Customer data platform (CDP)

A customer data platform (CDP) is software that collects customer data from every system a business uses, links each piece to the right person, stores it as a lasting profile and shares that profile with other systems. The CDP Institute defines a CDP as “software that creates and maintains a persistent, unified customer record that is accessible to other systems.”

What is a customer data platform?

That is the 2026 definition. The original 2013 version from David Raab, who named the category, included the word “packaged” (CDP Institute, CDP Basics, accessed October 2026). The new version drops it: CDP functions can be built into the CDP or assembled from outside systems such as a data warehouse.

The definition has two parts (CDP Institute, What Is a CDP, accessed October 2026):

• A persistent, unified profile: the CDP links information about the same customer from many systems and keeps enough detail to track behavior over time. The core data is first-party data about identified people.
• Accessible to other systems: the CDP adds calculated values such as model scores, then shares the results through APIs, database queries or file extracts.

This page is a short definition. For how to choose and roll out a CDP, read our customer data platform guide; for the Vietnamese market, see what experts say about CDPs in Vietnam.

Types of CDP

The CDP Institute sorts vendors into 4 groups, each including the functions of the one before:

1. Data CDPs: collect data, resolve identity, store profiles. This is the minimum to count as a CDP.
2. Analytics CDPs: add customer segmentation and sometimes predictive models.
3. Campaign CDPs: choose a different message or product recommendation for each person within a segment.
4. Delivery CDPs: send the messages themselves through email, website, app or ads.

CDP vs CRM vs DMP

In short, a CRM is where the sales team works, a DMP serves ad buying, and a CDP is the data layer that feeds customer profiles to analytics, marketing and service tools.

Example (hypothetical)

A retail chain has 40,000 website accounts, 25,000 app accounts and 60,000 loyalty phone numbers from its stores, 125,000 records in total. The CDP matches them by phone number and email and finds 80,000 real customers, so 45,000 records (36%) are duplicates. Without matching, a customer who buys both online and in store might get the same offer 2 or 3 times.

From those 80,000 profiles, the marketing team selects “bought 2 or more times in the past 12 months” and sends the segment to Google Ads Customer Match. Customer Match memberships expire after 540 days, and a list stays eligible only if at least 100 members are added or updated within that window (Google Ads Help, About Customer Match, accessed October 2026). The CDP handles that syncing automatically.

Legal note

A CDP holds names, phone numbers and purchase histories, which count as personal data under Vietnam’s Personal Data Protection Law No. 91/2025/QH15, in effect since January 1, 2026. When a customer withdraws consent, that status has to reach the CRM, the call center, Zalo messages and ad audiences. See our guide to the Vietnam personal data protection law for marketers.

Frequently asked questions

If I already have a CRM, do I need a CDP?

Not necessarily. If customer data sits neatly in the CRM and one or two other tools, the CRM is enough. A CDP is worth considering when customers leave traces across your website, app, stores, marketplaces and Zalo, and no system can merge them into one profile.

What is a composable CDP?

It is a CDP that works directly on customer data in the company’s data warehouse instead of keeping its own copy the way a packaged CDP does. Hybrid setups do both.

If you want to connect data from ads, your website and your CRM so you can measure each customer accurately, contact Adtek.