ABOUT CLIENT PORTAL ADVISOR PORTAL

Customer Data Platforms: A Market Landscape for Buyers

Landscape

Customer Data Platforms

An orientation map of the customer data platform market: packaged suites versus warehouse-native approaches, the identity resolution claims to test, and the questions that separate architecture from marketing.

The customer data platform category promises one trustworthy customer record with activation everywhere, and it is a category in the middle of an architectural argument: whether the CDP should be a product you buy or a layer you assemble on the data warehouse you already own. This landscape maps both camps and the claims to test before either receives your budget.

Last reviewed 3 July 2026 · Free and ungated

Pressure-test your shortlist

Selected senior operators from the Global Board have built both the packaged and the warehouse-native version of this stack. Put your shortlist to them confidentially before the architecture hardens.

Pressure-test your shortlist

How a client brief works · What you receive

The problem CDPs promise to end, and the argument about how

Customer data fragments across web, app, CRM, service and commerce systems, and marketing teams want segments, journeys and suppression lists without filing an engineering ticket for each one. Two philosophies now compete to solve this. The packaged CDP ingests events into its own store, resolves identities and ships connectors out of the box. The warehouse-native, or composable, approach keeps customer data in the cloud warehouse you already govern and adds modelling and activation on top. Packaged buys speed if you lack data engineering; composable avoids a second copy of customer data and a second governance regime, but assumes a warehouse and a standing data team that many marketing departments do not yet have behind them.

How to read this landscape

Category boundaries here are contested, and several of the vendor types below would dispute belonging to the same market at all. This landscape is based on public information and Digital Advisory's editorial assessment of the category. It is not a paid ranking, vendor inclusion does not imply endorsement and the landscape should be used as an initial orientation tool rather than a final procurement recommendation.

Evaluation criteria for a contested category

Criterion Why it matters in this category
Identity resolution tested on your data Match rates quoted in sales decks come from ideal conditions. Run the vendor's matching against a sample of your real records and inspect the false merges, not just the headline percentage.
Fit with your warehouse strategy If the organisation is consolidating onto a cloud warehouse, a packaged CDP creates a second copy of customer data with its own logic. Decide the architecture before shortlisting products.
Connector depth where you activate Connector counts flatter. What matters is depth on the five destinations you use daily: fields synced, latency, audience size limits and failure handling.
Consent propagation A CDP concentrates personal data by design, which concentrates regulatory exposure with it. Test how a consent change reaches every downstream destination, and how fast.
Who operates it day to day Some platforms assume marketers self-serve; others assume a standing data team. Buy for the operating model you have, or fund the one the platform needs.

Vendor camps in the CDP market

Provider type Typically strong at Watch for
Packaged enterprise CDPs Identity resolution out of the box, mature connectors, marketer self-service A second copy of customer data; pricing scales with profiles and events
Marketing-cloud CDP modules Native fit with the vendor's own engagement tools, single contract Strongest inside the vendor's walls; neutrality ends at the ecosystem edge
Warehouse-native and composable vendors No data duplication, transparency, engineering-friendly economics Assembly required; marketer self-service depends on your data team
Event collection and pipeline platforms Clean first-party data capture, developer tooling A pipeline is not a customer view; identity and activation can be thin or absent
Vertical and regional CDPs Preconfigured for one industry's data shapes and rules Concentration risk on a small vendor; narrower connector coverage

Traps this category sets for buyers

  • Buying a packaged CDP while the data team builds the warehouse that makes it redundant, on parallel budgets that never meet.
  • Judging identity resolution on the match rate a demo produced, then discovering it merged households, colleagues and strangers.
  • The single customer view becoming one more silo whose numbers disagree with the warehouse and the CRM.
  • Profile-based and event-based pricing that grows with data volume rather than with the value extracted from it.
  • Use cases that still require engineering tickets after purchase, which was the problem the CDP was bought to end.

Questions that cut through the architecture argument

  • Run your identity resolution on a sample of our real records: what is the match rate, and what did it merge that it should not have?
  • If we already run a governed cloud warehouse, what exactly does your product add, and what does it duplicate?
  • For our top five activation destinations, which fields sync, how fast, and what are the audience size and rate limits?
  • How does a consent withdrawal propagate: to which systems, how fast, and how is that evidenced for a regulator?
  • What does our cost look like at twice today's profiles and events, and which use cases can marketers run without engineering?

No vendor pays to appear in a Digital Advisory landscape. Read how landscapes are compiled.

Frequently asked questions

Who compiles this landscape and do CDP vendors have any say in it?

None. It is Digital Advisory's independent editorial assessment of the category, built from public information. Vendors cannot pay for inclusion, no scores are awarded and Digital Advisory has no stake in whether you buy packaged, build composable or do neither.

Packaged CDP or build on the warehouse: is there a sensible default?

Audit two things before choosing: the maturity of your warehouse and the availability of data engineers. With both in place, composable avoids duplicating customer data and usually costs less at scale. Without them, a packaged CDP buys speed this year, and you accept the second copy knowingly. The expensive mistake is deciding by conviction rather than by that audit.

How do we verify identity resolution claims before signing?

Insist on a proof of concept using a representative sample of your own records, and examine three outputs: the match rate, the false merges and the unmatched residue. Ask which matches are deterministic and which are probabilistic, and what confidence thresholds apply. A vendor that resists this test is telling you something useful about the claim.

Keep working the decision

Related decision resources

Comparison Build vs Buy Assessment Data Readiness Assessment Guide Before Selecting a Vendor Calculator Vendor Total Cost Calculator

One customer view, two architectures, one budget. Choose with evidence.

Pressure-test your shortlist