CDP Glossary: Turn MarTech Terms Into Systematic Action

Most CDP glossaries read like dictionary entries written by engineers for other engineers. You get technical definitions but no path forward. You learn what "identity resolution" means but not why it matters to your business or how to actually implement it.

Here's what I've noticed after years of building data systems: the companies that succeed with CDPs don't just understand the terms—they see how each concept connects to create a working system. They know that "data unification" isn't just cleaning up duplicate records. It's the foundation that makes personalization, attribution, and automation actually possible.

This glossary is different. Each term includes the "why it matters" and "what to do about it" that other resources skip. Because knowing definitions doesn't transform your business. Building systems does.

The Foundation Terms You Need First

Customer Data Platform (CDP)

A CDP is software that collects customer information from different places (your website, email system, store purchases, support tickets) and combines it into one unified view per person.

Why it matters: Right now, your marketing team sees one version of a customer, your sales team sees another, and your support team sees a third. A CDP connects these fragments so everyone works from the same truth.

Action step: Before shopping for CDPs, map where your customer data currently lives. List every system that touches customer information. This inventory reveals what needs connecting.

First-Party Data

Information you collect directly from your customers through your own channels—website visits, purchase history, email engagement, app usage, survey responses.

Why it matters: As third-party cookies disappear and privacy regulations expand, first-party data is the only reliable foundation for personalization and targeting. You own it, you control it, and customers gave it to you directly.

Action step: Audit what first-party data you're already collecting but not using. Most businesses capture valuable signals they never activate because the data sits isolated in separate tools.

Identity Resolution

The process of connecting different pieces of information about the same person across multiple devices and platforms to create one unified profile.

Why it matters: Your customer browses on mobile, adds items on desktop, and purchases in-store. Without identity resolution, these look like three different people. You can't personalize effectively when you're treating one customer as three strangers.

Action step: Start with email as your primary identifier. When someone provides an email address, use it to link their behavior across sessions and devices. This simple approach solves 70% of identity challenges before you need complex solutions.

Data Collection and Management

Data Ingestion

The process of bringing data from source systems (your website analytics, email platform, CRM, etc.) into your CDP.

Why it matters: If data doesn't flow automatically into your CDP, it becomes outdated immediately. Manual imports create gaps where customer actions happen but aren't recorded.

Action step: Prioritize real-time ingestion for behavioral data (website clicks, email opens) and daily batch imports for transactional data (purchases, support tickets). Not everything needs to be instant, but user behavior should update quickly.

Data Schema

The structure that defines what information you collect and how it's organized—like the blueprint for your customer data house.

Why it matters: A clear schema prevents chaos. Without it, different teams label the same thing differently ("customer ID" vs "user ID" vs "contact ID"), making it impossible to connect information accurately.

Action step: Create a simple data dictionary that defines your core customer attributes. Include what each field means, what format it uses, and which system owns the truth for that information.

ETL (Extract, Transform, Load)

The process of pulling data from source systems (Extract), cleaning and formatting it to match your schema (Transform), and moving it into your CDP (Load).

Why it matters: Raw data from different systems rarely matches. One tool formats dates as "MM/DD/YYYY" while another uses "DD-MM-YYYY." ETL makes everything speak the same language so your CDP can actually unify information.

Action step: Focus your transformation efforts on standardizing identifiers first (email addresses, phone numbers, customer IDs), then tackle attribute formatting. Clean identifiers unlock identity resolution; clean attributes improve segmentation.

Data Governance

The policies and processes that control who can access customer data, how it's used, and how you ensure privacy and compliance.

Why it matters: One privacy violation can cost you customer trust, regulatory fines, and brand reputation. Plus, clear governance prevents teams from accidentally using data they shouldn't or making decisions based on unreliable information.

Action step: Document three simple rules before you activate any CDP data: (1) Who owns customer consent management? (2) Which data requires customer permission before use? (3) How long do you keep different types of information?

Unification and Activation

Data Unification

Combining information from multiple sources into single, accurate customer profiles by matching records, removing duplicates, and resolving conflicts.

Why it matters: This is the core job of a CDP. Without unification, you just have a expensive data warehouse. With it, you transform fragments into insight and insight into action.

Action step: Start with unifying your three highest-volume data sources first. For most businesses, that's website behavior, email engagement, and transaction history. Get these three talking before adding complexity.

Golden Record

The single, most complete and accurate version of a customer profile after all data sources have been unified and conflicts resolved.

Why it matters: When your email system says a customer's name is "John Smith" but your CRM says "J. Smith" and your store system says "Jonathan Smith," which is correct? The golden record decides by applying rules you define.

Action step: Establish a source hierarchy for each attribute type. For contact information, your CRM might be the source of truth. For purchase behavior, your transaction system wins. Define this before conflicts arise.

Segmentation

Grouping customers based on shared characteristics, behaviors, or predicted value so you can target them with relevant messages and offers.

Why it matters: Sending everyone the same message is easy but ineffective. Segmentation lets you speak differently to new customers vs loyal advocates, high-value prospects vs bargain hunters, engaged users vs those at risk of leaving.

