CDPs solve data unification and activation problems effectively. They do not solve measurement, attribution, or data quality problems — and organisations that purchased them for those purposes are discovering the gap at significant cost.
The Problem CDPs Were Designed to Solve — and The Problems They Cannot
Customer data platforms emerged as a response to a specific and genuine problem: the fragmentation of customer data across multiple systems, creating an inability to activate that data consistently across marketing channels. CRM data was in one system. Website behaviour data was in another. Email engagement data was in a third. Point-of-sale transaction data was in a fourth. No single platform could read all four simultaneously, which meant that marketing communications were disconnected, audience targeting was based on incomplete signals, and the customer’s experience was inconsistent across touchpoints. The CDP was designed to solve this by creating a unified customer profile that could be activated across channels in real time.
The problem was real, and CDPs represent a genuine technical solution to it. But the market for CDPs — which grew rapidly from 2017 through 2023 and now includes over 150 vendors in various forms — has been shaped by vendor positioning that has systematically overstated the breadth of the problem that CDPs can solve. CDPs are positioned not merely as data unification tools but as the foundation for measurement, personalisation, analytics, and customer experience management simultaneously. The result is that many organisations have purchased and implemented CDPs to solve problems that the technology was not designed to address, and are discovering — often after significant investment — that the actual problem persists.
The most common mismatch is between the CDP as a data unification and activation tool and the CDP as a measurement platform. A CDP can consolidate customer data and activate unified profiles across channels. It cannot, by itself, measure whether that activation is generating incremental business outcomes. The attribution problem, the incrementality problem, and the causal inference problem that sit at the heart of marketing measurement are not solved by having a more complete customer profile — they require measurement methodologies that are entirely separate from the CDP’s core function.
What CDPs Are Genuinely Effective For
Within their genuine scope, CDPs provide valuable capabilities that do solve real operational problems for marketing organisations. Understanding where CDPs deliver genuine value is necessary for evaluating whether a particular investment is solving the right problem or addressing a symptom of a different underlying issue.
The Wrong Problem: When CDPs Are Purchased to Solve Measurement Failures
The most expensive wrong problem is when a CDP is purchased — at investment levels typically ranging from $500,000 to several million dollars over a multi-year contract — to address the organisation’s inability to measure marketing effectiveness. The reasoning is seductive: if the measurement problem stems from fragmented data, unifying the data in a CDP should solve the measurement problem. In practice, it does not.
A CDP can unify customer data and activate it across channels. It cannot determine whether that activation is generating incremental business outcomes — which is a different problem entirely.
The measurement failures that organisations experience — inability to attribute outcomes across channels, inability to measure long-run brand effects, inability to demonstrate incrementality — are not caused by data fragmentation in the sense that CDPs are designed to address. They are caused by the absence of causal measurement methodologies — incrementality testing, marketing mix modelling, brand tracking — that a unified customer profile does not provide. A more complete customer record fed into a last-click attribution model produces a more complete last-click attribution model. It does not produce a causal measurement of marketing effectiveness.
The second category of wrong problem is when CDPs are purchased to solve a data quality problem that the organisation does not have the governance structures to maintain. A CDP consolidates the data that feeds into it. If the contributing systems contain inaccurate, inconsistent, or incomplete data — which is the normal condition in most large Australian organisations — the CDP will consolidate and surface that poor quality data more efficiently without improving it. The clean room analogy is instructive: a CDP is a container; it does not clean the water.
The Alternative Problems Worth Solving First
For organisations that are considering CDP investment, the diagnostic question is whether the primary capability gap is in data activation — the inability to use available customer data consistently across channels — or in measurement and data quality — the inability to understand what the data means and whether it is accurate. CDP investment addresses the former effectively. The latter requires different investments.
Data quality problems are best addressed through a combination of master data management governance, data stewardship roles with cross-functional authority, and investment in the source systems that generate the data rather than in the consolidation layer that aggregates it. Measurement problems are best addressed through the methodologies described throughout this series: marketing mix modelling, incrementality testing, brand tracking, and the organisational capability to interpret and act on their outputs.
The Right Role for CDPs in a Sophisticated Marketing Stack
CDPs have a legitimate and valuable role in a sophisticated marketing technology stack — specifically in the data unification and activation layer. The organisations that get the most value from their CDP investment are those that have already solved the data quality and measurement methodology problems through other means, and are using the CDP to operationalise the targeting and personalisation decisions that the measurement infrastructure has informed. In this configuration, the CDP is the execution layer for a strategy that is validated and directed by a separate measurement capability.
For boards reviewing marketing technology investment proposals that include CDPs, the relevant question is not whether CDPs are legitimate technology — they are — but whether the problem being solved is genuinely a data unification and activation problem, or whether the CDP is being positioned to solve a measurement or data quality problem that it is not equipped to address. The distinction between these problem types is consequential, both for investment allocation and for the expected returns from the programme. Getting the diagnosis right before committing the capital is substantially cheaper than discovering the mismatch after implementation.