CDP
Also: CDP, Customer Data Platform, Plateforme de donnees clients, Customer data hub
A Customer Data Platform unifies customer data from all sources into persistent, actionable profiles that other systems can use.
What it is
A Customer Data Platform (CDP) is a system that collects customer data from every source a business has, resolves it into a single persistent profile per person, and makes those profiles available to other tools. Unlike a data warehouse (built for analysts) or a CRM (built for sales and manual entry), a CDP is packaged software designed to be owned by business teams and to feed data back out for activation.
The defining trait is identity resolution: matching records that belong to the same person across channels (web, mobile, email, point of sale, support) into one unified, continuously updated profile.
Why it matters
- One version of the customer. Fragmented data across silos leads to inconsistent messaging and bad decisions.
- Actionable, not just stored. A CDP is built to push audiences and traits into downstream tools (ad platforms, email, personalization engines).
- Governance and consent. Centralizing data makes it easier to enforce consent, retention, and privacy rules (GDPR, CCPA).
- Speed for business teams. Marketers can build segments without waiting on engineering.
How it is used in practice
A typical CDP workflow:
- Ingest: pull events and attributes from web SDKs, apps, transactional databases, and SaaS tools.
- Unify: apply identity resolution rules (deterministic on email or user ID, sometimes probabilistic).
- Model: compute traits and audiences (for example, "high value at churn risk").
- Activate: sync segments to destinations for targeting, suppression, or personalization.
- Analyze: measure results and feed them back into the profiles.
Two common patterns exist: packaged CDPs with their own storage, and composable CDPs that run on top of an existing warehouse.
Worked example
A retailer has a website, a loyalty app, and an email tool. Anna browses running shoes on the site (anonymous cookie), later logs into the app with her email, then buys in store using her loyalty card.
- Without a CDP: three disconnected records, and Anna gets an ad for the shoes she already bought.
- With a CDP: the cookie, app login, and loyalty ID resolve to one profile. The CDP computes a trait (`recent_purchase = running_shoes`), builds a suppression audience, and syncs it to the ad platform so Anna is excluded, while adding her to a cross sell segment for socks and apparel.
The result is coherent, consent aware, and measurable engagement across every channel.
See also
Frequently asked questions
What does a CDP actually do?
A Customer Data Platform collects customer data from every source a company has (web, mobile, email, point of sale, support), resolves those records into one persistent profile per person, and pushes those profiles back out to other tools. The defining capability is identity resolution: matching scattered records that belong to the same individual into a single, continuously updated profile.
What is the difference between a CDP, a CRM and a data warehouse?
A CRM is built for sales teams and relies largely on manual entry; a data warehouse is built for analysts and stores data for querying. A CDP is packaged software owned by business teams whose purpose is activation: it unifies data automatically and sends audiences back into ad platforms, email tools and personalization engines. The distinction is less about storage than about who uses the system and whether data flows back out.
Who needs a CDP inside a company?
Marketing and data leadership are the primary users: a CDP lets marketers build segments without waiting on engineering, and gives data teams a single place to enforce consent and retention rules. It matters most when customer data is fragmented across silos, which produces inconsistent messaging and unreliable decisions.
What is the difference between a packaged CDP and a composable CDP?
A packaged CDP comes with its own storage layer and handles ingestion, unification and activation in one product. A composable CDP runs on top of an existing data warehouse, using it as the source of truth rather than duplicating the data. The choice usually depends on whether the company already has a warehouse and the data team to run it.
What does a CDP change in a concrete retail scenario?
Take a shopper who browses running shoes anonymously on a website, later logs into the loyalty app with her email, then buys in store with her loyalty card. Without a CDP these are three disconnected records and she keeps seeing ads for shoes she already owns. With a CDP, the cookie, the app login and the loyalty ID resolve into one profile: a trait like recent_purchase = running_shoes is computed, she enters a suppression audience synced to the ad platform, and she is added to a cross sell segment for socks and apparel.