small conversion gain often beats a large spend increase on economics alone.

CTR (clickthrough rate) is the number of clicks an ad receives divided by the number of times it is shown (impressions): clicks / impressions. Five clicks on 100 impressions is a 5% CTR. A strong CTR signals that creative, targeting, and messaging are relevant to the audience, and in Google Ads it feeds expected CTR, a component of Ad Rank that influences both ad position and cost.

Churn rate is the percentage of customers — or recurring revenue — lost over a given period, calculated as customers lost divided by customers at the start of the period. It is the inverse of retention and the single biggest drag on LTV: high churn forces acquisition to keep refilling a leaking bucket. In subscription and SaaS models, the distinction between customer churn and revenue churn matters, since losing a few large accounts can hurt more than losing many small ones. Lowering churn is almost always cheaper than raising acquisition to offset it.

Connected TV (CTV) is television content delivered over the internet to a smart TV, streaming device, or game console, allowing advertisers to buy addressable, measurable video placements that combine TV-scale reach with digital targeting and attribution. It is the fastest-growing major video channel: the IAB projects U.S. CTV ad spend to grow 13.8% in 2026, and eMarketer expects 2026 to be the first year U.S. CTV upfront spend (about $17.73B) exceeds primetime linear TV upfront spend (about $16.98B). For full-funnel programs, CTV bridges upper-funnel awareness and lower-funnel performance when paired with retargeting and incrementality measurement.

Consent Mode v2 is Google's framework for passing a user's consent choices to Google tags (Google Ads, GA4, Floodlight) so measurement and ad personalization adjust to whether the user granted permission. Introduced in late 2023, it added two required signals, ad_user_data and ad_personalization, on top of the original ad_storage and analytics_storage. Advertisers serving users in the EEA must transmit these signals to keep using Google's audience and measurement features for that traffic; when consent is denied, Google can model conversions from aggregated, non-identifying data.

A conversion funnel is the staged path a prospect follows toward a desired action — typically awareness, consideration, conversion, and retention — modeled so each step's drop-off can be measured and improved. Mapping it shows exactly where prospects leak out, turning vague we need more sales into a specific fix at a specific stage. The funnel is a model, not a literal journey; real buyers loop, pause, and re-enter, especially as AI answer engines compress research into a single step. Used well, it focuses optimization effort where the largest gains actually sit.

Conversion rate (CVR) is the percentage of ad interactions that result in a conversion — calculated as conversions divided by total clicks (or interactions), then multiplied by 100. For example, 20 conversions from 600 clicks is a 3.3% CVR. It is the core efficiency signal for whether traffic and landing experiences are turning attention into action, and the lever conversion-rate optimization works on directly.

The Conversions API (CAPI) is Meta's server-side interface for sending web, app, offline, and messaging events directly from a business's server to Meta, rather than relying solely on the browser-based Meta Pixel. Because it is server-to-server, it is resilient to browser tracking restrictions, ad blockers, and signal loss, improving event match quality, attribution, and ad delivery optimization. Meta recommends pairing CAPI with the Pixel and deduplicating shared events; setup requires a dataset/Pixel ID, a Business Manager, and an access token.

Core Web Vitals are Google's set of three field metrics for measuring real-world user experience: Largest Contentful Paint (LCP) for loading, Interaction to Next Paint (INP) for responsiveness, and Cumulative Layout Shift (CLS) for visual stability. The good thresholds, measured at the 75th percentile of page loads, are LCP at or under 2.5 seconds, INP at or under 200 milliseconds, and CLS at or under 0.1. INP replaced First Input Delay as a Core Web Vital on March 12, 2024.

A customer data platform (CDP) is packaged software that builds and maintains a persistent, unified customer record by ingesting data from all of a brand's sources, resolving identities into single profiles, and making those profiles accessible to other systems for activation. The term was coined by David Raab, whose CDP Institute defines the category; a CDP differs from a data warehouse or CRM because it owns identity resolution and exposes governed, ready-to-use audiences to downstream channels. With third-party signals less reliable, CDPs anchor first-party and zero-party data strategies.

