The decision-first growth glossary.

An account hierarchy is the structured map of how companies relate: parent, subsidiaries, divisions, and locations, so signals, spend, and decisions roll up to the right entity. It is what makes account-level truth possible in B2B data.

Why it matters: Without hierarchy, ten subsidiaries look like ten small accounts, intent fragments, and the buying committee splits across records. Rolled up correctly, the real opportunity becomes visible and decisions land at the right level.

Related: Master data management · Buying committee

An After-Action Review (AAR) is the structured retro a decision loop runs after every cycle: what was decided and why, what happened, what the causal impact was, and what changes next cycle. iCustomer automates it: Decision Traces record the why, causal measurement scores the outcome, and the learning feeds outcome context.

Why it matters: Teams skip retros when they are manual, which is why most marketing never compounds. An automated AAR is how every dollar teaches your loop instead of only teaching the ad platforms.

Related: Causal measurement · Decision loop

Agentic commerce is buying where discovery itself runs through AI agents: buyers ask assistants like ChatGPT, Gemini, and Perplexity, agents research and compare across brands, and increasingly complete the purchase, instead of humans browsing websites page by page. Emerging protocols (ACP, UCP, MCP, A2A) define how agents discover products and transact.

Why it matters: Agent-mediated purchases arrive with no clicks and no sessions, so brands need durable identity, structured first-party data, and causal measurement to stay discoverable, chosen, and provable. See the Agentic Commerce Blueprint.

Related: Agentic marketing · AEO · Causal measurement

Agentic marketing is marketing operated by AI agents that plan, decide, and act toward goals under human oversight, rather than humans clicking through tools with AI-assisted steps. The human role shifts from operating software to setting goals, taste, and guardrails.

Why it matters: Agents act at machine speed, so ungoverned agents compound mistakes at machine speed too. The winners will pair agent execution with decision governance, not just better prompts.

Related: Decision loop · Human-in-the-loop AI · Agentic commerce

Answer engine optimization (AEO) is the practice of making your content the answer AI assistants quote: definition-first sentences, self-contained answers that name their subject, question-shaped headings, and FAQ schema. Where SEO earns a ranking on a results page, AEO earns the citation inside ChatGPT, Claude, Perplexity, and AI Overviews.

Why it matters: A growing share of buyer research never touches a results page. Brands that define their category's terms crisply get quoted; everyone else gets summarized, or skipped.

Related: Content engine · Agentic commerce

Audience context is everything a decision needs to know about the person or account it concerns: durable identity, live signals, interests, lifecycle stage, and history, assembled at decision time. It is one of iCustomer's four context pillars, alongside brand, domain, and outcome context.

Why it matters: Generic AI guesses; grounded AI decides. Audience context built on OneSource IDs is the pillar no competitor can copy, because it is made from your data and your relationships.

Related: Context pillars · Audience intelligence

Audience intelligence is the continuous practice of knowing who your audience is, what they care about right now, and what to do about it: identity, signals, interests, and scores unified into one living, actionable picture. It goes beyond analytics (what happened) to decision-ready knowledge (who, why, what next).

Why it matters: Every downstream motion, paid, lifecycle, outbound, and agents alike, is only as good as the audience picture it runs on. iCustomer is built as an audience intelligence platform: the Audience Interest Graph is the picture, decision loops are the action.

Related: Audience Interest Graph · Signals-based marketing

An Audience Interest Graph is a living map of your audience and its evolving interest: every person and account, resolved to a durable identity, connected to the signals, content, and behaviors that show what they care about right now. Unlike a static segment, it decays, refreshes, and sharpens as interest changes.

Why it matters: Audiences built last quarter target who people used to be. A graph that tracks evolving interest targets who they are today, and it lives in your data cloud as your asset.

Related: Signals Hub · FIRE score

Audience suppression removes the people who should not see a message, such as current customers, recent purchasers, open support cases, and opted-out users, from targeting before any spend happens. It applies equally to ads and outbound sequences.

Why it matters: Suppression is the cheapest wasted dollar to save and the easiest brand-trust failure to prevent. In outbound, synced suppression also protects deliverability.

Related: First-party data · Next best action

Buyer discovery is the work of finding who is actually in-market for you right now: revealing anonymous brand and website visitors, resolving them to real accounts and people, and surfacing the buyers your ICP predicts but your CRM has never met. It runs continuously, not as a quarterly list pull.

Why it matters: Most in-market buyers never fill out a form. Discovery through the Signals Hub turns invisible demand into targetable audience before your competitors see it.

