Marketing Data Glossary: 95 Terms Every Modern Team Should Know
Marketing runs on data—but the language around marketing data isn’t always consistent.
At its core, marketing data is the information marketers use to understand people and improve performance based on that understanding of who people are (attributes), what they do (behaviors), and what they’re likely to do next (predictions). It powers everything from building audiences and personalizing messaging to activating campaigns across channels, understanding consumer behavior, improving customer experiences and measuring what actually drives results.
But if you work in marketing ops, growth, media, analytics, or data science, marketing data terms are often used differently and have to be translated between business and technical teams.
Terms like people-based data, predictive modeling, behavioral data, and incrementality get used every day—but they don’t always mean the same thing to everyone in the room.
This glossary is designed to create a shared language for modern marketing teams. While most glossaries stop at definitions, this one goes further by explaining:
Why it matters (what it impacts in the real world)
How it’s used (where it shows up in workflows)
Common confusion (for the terms that get mixed up most)
Let’s get started.
Key Terms at a Glance
If you only learn a handful of terms, start here:
- Marketing data: The information marketers use to understand audiences and improve performance.
- Predictive modeling: Building models that estimate outcomes using patterns in historical data.
- People-based data: Data tied to stable person/household identifiers (not just devices) to support consistent audience understanding across channels.
- Deterministic data: Identity or attributes matched using direct, known or high-confidence identifiers.
- Predictive data: Modeled attributes that estimate the likelihood of a behavior or attribute.
- Motivation data: Signals that describe why people make decisions (decision drivers, values, attitudes) used to improve targeting and messaging relevance.
- Data enrichment: Adding attributes or signals to your existing customer/prospect records.
- Cognitive science: The interdisciplinary study of how people think, learn, perceive, and make decisions (drawing from psychology, neuroscience, behavioral economics, and more).
- Decision drivers: The underlying psychological factors (needs, priorities, motivations) that influence what someone chooses and when they act.
Marketing Data Types
| Term | Definition | Why It Matters | How It’s Used |
|---|---|---|---|
| 1. Demographic Data | Attributes that describe who a person is such as their age, gender, ethnicity, occupation, marital status, education, household composition, and more. | Demographics help with basic segmentation, messaging fit, and compliance-sensitive targeting choices. | Audience creation, personalization, suppression, and reporting. |
| 2. Behavioral Data | Signals based on actions someone has taken (e.g., browsing patterns, content consumption, engagement, lifestyle, interests, etc.). | Behaviors often correlate more directly with short-term intent than demographics. | Retargeting, interest audiences, churn risk signals, recency-based segments. |
| 3. Purchase Data (aka Transaction Data) | Data describing what people have bought, when, and sometimes how often. | Purchase history is one of the strongest indicators for future category demand. | Prospecting lookalikes, cross-sell/upsell, suppression (don’t target recent buyers), modeling inputs. |
| 4. Psychographic Data | Data describing attitudes, motivations, preferences, and decision drivers. | Psychographics help explain why people decide and is useful for messaging and creative alignment. | Persona creation, creative strategy, audience refinement. |
| 5. Motivation Data | Data that describes the underlying why behind consumer decisions including the attitudes, values, needs, and decision drivers that influence choices. | Two people can look identical demographically and behave similarly online, yet respond to entirely different messaging. Motivation data helps teams move beyond “who” and “what” to predict what will resonate and what they’re likely to do next. | Persona development, message and creative alignment, audience refinement, predictive modeling features, and prioritizing which segments to activate first. |
| 6. People-based Data | Data organized around people/households using stable identity, designed for consistent understanding across channels. | Devices change; people don’t (at least not as fast). People-based approaches reduce fragmentation. | Cross-channel audience building, suppression, measurement, personalization. |
| 7. Predictive Data | Model outputs (often scores) that estimate likelihood of a behavior or attribute. | Turns complex signals into usable targeting inputs. | High-propensity audiences, suppression of low-likelihood segments, prioritization. |
| 8. Audience Data | Data packaged into segments (audiences) that can be activated in marketing platforms and across channels. | Data only creates value when it becomes usable audiences in real channels. | Targeting, exclusions, sequencing, measurement. |