A data clean room is a secure environment where two or more parties combine their data for mutually agreed uses without either party accessing the other's raw, user-level records. The IAB Tech Lab defines it as a secure collaboration environment that enforces strict data-access limits; results are returned only in aggregate, with minimum-match thresholds that prevent individuals from being re-identified. Walled gardens and retail-media networks (e.g., Google Ads Data Hub, Amazon Marketing Cloud) use clean rooms for privacy-safe measurement, audience overlap, and attribution.

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trust, the framework Google's human Search Quality Raters use to assess content quality. The first E, Experience, was added in December 2022, asking whether content was produced with first-hand experience; of the four, Trust is the most important. E-E-A-T is not a direct ranking factor but describes the qualities Google's systems aim to reward, and it matters most for Your Money or Your Life (YMYL) topics like health and finance.

A featured snippet is a short answer Google extracts from a ranking page and displays in a highlighted box at the top of the results, often called position zero. It feeds voice results and AI features, and it remains a high-value target in 2026 even as AI Overviews expand. When an AI Overview appears Google sometimes suppresses the traditional snippet, but on many informational queries both can show, letting a page earn visibility twice.

First-party data is information a business collects directly from its own audience and customers — through its site, app, CRM, purchases, and consented interactions. It is the most accurate, durable, and privacy-resilient data a marketer can hold because it is owned, not rented from third parties. With third-party cookies deprecated and privacy rules tightening, first-party data has become the foundation for targeting, personalization, and attribution that still work. In 2026 it is also what powers reliable measurement of AI-driven and cross-channel journeys.

Full-funnel marketing is an approach that builds, measures, and optimizes across every stage of the customer journey — from awareness through consideration, conversion, and retention — as one connected system rather than isolated tactics. It rejects the trap of optimizing a single stage (like top-funnel reach or bottom-funnel conversion) at the expense of the whole. The payoff is compounding: demand created upstream converts more efficiently downstream when the stages are designed together. It requires unified data and attribution to see the full picture.

GEO (Generative Engine Optimization) is the practice of making your brand more likely to be cited and recommended inside generative AI responses from tools like ChatGPT, Perplexity, Gemini, and Claude. Where AEO targets AI search features, GEO targets the large language models that synthesize answers from across the web and name their sources. It works by building citable, well-structured content and the third-party authority signals these models trust. Because citation overlap between engines is low, GEO requires platform-specific strategy, not a single universal playbook.

Google AI Mode is a dedicated conversational search experience, powered by Gemini, that returns AI-generated answers with linked sources and supports follow-up questions and a query fan-out across related sub-queries. It launched to all US users in 2025 and has since expanded to more markets. AI Mode surfaces sources via inline links and supports user-set Preferred Sources, but its zero-click rate is far higher than standard search, making cited visibility critical.

Google Analytics 4 (GA4) is Google's current analytics platform and the replacement for the now fully sunset Universal Analytics, which stopped processing data in 2023 and had its data deleted starting the week of July 1, 2024. GA4 uses an event-based data model, in which every interaction is an event, rather than Universal Analytics' session-and-pageview model; it spans web and app, supports cross-platform measurement, and is built around consent signals and modeled conversions. It is the standard source for on-site behavior and conversion measurement.

Ideal Customer Profile (ICP) is a definition of the accounts or buyers most likely to convert, retain, and expand, built from firmographic, technographic, and behavioral traits of your best existing customers. Unlike a broad persona, an ICP narrows targeting and qualification so paid media, outbound, and content focus budget on high-fit demand. A sharp ICP is the foundation for account-based marketing, lead scoring, and product-led growth, since it determines who the funnel is built to serve.

Incrementality is the measure of outcomes (conversions, revenue, sign-ups) that happened because of marketing and would not have occurred otherwise. It is established with controlled experiments — most commonly geo lift tests or holdout/control groups — that compare an exposed group against an unexposed one to isolate causal lift rather than correlated touchpoints. As user-level tracking weakens, incrementality testing has become a primary way to validate channel value and calibrate models, answering which results did this spend actually cause?

