Customer data platform (CDP) and Data management platform (DMP) at a glance
A CDP is a store built for the marketing use of customer data. It ingests events and records from the website, app, transactions, email and support systems, resolves them into one profile per person keyed by email, customer ID or login, and lets marketers build segments that it pushes to channels through connectors. The martech categories guide places it in the data layer alongside the warehouse and consent management, and explains how it overlaps with marketing automation.
A DMP was built for advertising audiences. It collects browsing behaviour keyed to cookie or device identifiers, from the advertiser's own tags, publisher partners and third-party data providers, groups it into segments such as in-market or interest audiences, and syncs those segments to demand-side and supply-side platforms through cookie matching. The programmatic ecosystem guide sets out where it sits in the advertising supply chain and how cookie matching works.
The difference is not just technical. A CDP answers the question of who the business's customers are and what they did. A DMP answered the question of which anonymous browsers looked like buyers. The first question has grown more answerable as businesses collect consented first-party data; the second has grown less answerable as the identifiers it relied on have disappeared.
Compare table
Customer data platform (CDP) vs Data management platform (DMP)
| Row label | Customer data platform (CDP) | Data management platform (DMP) |
|---|---|---|
| Data type | Events, transactions and attributes from the business's own systems | Browsing behaviour and segment membership, often from third parties |
| Identity | Identified people: email, customer ID, login, phone | Anonymous browsers and devices: cookie and device identifiers |
| Retention | As long as the customer relationship and the retention policy allow | As long as the cookie or identifier survives, often weeks |
| Sources | CRM, website, app, point of sale, email, support | Advertiser tags, publisher partners, third-party data providers |
| Activation | Any channel: email, messaging, site personalisation and advertising through hashed lists | Mainly demand-side and supply-side platforms through cookie syncs |
| Privacy exposure | Personal data from a direct relationship, where consent can be recorded | Online identifiers, often collected by parties with no direct relationship to the person |
| Owner | Marketing, CRM or data teams, often with engineering | Media and programmatic teams, often through an agency |
| Outlook as identifiers change | Growing, as first-party data becomes the main targeting asset | Declining, as third-party cookie matching erodes |
The rows that decide the choice are identity and outlook. A CDP's records are tied to identifiers the business collects itself and that browsers do not remove. A DMP's are tied to identifiers that browsers increasingly block, so the same segment reaches fewer people each year.
When to choose Customer data platform (CDP)
Choose a CDP when the business's own customer data is fragmented across systems and that fragmentation is holding marketing back. If the email platform, the shop, the app and the CRM each hold a partial view of the same person, a CDP can join them into one profile and push consistent segments everywhere. The CDP versus CRM comparison covers the related decision about where working records of contacts and deals should live.
Choose it when activation across channels matters. Suppressing recent buyers from acquisition ads, sending a lapsed-customer segment to both email and paid social, or personalising the site for known visitors are CDP jobs, because they depend on knowing who someone is. Its output reaches advertising platforms as hashed customer lists, which those platforms match against their own signed-in users.
Choose it with the warehouse in mind. The data flow and integrations guide describes the usual division: the warehouse as the system of record for history and definitions, the CDP, or a reverse-sync tool playing its role, as the system of action. A CDP bought without that decision tends to duplicate the warehouse's logic rather than act on it.
When to choose Data management platform (DMP)
A DMP is now a narrow choice. It still fits where an advertiser or publisher has a large, established programme of cookie-based segments that remain matchable in the browsers and markets it serves, and where those segments have been tested against a control and shown to pay for their fees. Some publishers also use DMP-style tools to package their own audience data for sale through deals.
It may also persist inside other products. Audience features within demand-side platforms and publisher ad servers do some of the segment work that a standalone DMP once did, and an advertiser using those features may have the capability without the separate contract.
Choose it, in short, only with evidence. The ecosystem guide's advice is to check what share of a DMP's segments can still be matched before renewing, and to run third-party segments against a control line item before scaling, because each segment adds a fee to every impression it touches.
Decision table
| Situation | Choose | Why |
|---|---|---|
| Customer data scattered across shop, app, email and CRM | CDP | Unified profiles are its core job |
| Need to suppress existing customers from acquisition campaigns | CDP | Suppression depends on knowing who the customers are |
| Prospecting audiences for programmatic buying | Platform audience features, CDP seeds and publisher deals | Third-party cookie segments reach fewer people each year |
| Existing DMP contract up for renewal | Test before renewing | Matchable share and paid-for lift decide its value |
| Publisher packaging its own audience data for buyers | DMP-style tools or the ad server's audience features | The publisher holds the relationship and the inventory |
| Strong warehouse and data engineering team | Warehouse with reverse sync, CDP optional | The warehouse can act as the profile store |
| Regulated market with strict consent requirements | CDP built on recorded consent | First-party data from a direct relationship is easier to govern |
What replaces the DMP's job
The DMP did three things: it collected audience data, segmented it and moved it to buying platforms. Each job now has a different home. Collection has moved to first-party dataData a business collects directly from its own customers, visitors and users. gathered under the business's own consent, increasingly through server-side collection. Segmentation has moved into the CDP or the warehouse, where segments are defined once and shared. Movement to advertising now happens through hashed customer lists and platform connectors rather than cookie syncs.
For reaching people the business does not yet know, the replacement is not one tool. Demand-side platforms offer their own audience segments; publishers offer their logged-in audiences through deals; contextual targeting reaches people by what they are reading rather than who they are; and advertising platforms build similar audiences from a CDP's customer seeds. The first-party versus third-party data comparison covers the underlying choice of data source.
Data protection points the same way. The GDPR's definition of personal data includes online identifiers, so cookie-keyed segments are not outside the law simply because they carry no name. Data collected in a direct relationship, with consent recorded at the point of collection, is easier to govern than data assembled by parties the person never dealt with. That is a general direction rather than legal advice, and each market's rules need their own review.
An optician chain retires its DMP
Brindlecote Opticians, a chain of 45 branches, pays £2,800 a month for a DMP whose third-party segments feed its programmatic prospecting. An audit finds that only 30% of its segment audiences can still be matched in its DSP, and a four-week test against a contextual control shows no difference in appointment bookings. Meanwhile its appointment system, website and email platform hold 210,000 customer records that nothing connects. Brindlecote lets the DMP contract lapse, puts £2,000 a month towards a CDP that unifies those records, and uses its segments to suppress customers who booked in the last year and to send a seed of frequent customers to its advertising platforms.
Common mistakes
- Treating the two as the same tool because both hold audiences and both vendors use similar language.
- Renewing a DMP without testing match rates, so the business pays for segments that reach a shrinking share of browsers.
- Calling cookie segments anonymous and therefore unregulated, when online identifiers can be personal data.
- Buying a CDP to fix a data quality problem, which it will faithfully copy into every channel.
- Running a CDP and a warehouse with separate identity logic, so every discrepancy has two candidate causes.
- Expecting a CDP to supply prospecting audiences on its own, when it holds only people the business already knows.
- Choosing by a vendor's label, rather than by which job the tool actually does in the stack.