| 9. First-Party Data | Data you collect directly from your customers and owned touchpoints (site/app/CRM). | It’s usually the most accurate for your business and supports strong personalization. | Lifecycle marketing, suppression, measurement, modeling, customer analytics. |
| 10. Second-Party Data | Another company’s first-party data shared through a direct partnership. | Can be high quality and highly relevant—if the partnership is aligned and permissioned. | Joint campaigns, co-marketing audiences, collaboration programs. |
| 11. Third-Party Data | Data obtained from an external provider, not collected directly by you. | Expands reach beyond your customer base and supports acquisition scale. | Prospecting audiences, enrichment, modeling inputs, suppression. |
| 12. Zero-Party Data | Data a customer intentionally and proactively shares (preferences, needs, intentions). | It’s explicit—great for personalization—but often limited in scale. | Preference centers, onboarding surveys, guided experiences. |
| 13. Co-op Data (aka Cooperative Data) | Co-op data is marketing data contributed by multiple participating organizations into a shared database (a “co-op”), where the combined information is used to create insights, attributes, and audiences that individual members can use typically under defined governance and usage rules. | Because it pools signals across many contributors, co-op data can provide broader coverage, richer attributes, and stronger performance insights than a single brand’s first-party data alone—especially for acquisition and segmentation. | Prospecting audiences, modeled segments, suppression (where permitted), enrichment of customer/prospect files, and analytics to identify high-value characteristics that correlate with conversion. |
| 14. Intent Data | Signals indicating research or interest in a topic/category, often from content consumption or searches. | Helps find in-market audiences earlier in the funnel. | B2B account targeting, mid-funnel prospecting, content personalization, conversion. |
| 15. Contextual Data | Signals based on the environment (content/page/program) rather than a person’s identity. | Privacy-resilient and effective for aligning message to moment. | Contextual targeting, brand safety, program selection. |
| 16. Attitudinal Data | Data capturing opinions, preferences, beliefs, and stated intent (what people say they value or plan to do). | Attitudes can predict receptivity and brand fit even before behavior shows up—especially for emerging categories or infrequent purchases. | Persona work, message targeting, model inputs, survey-based segmentation, and creative testing. |
|
Common confusion (Attitudinal vs Behavioral): |
|||
| 17. Location Data | Data indicating where someone is or has been (often modeled or aggregated). | Useful for local relevance, footfall measurement, and geo-based segmentation. | Geo-targeting, store visit measurement, trade area analysis. |
| 18. Event Data | A record of actions taken (e.g., “added to cart,” “watched video,” “opened email”). | Enables automation and timely messaging. | Journeys, triggers, retargeting, attribution. |
| 19. Consumer Motivations | The internal needs or goals that push someone toward a decision (e.g., security, convenience, | Motivation explains why certain benefits convert and others fall flat—especially when behaviors are noisy or inconsistent. | Creative strategy, offer positioning, segmentation, channel sequencing (e.g., education-first vs urgency-first messaging). |
| 20. Decision Drivers | The specific psychological factors that most strongly influence an individual’s choices, tradeoffs, and timing (what they prioritize when deciding). | Decision drivers help you predict which message angle is most persuasive—and reduce wasted impressions from irrelevant creative. | Audience messaging frameworks, variant testing (creative x audience), personalization rules, and model features for propensity scoring. |
| 21. Consumer Psychology | The study of how people choose, buy, and form preferences—shaped by emotion, identity, social influence, and context. | Many campaigns fail not because targeting is wrong, but because the message doesn’t match how the audience evaluates value, risk, and trust. | Message mapping, creative development, persona creation, lifecycle communications, testing hypotheses about what will drive action. |
Identity + Matching
| Term | Definition | Why It Matters | How It’s Used |
|---|---|---|---|
| 22. Identity Resolution | The process of connecting identifiers (email, device IDs, etc.) to represent a person/household consistently. | Reduces duplication, improves frequency management, and enables cross-channel activation. | Audience building, suppression, measurement, personalization. |
| 23. Deterministic Matching | Identity matching based on direct, high-confidence identifiers (e.g., the same hashed email). | Typically yields higher accuracy than inference-based matching. | Onboarding, CRM matching, audience creation. |
| 24. Probabilistic Matching | Identity matching based on patterns and likelihood (e.g., signals suggesting two devices belong to the same person). | Can expand reach when deterministic identifiers are unavailable—at the cost of potential noise. | Cross-device mapping, reach extension, identity graphs. |