Interaction to Next Paint (INP) is a Core Web Vital that measures a page's overall responsiveness by observing the latency of all user interactions, from input delay through event processing to the next visual update. A good INP is 200 milliseconds or less at the 75th percentile; above 500 ms is poor. INP replaced First Input Delay as a Core Web Vital on March 12, 2024, because it captures full interaction latency rather than only the first interaction's delay.

LTV (Lifetime Value) is the total revenue — or profit — a business can expect from a single customer across the entire relationship. It reframes growth around long-term worth rather than the first sale, which is what makes acquisition spend justifiable. Compared against CAC, LTV sets the ceiling on what you can profitably pay to acquire a customer. Improving LTV through retention and expansion is often cheaper than lowering CAC.

A landing page is a standalone page built for a single objective — capturing a lead or driving one action — that a visitor reaches from an ad, email, or search result. Stripping away site navigation and competing links keeps attention on one offer and one call to action, which is why dedicated landing pages typically convert better than sending paid traffic to a homepage. Message match between the ad and the page is the largest driver of performance, followed by load speed and a frictionless form. It is where paid spend either converts or evaporates.

A large language model (LLM) is an AI model trained on vast amounts of text to predict and generate human-like language, powering tools like ChatGPT, Gemini, and Claude as well as AI search features. LLMs answer from patterns learned during training, so they can produce fluent but inaccurate output (hallucinations) and have a knowledge cutoff. Techniques such as retrieval-augmented generation ground their responses in external, current sources to improve accuracy.

Lead scoring is the practice of assigning numeric or grade-based values to leads based on fit (how closely they match the ICP) and engagement (their behavior, such as content downloads, demo requests, or product activity) to prioritize follow-up and route the highest-intent leads to sales. Modern scoring is often predictive, using historical conversion data to weight signals automatically rather than relying on fixed point rules. It sits inside the marketing automation and CRM stack and directly feeds the MQL-to-SQL handoff.

Lifecycle marketing is the practice of tailoring messaging to a customer's stage in the relationship — from onboarding and activation through retention, expansion, and win-back — rather than treating every contact the same. It uses behavioral and first-party data to trigger the right message at the right moment, which compounds LTV by moving customers up instead of letting them lapse. In 2026, lifecycle programs increasingly run on agentic automation that personalizes timing and offers at the individual level rather than by broad segment. It is where retention and revenue growth are actually engineered.

A lookalike audience is a targeting type where an ad platform models a seed audience — such as a customer list or high-value converters — and finds new users who share similar attributes and behaviors. As of 2026, the tactic has shifted on Meta: with smaller, noisier pixel signals post-iOS tracking changes, classic 1% lookalikes often perform only marginally better than broad targeting, and Meta has folded lookalike expansion into Advantage+ audiences (its machine-learning targeting layer that treats your seed as a suggestion rather than a hard boundary). The durable best practice is to supply rich first-party seeds and let the system expand from there.

llms.txt is a proposed plain-text Markdown file placed at a site's root to give large language models a concise, curated guide to its most important pages. Proposed by Jeremy Howard of Answer.AI in September 2024, it remains an unofficial proposal: Google has confirmed it does not use llms.txt and that it is not a ranking signal, and no major AI provider has committed to it. Adoption is low, with most published files reportedly receiving no requests as of 2026.

MQL vs SQL describes the two key stages a lead moves through on the way to becoming a customer. An MQL (Marketing Qualified Lead) is a lead whose engagement — content downloads, site visits, demo interest — signals enough intent for sales to take a look, while an SQL (Sales Qualified Lead) is a lead that sales has vetted and accepted as a real opportunity worth pursuing. The handoff between them is where revenue is won or lost, which is why a shared, agreed definition of each stage matters. Misaligned MQL and SQL criteria are a leading cause of wasted leads and sales-marketing friction.

