Composable CDP Strategy
Key insights from our webinar with Adam Greco (Hightouch) & Jörg Hopmann:
- A Composable CDP activates customer data directly from the data warehouse. No copies, no proprietary schemas, no vendor lock-in.
- Packaged CDPs typically take 6–12 months to implement. A Composable CDP proof of concept is achievable in two to three weeks.
- Use cases that are impossible with a packaged CDP, such as contribution-margin-driven offers, inventory-guided advertising, and B2B buying team behavior, become straightforward when the CDP sits on top of your warehouse.
- A unified data foundation is the prerequisite for reliable AI in marketing.
This post summarizes the key insights from our webinar with Adam Greco (Hightouch) and our CEO Jörg Hopmann. Watch the full recording here, free of charge.
A scenario that comes up regularly: a marketing team sends a retargeting campaign to a customer who has just filed a support ticket about the very product they couldn’t purchase. Or a retailer continues targeting a buyer online for an item they purchased in-store twenty minutes earlier. Not malicious, just the CDP not knowing what the warehouse already had.
That problem is exactly what a Composable CDP strategy is designed to address. On June 17, 2026, we hosted a webinar together with Adam Greco, Product Evangelist at Hightouch, exploring why more and more companies are moving away from traditional, packaged Customer Data Platforms, and what a thoughtful marketing analytics strategy needs to deliver in that context.
“The minute you’re only using parts of the data that are in the packaged CDP, you can actually make mistakes – or do more harm than good.”
Adam Greco, Product Evangelist, Hightouch
“It’s one of the most impactful projects you can have as a consultancy – introducing CDPs, optimizing CDPs – because everything is measurable.”
Jörg Hopmann, CEO, Hopmann Marketing Analytics
What Is a Composable CDP, and Why Now?
A Customer Data Platform (CDP) collects, unifies, and activates customer data for marketing and sales. Classical, packaged CDPs do this by importing data into their own system. That was the standard for years.
The problem doesn’t appear at implementation. It grows over time. Cloud data warehouses like Snowflake or BigQuery have become the central storage layer for customer data. They hold far more data than any packaged CDP. The result: marketing teams work from a partial view of their customers, while the complete picture sits unused in the warehouse.
Source: Adam Greco, Hightouch
The Composable CDP inverts this logic. Instead of exporting data into an external system, it connects directly to the existing data warehouse and activates from there. Hightouch has been instrumental in shaping this approach. The term “composable” reflects the fact that CDP functionality, including identity resolution, audience building, and data activation, is assembled from best-of-breed components rather than delivered as a monolithic package.
10 Reasons Organizations Are Moving to a Composable CDP Strategy
Adam Greco presented ten reasons in the webinar, documented in full by Hightouch, why organizations are switching. The following stand out most clearly from our daily consulting practice.
Access to All Customer Data
Packaged CDPs only contain data that has been explicitly imported into them: often marketing data, rarely CRM data, almost never transactional or service history. The warehouse, by contrast, has everything. Building audiences on the basis of complete customer profiles requires a warehouse-first approach.
Shorter Implementation Timelines
Packaged CDPs require extensive tagging, data migration, proprietary data modeling, and continuous synchronization. A Composable CDP builds on data that already exists in the warehouse. Packaged CDPs typically take 6–12 months to implement and carry heavy ongoing maintenance costs (Hightouch, 2026). Jörg Hopmann shared concrete timelines from our own projects: a proof of concept is achievable in two to three weeks; a full rollout with the first productive use cases takes around six weeks.
Data Schema Flexibility
Packaged CDPs impose a schema that has to work for thousands of customers. Adam Greco described a concrete example in the webinar: a pet supply company had built a complex data model linking pets, owners, and products in ways no packaged CDP could replicate. The Composable CDP simply mirrors the existing warehouse schema without forcing or simplifying it.
Better Identity Resolution
The more identifiers a CDP can use for profile matching, the more accurate the result. Because the warehouse holds the most customer identifiers, the quality of identity resolution is naturally higher there than in a system that only sees part of the data.