|
Common confusion (Deterministic vs. Probabilistic): Deterministic is direct match; probabilistic is inferred match. Deterministic often wins on precision; probabilistic can help on scale. |
|||
| 25. Identity Graph | A database that links identifiers (emails, devices, households) to represent relationships across channels. | Enables consistent targeting and measurement across fragmented identifiers. | Onboarding, cross-channel activation, suppression, frequency management. |
| 26. Household Graph | A data framework that connects devices and identifiers belonging to the same household | Enables cross-device targeting and measurement. | Unifies TV, mobile, and digital signals for audience activation and attribution. |
| 27. Device Graph | A type of identity graph focused on connecting devices to individuals or households. | Helps reduce duplicate reach and improve sequencing. | Cross-device targeting, measurement, frequency control. |
| 28. Onboarding | The process of converting offline/CRM identifiers into platform-usable audiences. | Turns customer data into actionable segments in ad platforms. | Custom audiences, suppression, lookalikes. |
| 29. Hashing | Converting identifiers (like email) into a fixed string so they can be matched without exposing the raw value. | Supports privacy-minded matching workflows. | Data onboarding, platform matching. |
| 30. Hashed Email (HEM) | An email address that’s been converted into a fixed, non-readable string using a one-way hashing method, so it can be used as a privacy-protective identifier for matching records across systems without sharing the raw email. | Common identifier for audience matching and onboarding. | Custom audience creation, identity resolution. |
| 31. Mobile Ad ID (MAID) | A device-level identifier on mobile (e.g., IDFA/GAID) used for advertising and measurement (where available). | Historically important for mobile targeting/measurement; availability and usage vary by ecosystem. | Mobile audience targeting, measurement, attribution. |
| 32. Cookie | A browser-based identifier used to recognize users on a site or across sites (depending on type/permissions). | A long-time foundation for digital targeting and measurement; increasingly constrained. | Site analytics, personalization, retargeting. |
| 33. Consent | Permission given by a user for data collection and/or use under applicable policies/laws. | Determines what you can legally and ethically do with data. | CMPs, activation restrictions, governance. |
| 34. Clean Room | A controlled environment where parties can analyze or match data with privacy protections and restrictions. | Enables collaboration while limiting direct data sharing. | Measurement, audience insights, partner collaboration. |
|
Common confusion (Identity Graph vs. Clean Room): |
|||
Analytics + Modeling
| Term | Definition | Why It Matters | How It’s Used |
|---|---|---|---|
| 35. Predictive Analytics | Using data and statistical methods to forecast likely future outcomes (e.g., propensity to purchase). | Helps prioritize spend and tailor messaging based on likelihood—not guesses. | Audience scoring, targeting prioritization, churn prevention. |
| 36. Predictive Modeling | Building models that estimate outcomes using patterns in historical data. | Can outperform rules-based segmentation when done and validated correctly. | Propensity scores, risk scores, next-best-action, response scores, conversion scores, and attribute, behavior, or motivation predictions. |
| 37. Cognitive Science | The interdisciplinary study of how people think, learn, perceive, and make decisions (drawing from psychology, neuroscience, behavioral economics, and more). | Marketing performance often improves when you align campaigns with how people actually decide—not how we assume they decide. Cognitive science gives you evidence-based ways to understand attention, memory, and choice. | Designing segmentation strategies, choosing persuasion levers, creating research-based attributes, and building models that reflect human decision-making patterns. |
| 38. Cognitive Bias | A predictable shortcut in thinking that influences decisions (often unconsciously), such as favoring familiar brands or over-weighting recent experiences. | Biases affect conversion, loyalty, and response to offers—understanding them helps you craft messaging that matches real human behavior. | Creative strategy, offer framing, UX messaging, segmentation hypotheses, and interpreting performance patterns. |
|
Common confusion (Bias vs. Data bias): |
|||
| 39. Propensity Score | A score estimating the likelihood that someone will take a specific action (e.g., purchase, subscribe). | Helps allocate budget toward the most responsive audiences. | Prospecting, retargeting prioritization, personalization. |
| 40. Lift | The improvement in performance attributable to a tactic or audience relative to a baseline. | Shows whether something actually improved results versus doing nothing or doing “business as usual.” | Audience testing, creative testing, media optimization. |
| 41. Holdout Group | A portion of an audience intentionally excluded from marketing to create a comparison baseline. | Enables incrementality measurement. | Lift tests, incrementality testing, causal analysis. |