Marketing attribution is the practice of assigning credit to the marketing touchpoints that influence a conversion, so you can see which channels and campaigns actually drive revenue. Models range from simple first- or last-touch to multi-touch and data-driven approaches that weight every interaction along the path. In 2026 the hard part is the growing blind spot: traffic from AI answer engines often arrives without referrer data and lands in GA4 as Direct, undercounting AI-driven influence. Reliable attribution now leans on first-party data, server-side tracking, and incrementality testing rather than cookies alone.

Marketing automation is the use of software to run repetitive marketing tasks — email sequences, lead scoring, segmentation, and multi-step campaigns — triggered by customer behavior and data rather than manual sending. Through 2026 the category is shifting from rule-based workflows, which execute fixed logic until a human rewrites them, toward agentic systems that take an objective like reduce churn and decide the steps themselves. The practical payoff is consistency and scale: timely, personalized touches across the lifecycle without proportional headcount. It only works as well as the first-party data and clean systems underneath it.

Marketing mix modeling (MMM) is a statistical method, typically regression-based, that uses aggregated historical data to estimate how each marketing channel and other factors contribute to sales or other outcomes. Because it relies on aggregated rather than user-level data, MMM is privacy-safe by design and has seen renewed adoption as cookies and cross-platform tracking decline; open-source frameworks such as Google's Meridian and Meta's Robyn have lowered the barrier to building models. MMM is often paired with incrementality experiments for calibration and works alongside attribution for a fuller measurement picture.

MER (media efficiency ratio, also marketing efficiency ratio or blended ROAS) is total revenue divided by total marketing spend across all channels, measured at the business level rather than per campaign. Unlike channel-level ROAS, it does not rely on click attribution, so it sidesteps the over- and under-counting introduced by cookie loss and walled gardens — making it a steadier north-star for budget decisions in a privacy-first era. Typical 2026 DTC benchmarks run roughly 3x-5x, with mature subscription brands often above 6x.

Monthly Recurring Revenue (MRR) is the predictable subscription revenue a business normalizes to a monthly amount from all active recurring contracts, excluding one-time charges, setup fees, and variable usage. It is the core operating metric for subscription businesses, tracked as movements — new, expansion, contraction, and churned MRR — to show how the revenue base changes month over month. Annualized, MRR becomes ARR (MRR x 12).

NPS (Net Promoter Score) measures customer loyalty by asking how likely someone is to recommend a brand on a 0-10 scale, then subtracting the percentage of detractors (0-6) from the percentage of promoters (9-10) to produce a score from -100 to +100. Developed by Fred Reichheld and Bain & Company, it is widely used as a single, trackable read on loyalty and a leading indicator of retention and growth. Its credibility depends on acting on the follow-up why — Bain's Net Promoter 3.0 pushes teams to tie scores to actual earned growth rather than treat the number as a vanity metric. It is a signal to investigate, not a verdict on its own.

Net Revenue Retention (NRR) measures how much recurring revenue a cohort of existing customers generates over a period relative to the start of that period, after expansion, contraction, and churn, and excluding revenue from new customers. It is calculated as (starting MRR + expansion - contraction - churn) / starting MRR; above 100% means the existing base grows on its own even before new sales. In 2026 NRR is widely treated as the defining SaaS quality metric, with best-in-class public companies averaging roughly 120-125% and enterprise medians near 118%.

A North Star Metric is the single measure that best captures the core value a product delivers to customers and that, when it grows, reliably predicts sustainable revenue growth — for example, weekly active teams or messages sent. It aligns product, marketing, and growth teams around one leading indicator of value rather than a lagging financial output. A good North Star reflects customer value delivered, not just usage volume, so that optimizing it compounds into retention and expansion.

Omnichannel marketing is the practice of delivering a single, consistent customer experience across every channel and device — paid, organic, email, social, and offline — so each touchpoint reinforces the others. Unlike multichannel marketing, which simply runs many channels in parallel, omnichannel connects them around the customer so context carries from one to the next. It depends on unified data and integrated systems to recognize the same person across touchpoints. Done right, it removes friction and makes the whole journey feel like one conversation.