Related: Signals Hub · ICP

A buying committee is the group of people who collectively decide a B2B purchase: champions, economic buyers, technical evaluators, end users, and blockers, often six to ten people or more. They research on different surfaces at different times.

Why it matters: Deals stall when only one contact ever hears from you. Account-level decisioning reaches the whole committee across its attention surfaces, ads, email, and sales touches, at once.

Related: Account hierarchy · Buyer discovery

Causal measurement answers what a marketing action actually caused, its incremental impact, rather than what it merely touched. Where last-touch attribution credits whatever was nearby when a sale happened, causal AI compares against what would have happened anyway.

Why it matters: You cannot compound what you cannot measure truthfully. Causal measurement is what lets a decision loop learn: every outcome feeds back as ground truth, so the next cycle targets sharper and wastes less.

Related: Decision loop · Decision Activation Gap

A composable CDP delivers customer data platform capabilities, such as profiles, segments, and activation, directly on your data warehouse instead of copying your data into a vendor's cloud. You assemble best-of-breed components on one source of truth rather than buying a monolithic bundle.

Why it matters: Every copy of customer data is cost, lag, and governance risk. Composable architecture keeps the warehouse as the system of record; iCustomer is built to activate the CDP investments you already made, not replace them.

Related: Warehouse-native activation · Identity resolution

Consent capital is the accumulated, provable permission a brand has earned from its audience: every opt-in, preference, and consent state treated as a compounding asset rather than a compliance checkbox. It grows with trust and is spent, or destroyed, by misuse.

Why it matters: In a privacy-first era, brands that can prove consent can activate everywhere; brands that cannot are locked out of their own audience. iCustomer carries consent with every payload, so activation never spends capital you have not earned.

Related: First-party data · Decision governance

A content engine is the repeatable system that turns strategy and expertise into a steady flow of content: pillars, briefs, production, distribution, and measurement running as one loop rather than one-off campaigns. Cadence and reuse beat bursts of volume.

Why it matters: Consistency compounds, and volume without a system is slop. In an AEO era the engine's job widens: teach humans and stay quotable to machines at the same time.

Related: AEO · Customer lifecycle

The four context pillars are the grounding every iCustomer decision draws on: audience context (who this person or account is, powered by OneSource immutable IDs and live signals), brand context (your goals, taste, guardrails, and taxonomy), domain context (industry ontology and proven plays), and outcome context (Decision Traces and outcome memory from every prior cycle).

Why it matters: Generic AI knows none of these; that is why generic AI output reads generic. Context is the difference between automation and decisioning, and audience context, identity you own, is the pillar competitors cannot rent.

Related: OneSource · Decision Fabric

The customer lifecycle is the full arc of a customer's relationship with a brand: awareness, consideration, purchase, onboarding, retention, expansion, and advocacy, or churn and win-back. Each stage produces different signals and calls for different decisions.

Why it matters: Most stacks over-invest in acquisition and under-decide everything after it. Decision loops run per lifecycle goal, churn, cross-sell, win-back, so no stage runs on autopilot.

Related: Decision loop · Next best action

A data clean room is a privacy-safe environment where two parties match and analyze their audiences without either side seeing the other's raw data. Common uses: brand-retailer partnerships, publisher matches, and measurement inside walled gardens.

Why it matters: Clean rooms let you collaborate on audiences you could never legally or safely share. They pair naturally with an identity spine you own and consented, minimal payloads.

Related: Identity resolution · Retail media network

The Decision Activation Gap is the distance between owning first-party data and activating it: brands store rich customer data in their warehouses, yet less than 5% of ad spend is informed by it. The data is in your warehouse; the decisions live nowhere.

Why it matters: Closing this gap is worth more than any new tool: it means every dollar of spend is informed by data you already paid to collect. iCustomer exists to close it with decision loops that read your data where it lives and activate it in the channels you already run.

Related: Decision loop · Warehouse-native activation

A Decision Fabric is a trusted decision layer that both humans and AI agents work from: one place where decisions are made, governed, and audited. iCustomer's Decision Fabric combines a decision engine (causal AI plus a Decisioning Waterfall), a Harness of guardrails, evals, and controls for every agent action, and four context pillars, audience, brand, domain, and outcome, plus Decision Traces auditing every action, so every decision is grounded rather than guessed.

Why it matters: In the human-and-agent era, agents without shared context and guardrails create sprawl and risk. A fabric gives every human and agent the same ontology, the same rules, and the same audit trail.