| 42. Incrementality | The results that happened because of marketing—beyond what would have happened anyway. | Prevents you from rewarding channels for conversions they didn’t truly drive. | Testing frameworks, budget allocation. |
| 43. Backtesting | Testing a model on historical data to evaluate performance. | Helps catch models that look good in theory but fail in reality. | Model validation, comparison, QA. |
| 44. Overfitting | When a model performs well on training data but poorly on new data because it learned noise, not signal. | Overfit models waste spend and degrade performance when scaled. | Model QA, feature selection, validation planning. |
| 45. Feature Engineering | Creating or transforming input variables to improve model performance. | Often determines whether models are robust or brittle. | Modeling workflows, analytics pipelines. |
| 46. Behavioral Economics | A field that combines psychology and economics to explain why real-world decisions often deviate from “perfectly rational” behavior. | Helps teams understand why people procrastinate, stick with defaults, fear losses, or avoid uncertainty—critical for conversion strategy. | Framing offers (loss vs gain), building nudges into messaging, designing tests, and interpreting why audiences respond unexpectedly. |
| 47. Model Explainability | The ability to understand and communicate why a model produced a given score or outcome (at a level appropriate for the user). | Explainability helps build trust and supports governance—especially when model influences spend, eligibility, or customer experience. | Model documentation, stakeholder alignment, debugging, and compliance reviews. |
| 48. Values-Based Segmentation | Segmenting audiences by core values and priorities (e.g., status, stability, altruism, independence) rather than only demographics or behaviors. | Values often predict brand affinity and message resonance better than broad demographics. | Audience strategy, creative themes by segment, personalization, and long-term loyalty positioning. |
| 49. Likert Scale | A standardized rating scale (often 1–5 or 1–7) used to quantify attitudes, agreement, likelihood, or intensity of a trait. | It transforms “soft” concepts like motivation into structured signals that can be segmented, modeled, and activated. | Building motivation-based attributes, creating propensity-like segment thresholds, prioritization and ROI management, and generating consistent inputs for modeling/analytics. |
| 50. Psychometric Modeling | Modeling approaches used to quantify psychological characteristics (attitudes, motivations, traits) from observed data or structured measures. | It enables scalable, consistent measurement of “why” factors that are otherwise hard to operationalize. | Creating motivation/decision-driver variables, segmentation, personalization frameworks, and predictive modeling features. |
AI in Marketing
| Term | Definition | Why It Matters | How It’s Used |
|---|---|---|---|
| 51. AI in Marketing | The use of machine learning and other AI techniques to improve marketing decisions—such as who to target, what to say, where to spend, and how to measure impact. | AI can turn complex, messy data into practical decisions at scale—especially when audiences, channels, and content options explode. | Predictive audiences, personalization, media optimization, experimentation, automated insights, and forecasting. |
| 52. Machine Learning (ML) | A type of AI where algorithms learn patterns from data to make predictions or decisions without being explicitly programmed for every rule. | ML powers many “predictive” capabilities marketers rely on—propensity, churn risk, next-best-action, and more. | Predictive modeling, scoring, segmentation, anomaly detection, and optimization. |
| Common confusion (AI vs ML): AI is the umbrella term; ML is one common approach inside AI. |
|||
| 53. Predictive AI | I/ML used to forecast future outcomes like likelihood to convert, churn, respond, or purchase a category. | Predictive AI helps you prioritize spend and focus on the audiences most likely to drive outcomes. | Propensity scoring, audience ranking, suppression of low-likelihood segments, and scenario planning. |
| 54. Generative AI (GenAI) | AI that creates new content such as text, images, audio, video, or code based on patterns learned from large datasets. | GenAI can accelerate content creation and iteration, but its value depends on the quality of the inputs (brand rules, audience insights, claims) and validation. | Drafting ad copy and emails, generating creative variants, summarizing research, producing outlines, and creating on-brand content at scale. |
| 55. AI-Ready Data | Data that’s structured, governed, and documented well enough to be used reliably in AI/ML systems (clear definitions, consistent formats, usable IDs, and known limitations). | AI doesn’t “fix” messy data—poor inputs often produce unreliable outputs, at scale. | Preparing customer/prospect data for modeling, scoring, segmentation, and measurement; requires human input. |
| 56. Training Data | The historical data used to teach a model patterns and relationships. | If training data is incomplete, biased, or outdated, model outputs will reflect those constraints. | Building propensity models, classifiers, and recommendation systems. |