PPC (pay-per-click) is an advertising model in which advertisers pay only when a user clicks their ad, rather than for impressions. It spans paid search (e.g., Google Ads), shopping, and many social and display placements, with ad delivery and pricing typically set by real-time auctions that weigh bid and ad quality. PPC is the workhorse of intent-driven, measurable acquisition, where spend ties directly to clicks and downstream conversions.

Pipeline velocity is the rate at which revenue moves through the sales pipeline, calculated as (number of qualified opportunities x average deal value x win rate) / average sales cycle length. It quantifies how quickly the funnel converts demand into revenue and pinpoints whether to improve deal volume, deal size, win rate, or cycle time. Because it ties marketing-sourced pipeline to closed revenue, it is a primary lens for revenue operations and lead-handoff health.

The Privacy Sandbox is a Google initiative, begun in 2019, to develop web and Android technologies intended to support advertising and measurement with stronger privacy. Its direction changed substantially: in April 2025 Google decided to keep third-party cookies in Chrome under existing user controls rather than introduce a new prompt, and on October 17, 2025 it announced it would retire most of the advertising and measurement APIs — including Topics, Protected Audience, Attribution Reporting, Private Aggregation, and IP Protection — citing low adoption. Google continues to support privacy and identity features such as CHIPS, FedCM, and Private State Tokens, and is pursuing an interoperable Attribution standard through the W3C.

Product-Led Growth (PLG) is a go-to-market motion in which the product itself drives acquisition, activation, and expansion — typically through free trials, freemium tiers, or self-serve onboarding — so users experience value before talking to sales. Growth marketing in a PLG model focuses on driving qualified sign-ups, improving activation and time-to-value, and identifying product-qualified leads (PQLs) for sales or expansion. It often complements, rather than replaces, sales-led motions for larger accounts.

POAS (profit on ad spend) is the gross profit generated per dollar of ad spend, calculated after deducting cost of goods sold and other variable costs — not gross revenue. It corrects the blind spot in ROAS: a 4.0 ROAS on a product with a 20% margin is actually a losing campaign once costs are counted. By steering bids and budgets toward profit rather than top-line return, POAS aligns paid media with the actual business outcome and is increasingly fed into smart-bidding via profit-margin data.

Programmatic advertising is the automated buying and selling of digital ad inventory using software, data, and auctions, connecting advertisers' demand-side platforms (DSPs) with publishers' supply-side platforms (SSPs) through ad exchanges. Much of it runs via real-time bidding (RTB), where each impression is auctioned in milliseconds as a page or app loads, on standards such as IAB Tech Lab's OpenRTB. By 2026 the large majority of digital display spend transacts programmatically, and the channel is adapting to cookie deprecation through first-party data, contextual signals, and privacy-preserving identifiers.

ROAS (Return on Ad Spend) is the revenue generated for every dollar spent on advertising, calculated as revenue from ads divided by ad spend. A 4:1 ROAS means $4 in revenue for every $1 invested. It is the fastest read on whether a campaign or channel is paying off, though it measures gross efficiency rather than profit — so it is strongest when paired with margin, CAC, and LTV. Used alone, a high ROAS can still hide an unprofitable funnel.

A retail media network (RMN) is an advertising business run by a retailer that lets brands buy ad placements across the retailer's owned channels — on-site search and product pages, its app, and off-site or in-store inventory — using the retailer's first-party purchase data for targeting and closed-loop sales measurement. RMNs are one of the fastest-growing ad channels: eMarketer forecasts U.S. retail media ad spend to reach about $69.33 billion in 2026, up roughly 18% year over year and approaching 18% of total U.S. digital ad spend. Their value lies in proximity to the point of purchase and first-party data that survives signal loss from cookie deprecation.

Retargeting is the practice of serving ads to people who previously interacted with a brand — site visitors, cart abandoners, or app users — to bring them back to convert. Historically powered by third-party cookies for cross-site tracking, it has been reshaped by browser privacy changes and cookie deprecation, which shrink the reachable pools and blur the old retargeting/remarketing distinction. In 2026 effective retargeting leans on consented first-party data, server-side tracking, and platform audiences built from owned signals.