Related: Context pillars · Decision loop

Decision governance extends data governance to the decisions made on data: who or what, human or agent, may decide, within which guardrails, and with what audit trail. Data governance controls access to information; decision governance controls action.

Why it matters: The moment agents act on customer data, "who can see it" is no longer the hard question; "who approved that action, and can we trace it" is. iCustomer's Harness and Decision Traces exist to make every automated decision governed and auditable.

Related: Decision Fabric · Human-in-the-loop AI

A decision loop is a goal-oriented, always-on system: give it a goal and a target, and it observes, orients, decides, acts, then learns from every outcome and repeats. In growth marketing, that means reading audience signals, deciding who to reach, when, and with what, acting in your channels, and letting every cycle teach the next. A brand runs many decision loops at once, one per goal, across acquisition, conversion, retention, churn, and expansion.

Why it matters: Campaigns end; loops compound. Every cycle a loop runs, targeting sharpens and waste drops, and the learning stays your asset instead of a vendor's. iCustomer builds and runs decision loops on your data cloud.

Related: Decision Activation Gap · Decision Fabric · Causal measurement

FIRE is iCustomer's real-time scoring framework: Fit (how well someone matches your ideal profile), Intent (buying signals right now), Recency (how fresh those signals are), and Engagement (depth of interaction). Every person and account carries a composite FIRE score, recomputed continuously from your first-party signals.

Why it matters: FIRE solves the intent problem: knowing who is genuinely in-market today, not who landed in a segment last quarter. Those scores drive who to reach, when, with what, by segment and one-to-one.

Related: Signals Hub · Audience Interest Graph

First-party data is the data a brand collects directly from its own audience, with consent: site and product behavior, purchases, CRM records, email and SMS engagement. It is owned, permissioned, and unavailable to competitors, unlike second-party data (a partner's first-party data) or third-party data (aggregated and rented).

Why it matters: As third-party signals disappear, first-party data is the only durable edge, yet less than 5% of ad spend is informed by it. Activating it is the whole point of a decision loop.

Related: OneSource · Decision Activation Gap

Human-in-the-loop AI keeps people at the decision points that matter: agents propose and execute within approved bounds, while humans set goals, approve exceptions, and audit outcomes. The loop is designed so oversight scales without becoming a bottleneck.

Why it matters: Your team's taste is the edge no competitor can copy, and it only compounds if it stays in the loop. Full autonomy without checkpoints is how brands ship AI slop.

Related: Decision governance · Decision Fabric

iConnect is iCustomer's meta orchestrator: one router that takes each decision from a loop and sends it to whichever connected tool should execute it, across 200+ integrations, inside the Harness guardrails. Teams and agents call one API, via the MCP Server, the CLI, or the app, instead of wiring every tool to every other tool.

Why it matters: Tool bloat compounds: every new tool multiplies integrations, permissions, and leak surface. One governed router means one contract, one audit trail (every action leaves a Decision Trace), and no orphaned automations.

Related: Decision Fabric · Decision loop

An ideal customer profile (ICP) is the definition of the accounts and people most likely to buy, succeed, and expand: firmographics, technographics, and behaviors drawn from your best customers. It is a hypothesis to keep testing, not a document to file.

Why it matters: Fit is the F in FIRE. A sharp ICP concentrates spend on winnable audience; a vague one quietly subsidizes audiences that were never going to convert.

Related: FIRE score · Buyer discovery

Identity resolution links the many identifiers one person or account produces, including emails, devices, cookies, CRM IDs, and offline records, into a single durable profile. Done well, it is deterministic where possible and probabilistic only where necessary.

Why it matters: Everything downstream inherits identity quality: match rates, scores, personalization, and measurement all break on a fractured profile. OneSource anchors identity in immutable IDs you own in your cloud.

Related: OneSource · Match rate

Intent data is behavioral evidence that a person or account is actively researching a problem or product: content consumption, search patterns, review-site activity, and site visits, collected first-, second-, or third-party. Its value decays in days, not quarters.

Why it matters: Intent is the I in FIRE, and stale intent is noise. The signals waterfall keeps it fresh across 50+ pre-unified sources instead of one vendor's panel.

Related: Signals-based marketing · FIRE score

A lookalike audience is a platform-generated audience of new people who resemble a seed audience you provide, typically your best customers. Quality follows the seed: a clean, high-value seed models your best customer; a stale one models the wrong customer at scale.