| 57. Feature (in Machine Learning) | An input variable a model uses—such as a demographic attribute, behavior signal, purchase indicator, or motivation measure. | Features are where “data strategy” becomes “model performance.” Better features often matter more than fancier algorithms. | Feature engineering, model training, and ongoing improvement. |
| 58. AI Hallucination (GenAI) | When a generative AI system produces information that sounds plausible but is incorrect or unsupported. | In marketing, hallucinations can create brand risk—incorrect claims, wrong stats, or inaccurate product details. | As a risk concept that drives content review workflows, citations, and human QA for AI-assisted writing. |
| 59. Synthetic Data | Artificially generated data designed to resemble real data patterns, often used to test systems or protect privacy. | It can help teams experiment and develop workflows when real data access is limited—but it’s not automatically “representative.” | Testing pipelines, privacy-preserving prototyping, model experimentation, and QA. |
| Common confusion (AI vs ML): AI is the umbrella term; ML is one common approach inside AI. |
|||
| 60. AI Optimization | Using AI systems to automatically adjust decisions (bids, budgets, audience allocation, creative rotation) toward a goal. | Optimization can improve efficiency, but it requires clean inputs, clear goals, and guardrails to avoid chasing the wrong metric. | Media buying platforms, experimentation frameworks, creative selection, pacing, and conversion optimization. |
| 61. AIEO (AI Engine Optimization) | The practice of structuring content so it’s more likely to be surfaced or summarized accurately by AI-driven discovery experiences (AI search, assistants, summaries), not just classic SEO rankings. Encompasses GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization). | As discovery shifts toward AI summaries, content needs clear structure, strong definitions, and trustworthy sourcing to earn visibility and accurate representation. | Creating “definition-first” intros, structured headings, FAQ blocks, clean internal linking, and citation-friendly content formatting. |
| Common confusion (AIEO vs SEO): SEO focuses on ranking in search results; AIEO focuses on being understood and selected by AI-driven answer systems. They overlap heavily in best practices. |
|||
| 62. RAG (Retrieval-Augmented Generation) | A GenAI approach that retrieves relevant information from approved sources and uses it to generate a response grounded in that material | RAG reduces hallucinations and helps keep AI outputs aligned to current, approved facts—useful for brand-safe content and support experiences. | Brand-safe content drafting, sales enablement assistants, internal knowledge bots, and customer support tools. |
Activation + Measurement
| Term | Definition | Why It Matters | How It’s Used |
|---|---|---|---|
| 63. Custom Audiences | A platform-ready audience built from your identifiers (e.g., CRM list) or defined segment criteria. | Powers high-intent targeting and suppression. | Paid social, CTV, display, email matching. |
| 64. Lookalike Audience | An audience of new prospects who resemble a seed audience (e.g., your best customers). | Expands acquisition while preserving some similarity to converters. | Prospecting, scaling campaigns. |
| 65. Suppression | Excluding people from targeting (e.g., recent buyers, current customers, ineligible audiences). | Reduces wasted spend and improves customer experience. | Paid media exclusions, direct mail suppression, channel coordination. |
| 66. Frequency | How often someone sees your ad in a given time period. | Too low = no impact; too high = waste and annoyance. | Media planning, cross-channel coordination, optimization. |
| 67. Frequency Capping | A control that limits how many times the same viewer or household sees an ad. | Prevents ad fatigue and improves campaign efficiency. | Manages exposure across digital and CTV campaigns. |
| 68. Reach | The number of unique people exposed to your campaign. | Reach is foundational for awareness goals and for understanding duplication across platforms. | Planning, reporting, incremental reach analysis. |
| 69. Incremental Reach | Measuring the additional unique audience reached by adding a new channel to a campaign, often measured when platforms like CTV reach viewers not exposed to linear TV ads. | Helps marketers understand the true value of expanding media channels. | Used to evaluate how channels like CTV extend reach beyond linear TV. |
| 70. Attribution | Methods used to assign credit for conversions to marketing touchpoints. | The attribution model you choose can dramatically change perceived ROI. | Reporting, optimization, budget allocation. |
| 71. Marketing Mix Modeling (MMM) | A statistical approach that estimates how different marketing activities contribute to outcomes over time. | Useful for strategic budget allocation and understanding channel contribution at a macro level. | Quarterly planning, scenario modeling, investment decisions. |