Retention rate is the percentage of customers a business keeps over a period, calculated as customers retained divided by customers at the start, excluding new ones acquired during the window. It is the inverse of churn and the engine behind LTV: small, sustained gains in retention compound into outsized revenue because retained customers cost nothing more to acquire and tend to spend more over time. It is usually the highest-leverage metric a growth team can move, since improving it lifts both profit and the budget you can justify for acquisition. Retention is built through onboarding, product value, and lifecycle programs, not discounts alone.

Retrieval-augmented generation (RAG) is an AI technique that retrieves relevant information from an external knowledge source at query time and feeds it to a large language model as context before it generates an answer. By grounding output in current, authoritative data, RAG improves accuracy, reduces hallucinations, and lets models use information beyond their training cutoff without retraining. It underpins many AI search and assistant products that cite their sources.

RevOps (Revenue Operations) is the discipline of aligning marketing, sales, and customer success around a single revenue process, shared data, and connected systems. It breaks down the silos where leads get lost between teams and where reporting contradicts itself. RevOps owns the tooling, data hygiene, and handoffs that let the whole go-to-market motion run as one engine. The result is cleaner forecasting, faster cycles, and fewer leaks between first touch and renewal.

A SERP (search engine results page) is the page a search engine returns in response to a query. Modern SERPs blend organic listings, paid ads, and an expanding set of features, including AI Overviews, featured snippets, knowledge panels, and local packs, that often answer the query directly on the page. As these features grow, a rising share of searches end without a click to any website, reshaping how visibility is measured.

Server-side tracking (server-side tagging) moves the collection and distribution of measurement data from the browser or app to a server-side container you control, which then forwards events to analytics and advertising platforms. Running in a first-party context keeps data and cookies on the brand's own domain, lets teams strip or normalize data (including PII) before it reaches third parties, and improves page performance and data reliability against ad blockers and browser restrictions. It commonly underpins implementations of GA4 and Meta's Conversions API.

Structured data is standardized code, usually schema.org vocabulary added in JSON-LD, that describes a page's content so search engines can understand it and display rich results such as reviews, FAQs, products, events, and breadcrumbs. Google recommends JSON-LD and maintains its own list of supported features, which it treats as definitive over schema.org for Search behavior. The supported set changes over time; for example, Google removed the practice-problem feature in January 2026.

A third-party cookie is a cookie set by a domain other than the one a user is visiting, historically used for cross-site advertising, retargeting, and conversion tracking. As of June 2026 they remain available in Chrome: in April 2025 Google decided to maintain its existing approach of offering users third-party cookie controls in Chrome's settings rather than deprecating cookies by default or adding a standalone prompt — reversing its earlier phase-out plans. They are already blocked by default in Safari, Firefox, and Chrome's Incognito mode, so most marketers continue shifting to first-party and server-side measurement.

UGC (User-Generated Content) is content created by customers and creators rather than the brand — reviews, photos, videos, and social posts — used in organic and paid channels. It outperforms polished brand assets in many feeds because it reads as authentic and provides social proof, which is why creator-style UGC has become a staple of paid social creative. In practice, brands now commission UGC-style content at scale and rotate it constantly to fight ad fatigue. Its strength is trust; its risk is rights and disclosure, which need to be managed deliberately.

A zero-click search is a search that ends without the user clicking through to any website, because the answer is satisfied on the results page itself by features like AI Overviews, featured snippets, or knowledge panels. According to a SparkToro study published in June 2026, 68% of US Google searches ended without a click in early 2026, up from about 60% in 2024. The trend pushes brands toward optimizing for on-SERP visibility and AI citations rather than clicks alone.

Zero-party data is information a customer intentionally and proactively shares with a brand — such as preferences, purchase intentions, and how they want to be recognized — collected through preference centers, surveys, quizzes, and similar opt-in tools. A term popularized by Forrester, it is distinguished from first-party (behavioral) data because it is explicitly volunteered rather than observed or inferred, making it accurate, consent-based, and well suited to privacy-first personalization. It is a core input for CDPs and retention programs as tracking-based signals decline.