Action step: Build your first three segments around behavior, not just demographics. Try: (1) Active customers (purchased in last 90 days), (2) At-risk customers (no purchase in 90+ days but engaged with emails), (3) Dormant customers (no interaction in 180+ days). These segments drive clear action.

Data Activation

Pushing unified customer profiles from your CDP to the tools where you take action—your email platform, ad systems, personalization engines, analytics tools.

Why it matters: A unified customer profile has zero value if it sits in your CDP unused. Activation means your email system can personalize based on website behavior, your ads can exclude recent purchasers, your website can show relevant content based on purchase history.

Action step: Choose one activation use case to prove value fast. Suppressing recent purchasers from acquisition ads usually delivers immediate ROI and requires minimal setup—just sync purchase data to your ad platforms daily.

Advanced Capabilities

Real-Time Personalization

Changing website content, offers, or messages instantly based on who the customer is and what they're doing right now.

Why it matters: When a customer who just browsed premium products visits your homepage, showing them entry-level offerings wastes the moment. Real-time personalization matches experience to intent while interest is highest.

Action step: Start with one high-traffic page and two audience segments. For example, show different homepage heroes to first-time visitors vs returning customers. Measure the difference before expanding to more complex personalization.

Predictive Analytics

Using historical data patterns to forecast future customer behavior—who's likely to purchase, who might leave, which products someone will want next.

Why it matters: Reactive marketing responds to what customers already did. Predictive marketing anticipates what they'll do next and positions your message at the perfect moment.

Action step: Begin with churn prediction for your top customer segment. Identify patterns shared by customers who left, then flag current customers showing those same signals. Reach them with retention offers before they decide to leave.

Lookalike Audiences

Finding new prospects who share characteristics with your best existing customers, then targeting them with ads or outreach.

Why it matters: Acquiring random new customers is expensive. Acquiring people who look like your best customers has higher conversion rates and better long-term value because they match proven patterns.

Action step: Export your top 20% of customers by lifetime value to your ad platforms and create lookalike audiences. Run split tests comparing these audiences to your current targeting to quantify the improvement.

Event Streaming

Capturing and processing customer actions (clicks, purchases, form submissions) immediately as they happen instead of in batches.

Why it matters: Batch processing means you learn a customer abandoned their cart at midnight when you import data the next morning. Event streaming knows instantly and triggers a recovery message within minutes while intent is fresh.

Action step: Prioritize streaming for high-intent events: cart abandonment, product views, pricing page visits, free trial signups. These moments have short windows where immediate response dramatically improves conversion.

Architecture and Integration

Composable CDP

Building your customer data platform by connecting specialized tools (a data warehouse, identity resolution service, activation layer) instead of using one vendor's complete solution.

Why it matters: Packaged CDPs lock you into one vendor's roadmap and pricing. Composable approaches let you swap components as better options emerge and avoid vendor lock-in.

Action step: Evaluate whether you have the technical resources to manage multiple integrations. Composable CDPs offer flexibility but require more internal expertise. Packaged solutions cost more but include support.

API (Application Programming Interface)

The method that lets different software systems talk to each other and exchange information automatically.

Why it matters: Without APIs, you're manually exporting data from one system and importing it to another—slow, error-prone, and impossible to scale. APIs create the automatic data flow that makes CDPs work.

Action step: When evaluating any MarTech tool, ask about API capabilities first: What data can it send? What can it receive? How often does it sync? These answers determine integration possibilities.

Reverse ETL

Pushing data from your CDP or data warehouse back to operational tools (CRM, email, ads) to activate insights where teams work.

Why it matters: Your data team builds great customer scores and segments in your warehouse, but sales reps work in the CRM. Reverse ETL puts those insights directly into tools people actually use daily.

Action step: Identify one valuable insight trapped in your warehouse or analytics tool. Push it to one operational system where it changes decisions. For example, sync lead scores to your CRM so sales prioritizes high-value prospects.

Data Warehouse

Centralized storage that holds large amounts of structured customer and business data optimized for analysis and reporting.

Why it matters: CDPs excel at real-time activation but data warehouses excel at historical analysis and complex reporting. Many businesses need both—CDPs for doing, warehouses for understanding.

Action step: Use your CDP for operational segments and activation. Use your warehouse for deep analysis, attribution modeling, and reporting across long time periods. Each tool does different jobs well.

Privacy and Compliance

Consent Management

Systems and processes for collecting, storing, and honoring customer permissions about how you use their data.

Why it matters: Privacy regulations (GDPR, CCPA, and others) require explicit consent for many data uses. Plus, respecting customer preferences builds trust that drives better long-term relationships than aggressive data collection.

Action step: Implement a consent management platform that integrates with your CDP before you activate any data. It's easier to build right from the start than to retrofit compliance later.

PII (Personally Identifiable Information)

Data that can identify a specific person—names, email addresses, phone numbers, home addresses, government ID numbers.