Why it matters: Lookalikes are the fastest legitimate way to grow beyond your list. Seeding them from FIRE-scored, identity-resolved audiences is what keeps expansion from becoming waste.

Related: Match rate · First-party data

Master data management (MDM) is the discipline of maintaining one authoritative, deduplicated record, often called a golden record, for core entities like customers, accounts, and products across every system that touches them. It is the enterprise answer to "which record is true."

Why it matters: Decisions grounded in conflicting records are guesses. MDM gives the enterprise trustworthy entities; a decision layer turns those entities into prioritized action. iCustomer reads MDM platforms like Reltio as a source.

Related: Identity resolution · Decision governance

Match rate is the percentage of an uploaded audience that an ad platform can match to real, targetable users. Raw CRM lists typically match 20-40%; resolving identities and formatting identifiers correctly can lift the same list to 70-90%.

Why it matters: Match rate is the multiplier on every paid dollar: a 30% match means 70% of your audience, and the budget aimed at them, never reaches the platform. It is the cheapest performance lever most teams never touch.

Related: Identity resolution · First-party data

Media mix modeling (MMM) is a top-down statistical method that estimates each channel's contribution to revenue from aggregate spend and outcome data, with no user-level tracking required. It regained favor as privacy changes weakened tracking-based attribution.

Why it matters: MMM sees the forest, incrementality tests verify the trees. Mature teams triangulate: MMM for budget allocation, causal experiments for decisions, and neither replaces the other.

Related: Causal measurement

A next best action (NBA) is the single highest-value thing to do for a specific person or account right now: an offer, a message, a channel switch, or deliberate silence, chosen from every possibility by a decision engine. It is decisioning at the individual level, continuously refreshed.

Why it matters: Segments say who; NBAs say what and when. They are the concrete output a decision loop delivers into your CRM, lifecycle, and sales tools every day.

Related: Decision loop · FIRE score

OneSource is iCustomer's identity solution: immutable, HMAC-anchored IDs you own in your data cloud, extended by the signals waterfall's 50+ pre-unified third-party sources. It is how a brand owns its audience rather than renting a network key.

Why it matters: Identity networks make your audience knowledge a shared asset; owned identity keeps it yours. OneSource anchors identity in your warehouse and still reaches the match rates networks promise, typically lifting 20-40% to 70-90% on paid channels.

Related: Audience Interest Graph · Warehouse-native activation

A retail media network (RMN) is a retailer's advertising business: it packages the retailer's shopper audiences and sells brands access to them, on its own properties and across the open web. RMNs are the fastest-growing segment of digital advertising.

Why it matters: An RMN is only as good as the audience data underneath it. Retailers need clean identity, fresh signals, and privacy-safe packaging to monetize; brands need measurement to know it worked.

Related: First-party data · Data clean room

Reverse ETL syncs data out of the warehouse into the operational tools where work happens, such as CRM, ad platforms, and email, turning the warehouse into a source of action rather than only analysis. It is the plumbing of the composable stack.

Why it matters: Pipes move data; they do not decide what to do with it. Reverse ETL answers "how do I sync this segment," while a decision layer answers "who belongs in it, when, and why."

Related: Warehouse-native activation · Next best action

A Signals Hub reveals and collects the signals your audience produces, including website and brand visitors you could not previously identify, and a signals waterfall resolves those signals in priority order across 50+ pre-unified sources so every profile stays current. Together they keep FIRE scores dynamic and audiences always on.

Why it matters: Most brands see a fraction of their own traffic and buy stale intent data. A hub-plus-waterfall design means fresher signals, from sources you choose, feeding scores that change the moment behavior does.

Related: FIRE score · OneSource

Signals-based marketing replaces static lists and campaign calendars with a continuous read of buying signals, such as site visits, intent surges, hiring, product usage, and engagement, and acts the moment a signal fires. The unit of work shifts from the campaign to the signal.

Why it matters: Buyers show intent long before they fill out a form. Teams that read and act on signals reach them first; iCustomer's signals waterfall and FIRE scoring exist to make this continuous.

Related: Signals Hub · FIRE score

Warehouse-native activation reads and activates your customer data where it already lives, Snowflake, Databricks, BigQuery, with zero data copies and zero raw-data egress, instead of duplicating it into a vendor platform. Composable means it works alongside your existing CDP, CRM, and tools rather than replacing them.

Why it matters: Every copy of your data is a liability and a lag. Computing in place means governance stays yours, PII stays protected, and only consented activation payloads ever leave, no migration, no rip-and-replace.

Related: Decision Activation Gap · OneSource