| 72. Multi-Touch Attribution (MTA) | Attribution models that attempt to assign fractional credit across multiple touches leading to conversion. | Helpful for understanding customer journeys, but quality depends on data completeness and assumptions. | Channel optimization, journey analysis. |
Advanced TV
| Term | Definition | Why It Matters | How It’s Used |
|---|---|---|---|
| 73. Advanced TV | The umbrella term for data-enabled TV advertising across streaming, CTV, addressable, and programmatic environments. | Brings targeting and measurement to television, enabling advertisers to reach households across devices. | Planning and executing audience-based campaigns beyond traditional linear buying. |
| 74. Addressable TV | TV advertising that uses data to deliver an ad to a specific household on a TV screen. | Technology allows different households watching the same program to see different ads. | Audience targeting, segmentation, and tailored messaging with each advertisement. |
| 75. Connected TV (CTV) | Television devices connected to the internet that stream video content and allow advertisers to deliver targeted ads within streaming environments (e.g., Smart TV, Roku, Apple TV). | Combines the scale and impact of TV with programmatic targeting, measurement, and optimization. | Audience targeting and measurement through streaming apps and platforms. |
| 76. CTV Targeting | Audience targeting for connected TV advertising using available identifiers and segment definitions. | Offers digital-style targeting to premium video environments. | Prospecting, sequential messaging, household-focused campaigns. |
| 77. Data-Driven Linear TV | Traditional television that leverages data for planning of a media buy. More of an offline planning activity that leverages data from varying channels. | Improves planning and optimizes advertising within traditional broadcast or cable TV, reducing waste in linear campaigns. | Understanding what percentage of viewers of a specific channel or program are also likely to be buyers of a specific brand or product. Used to select networks, dayparts, and programs that best align with audience insights. |
| 78. Linear TV | Traditional television viewing through broadcast, cable, or satellite TV with a set schedule. | Delivers large-scale reach for brand campaigns mass awareness and broad audience exposure. | Audience targeting through scheduled programming, with planning informed by audience data and campaign measurement. |
| 79. FAST (Free Ad-Supported TV) | Linear, ad-supported content delivered via streaming (e.g., Pluto TV, Tubi). | Expands premium inventory in streaming environments to reach cord-cutters and streaming-first audiences at scale. | Targeted advertising that reaches streaming viewers through ad-supported channels that run continuous, linear-style programming. |
| 80. Over-the-Top-TV (OTT) | Video content delivered via the internet, bypassing traditional cable/satellite. Distribution model for services like streaming apps and connected TV platforms (e.g., Hulu, Netflix). | Creates new advertising inventory within streaming platforms where audiences now spend a growing share of viewing time. | Extends audience targeting and measurement into streaming environments and devices. |
| 81. Programmatic TV | The automated buying and selling of TV advertising using software and data signals, allowing media to be transacted and optimized in real time across streaming and digital video environments. | Increases efficiency, targeting precision, and campaign optimization | Used to transact CTV and digital video inventory through demand-side platforms. |
| 82. Smart TV | A television with built-in internet connectivity and streaming apps that allows viewers to access digital content directly without external devices. No sticks or dongles are required (e.g., LG, Samsung, Vizio). | Primary gateway to streaming content, expanding opportunities for advertisers to reach audiences where CTV advertising is delivered. | Targeted video advertising through streaming apps and platforms accessed directly on smart TVs, often using audience data and programmatic buying. |
| 83. MVPD (Multichannel Video Programming Distributor) | Traditional cable or satellite providers. Refers to any service provider that delivers video programming services (e.g., Comcast, DirecTV). | Represents the traditional cable and satellite TV ecosystem that still delivers large-scale reach and premium programming inventory for advertisers. | Audience targeting and measurement through traditional TV networks. |
| 84. vMVPD (Virtual Multichannel Video Programming Distributor) | Internet-based, streaming pay-TV subscription services that offer live, traditional TV channels (e.g., Sling TV, Hulu + Live TV). | Provides access to traditional linear TV networks and on-demand content delivered over the internet without the traditional set-top box infrastructure. | Reach audiences who watch live television through streaming services instead of cable at scale. |
| 85. AVOD (Ad-Supported Video on Demand) | Free content supported by ads. Streaming video content is available for free to viewers in exchange for watching ads (e.g., YouTube, Tubi) | Provides scalable advertising inventory in streaming environments | Target and activate audiences through ad-supported platforms. |