Why it matters: PII requires stricter security and privacy controls than anonymous behavioral data. Regulations impose heavy penalties for PII breaches, and customers expect careful handling of their personal information.

Action step: Separate PII from behavioral data when possible. Store names and contact information in secure systems with limited access. Use anonymous identifiers for analysis and testing.

Data Retention Policy

Rules defining how long you keep different types of customer data before deleting it.

Why it matters: Keeping data forever increases security risk, storage costs, and privacy compliance complexity. Most data loses value over time—five-year-old browsing behavior doesn't predict current intent.

Action step: Set retention periods by data type and use case. Keep transactional data for accounting requirements (often 7 years). Delete behavioral data after 24 months unless customers are active. Remove contact information for people who haven't engaged in 36 months.

Right to Deletion

Customer rights (required by regulations like GDPR and CCPA) to request that you delete all personal information you've collected about them.

Why it matters: You need systems that can find and remove all instances of a customer's data across every tool in your stack—not just your CDP but every connected system. Manual processes don't scale and miss data.

Action step: Map your data flow to understand where customer information lives. Build automated deletion workflows that cascade from your CDP through all connected systems when a customer requests removal.

Measurement and Optimization

Attribution

Determining which marketing touchpoints and channels deserve credit for driving customer actions like purchases, signups, or conversions.

Why it matters: Without attribution, you can't answer "which marketing works?" You might cut effective channels because they don't get last-click credit, or over-invest in channels that look good but don't actually drive results.

Action step: Start with multi-touch attribution that gives partial credit to each interaction in the customer journey. It's not perfect, but it's dramatically better than last-click attribution that ignores everything except the final touchpoint.

Customer Lifetime Value (CLV)

The total revenue you can expect from a customer over their entire relationship with your business.

Why it matters: Acquisition cost only makes sense in context of lifetime value. Spending $200 to acquire a customer is great if they're worth $2,000, terrible if they're worth $150. CLV focuses decisions on long-term value, not just initial conversion.

Action step: Calculate simple CLV for your top customer segments: (average purchase value) × (purchase frequency per year) × (average customer lifespan in years). This basic model is good enough to guide acquisition spending and retention priorities.

Incrementality Testing

Measuring whether your marketing actually causes behavior change or just reaches people who would've converted anyway.

Why it matters: Your "high-performing" email campaign might target people already planning to purchase. It looks effective but doesn't actually increase sales. Incrementality testing separates real impact from correlation.

Action step: Run holdout tests where you don't contact a random sample of your target audience, then compare their behavior to the group that received your campaign. The difference is your true incremental impact.

How to Actually Use This Glossary

Understanding terms doesn't build systems. Here's how to move from definitions to implementation:

If you're starting from zero: Focus on "Data Ingestion," "Identity Resolution," and "Data Unification" first. These foundation capabilities must work before anything else matters. Map your current data sources, choose identifiers, and define your unification rules.

If you have a CDP but aren't seeing results: Review "Data Activation" and "Segmentation." Most CDP failures happen here—the platform collects and unifies data fine, but nobody activates it effectively. Build three behavioral segments and create one activation use case that delivers measurable business value.

If you're optimizing an existing system: Study "Predictive Analytics," "Attribution," and "Incrementality Testing." You've mastered the basics; now layer in advanced capabilities that compound your advantage.

If you're evaluating vendors: Use the architecture terms ("Composable CDP," "API," "Reverse ETL") to ask detailed questions that reveal real capabilities versus marketing promises. Vendors who can't clearly explain these concepts probably haven't built robust systems.

The Pattern Most People Miss

Every term in this glossary connects to a larger system. Companies that struggle with CDPs treat each capability as separate:

• They implement identity resolution without thinking about data governance
• They build segments without planning activation
• They collect data without defining retention policies
• They activate insights without measuring incrementality

Companies that succeed see the whole picture. Identity resolution enables accurate segmentation. Segmentation drives personalized activation. Activation requires consent management. Measurement validates incrementality. Each piece reinforces the others.

This systematic thinking—seeing how pieces connect to create outcomes—is what separates businesses that extract value from their customer data platform and those that just pay the subscription.

Next Steps: From Terms to Systems

You now understand the language of customer data platforms. The question is: what system will you build with this vocabulary?

Start with your biggest gap. If you can't connect customer behavior across channels, focus on identity resolution and data unification. If data sits unused in your CDP, prioritize activation and segmentation. If you're activating data but can't prove impact, implement attribution and incrementality testing.

House of MarTech helps businesses build systematic customer data platforms that unify insights and drive measurable growth. We don't just implement tools—we design data systems where each component works together to solve your specific business challenges.

Whether you're evaluating CDPs, optimizing an existing platform, or building a composable architecture, we bring strategic frameworks that turn technical complexity into clear action.

The businesses winning with customer data aren't the ones with the biggest budgets or the most advanced tools. They're the ones who see how each piece fits together and build systems that compound value over time.

Ready to turn these terms into a working system? Let's map your customer data architecture and identify the highest-impact opportunities for unification and activation. Connect with House of MarTech to start building your systematic approach to customer data.