| 86. SVOD (Subscription Video on Demand) | Ad-free, subscription-based content. Streaming services where viewers pay a recurring fee to access content without traditional ads (e.g., Netflix, Disney+). | Creates premium streaming environments where high-quality content attracts large, highly engaged audiences. | Audience targeting and measurement within subscription-based streaming platforms. |
| 87. TVOD (Transactional Video on Demand) | A streaming model where viewers pay for individual purchases or rentals rather than subscribing to a service (e.g., iTunes). | Provides consumers with flexible access to premium content without requiring a subscription, expanding how audiences access digital video. | Advertising opportunities are generally minimal, though brands may participate in promotional partnerships tied to movie releases or premium content. |
| 88. PVOD (Premium Video on Demand) | streaming model where viewers pay a premium price to access newly released or early-window content at home, often shortly after or alongside theatrical release. | Expands digital distribution for premium film releases and gives advertisers new ways to reach audiences outside traditional theatrical windows. | Targeted advertising that reaches viewers who are willing to pay a premium to watch them at home before they are widely available on other platforms. |
| 89. TVE (TV Everywhere): | Authentication services that allow cable or satellite subscribers to watch network content across digital devices | Extends traditional TV access to streaming environments. Used by networks to provide mobile and streaming viewing for existing subscribers. | Advertisers reach authenticated viewers through ads delivered within network apps and digital platforms that stream TV programming. |
Data Quality + Governance
| Term | Definition | Why It Matters | How It’s Used |
|---|---|---|---|
| 90. Data Quality | Data quality is how well a dataset is fit to support a specific marketing use case—based on factors like accuracy, completeness/coverage, freshness, consistency, and clear documentation of how the data was created and can be used. | Even “more” data can reduce performance if it’s noisy, stale, or inconsistent. High-quality data improves targeting precision, reduces wasted spend (and bad customer experiences), and makes measurement and modeling more trustworthy. | As a vendor and dataset evaluation lens (scorecards/rubrics), to set refresh and governance standards, to QA audiences before activation, and to validate impact through testing (e.g., lift or holdout-based incrementality). |
| 91. Data Coverage | How much of your target population can be represented or matched in a dataset. | Low coverage limits scale and can skew performance. | Vendor evaluation, planning, match-rate analysis |
| 92. Data Accuracy | How often an attribute correctly reflects reality. | Inaccurate data creates waste and weakens trust in analytics. | Validation, QA, vendor selection. |
| 93. Data Freshness | How recently data was updated and how quickly it reflects real-world changes. | Stale data can miss life changes, intent shifts, and eligibility changes. | Campaign planning, vendor evaluation, modeling inputs. |
| 94. Latency | The delay between a real-world event and when it appears in a dataset or system. | High latency reduces relevance for time-sensitive use cases. | Data pipelines, activation readiness, measurement. |
| 95. Compliance (Privacy + Governance) | The policies, legal requirements, and internal rules that govern how data can be collected, used, and shared. | Compliance reduces risk and builds long-term trust with customers and partners. | Vendor assessment, activation restrictions, documentation, consent management. |
FAQ
What’s the difference between marketing data and audience data?
Marketing data is the broad set of inputs (customer, prospect, behavioral, purchase, etc.). Audience data is that information packaged into segments you can activate in platforms.
Is deterministic data always better than predictive data?
Not always—deterministic often wins on precision, predictive can win on propensity and scale. Many strategies use both.
What’s the fastest way to tell if a dataset is “good”?
Start with a simple rubric: fit for use case, match methodology, attribute definitions, freshness, and validation approach. Consider the data quality scorecard.
What terms do teams most commonly confuse?
Deterministic vs probabilistic, enrichment vs modeling, identity graph vs clean room, lift vs incrementality, cognitive science vs consumer psychology.
Should I build audiences the same way for CTV as I do for digital?
The strategy can be similar, but the identifiers, household dynamics, and measurement approaches often differ.
How should marketers work with data science teams on modeling?
Align first on the outcome metric, the evaluation method (holdout/lift), and the operational constraints (where the audience must activate).
Next Steps
If you’re using this glossary as a starting point, here are two practical follow-ups:
- Evaluate your current data with a scorecard (see above)
- Go deeper on people-based data and modern audience targeting
Ready to take your marketing data expertise to the next level? Schedule your free data consultation or secure your free data test today.