Privacy and Data Security
Every copy of customer data represents additional risk, both regulatory and from a security standpoint. Composable CDPs don’t create data copies in the classical sense. This matters especially in the DACH region: by avoiding additional data silos, Composable CDPs simplify GDPR right-to-erasure requests considerably, since there’s only one central data source to act on.
No Vendor Lock-in
Identity graphs, audience definitions, and journey logic built inside a packaged CDP belong to the vendor, not the company. When the vendor changes, that knowledge is lost. In the composable approach, all of these outputs exist as tables in the warehouse. The vendor can be replaced at any time without losing the data intelligence that has been built up.
AI Readiness
AI models need complete, consistent, and clean data. Customer data fragmented across multiple systems makes accurate AI outputs practically impossible. Any company planning to introduce AI-powered personalization or predictive analytics needs a unified data foundation first. The Composable CDP is not just a MarTech decision. It’s a building block of an AI data strategy.
What Only the Composable CDP Can Do: Use Cases from the Warehouse
Jörg Hopmann presented four use case types in the webinar that are either impossible or extremely difficult to implement with packaged CDPs.
Contribution-margin-driven offers: Product margins are highly sensitive data that no company wants to export to an external system. In the warehouse, they’re available. A Composable CDP can steer offers and campaigns based on contribution margins, not just revenue or click-through rates. That changes the economic impact of campaigns fundamentally.
Inventory-guided advertising: Packaged CDPs are built around person-centric entities. Inventory data, product availability, and campaign timing are not people. In the warehouse, these non-person entities can be linked directly to customer data and used to control campaigns accordingly.
B2B buying teams: In B2B contexts, purchasing decisions are rarely made by a single person. They’re made by teams. A CDP that only knows individual person profiles cannot model buying team behavior. The warehouse holds these cross-person relationships, and a Composable CDP can use them directly.
Long-term customer history: Some packaged CDPs delete inactive customers after 90 days. That makes them structurally unsuitable for industries with long purchase cycles, such as automotive, insurance, or industrial equipment. The warehouse stores the full history, independent of activity windows.
What a Composable CDP Implementation Looks Like in Practice
A three-phase approach has proven effective in consulting projects, and one we outlined in the webinar.
Phase 1: Evaluate quickly. Grant access to selected warehouse tables, define and build the first use case, connect existing tools. No new tagging, no migration. Companies that want to test without their own dataset can start with a demo warehouse.
Phase 2: Adopt from day one. Marketing teams work in an intuitive interface, no SQL knowledge required. Data teams stay in the warehouse environment they already know. No parallel data management, no new system on the data side.
Phase 3: Prove the value. Since the warehouse and the CDP share the same data source, control groups can be set up cleanly. Results are measurable directly in the warehouse. That enables genuine incrementality tests, not just before-and-after comparisons.
Want to see the full session? Adam Greco and Jörg Hopmann walk through all ten reasons, the use case framework, and the AI readiness architecture in detail.
Watch the webinar recording, free of chargeHow to Prioritize Your Composable CDP Use Cases
Which use case should come first? The framework we presented in the webinar describes each use case along four dimensions: target audience, channels and actions, required data, and primary KPI. Impact and feasibility are then scored, and use cases prioritized accordingly.
Typical starting use cases: win-back of high-value inactive customers, churn prevention, cross- and upsell, and paid media efficiency through suppression of known buyers. Each works best when drawing on complete warehouse data, not an imported subset.
The Hybrid Question: Is There a Middle Ground?
The webinar Q&A surfaced a question that comes up regularly in evaluation projects: how sensible is a hybrid approach, where a packaged CDP is enriched with warehouse data?
Adam Greco’s answer was direct: the moment data is copied into an external system, the problems return. Vendor lock-in, data discrepancies, doubled security requirements, limited schema flexibility. A hybrid model sounds like a compromise, but in practice often delivers the worst of both worlds. The recommendation: either fully packaged or fully composable.
What This Means for Your Customer Data Platform Strategy
A Composable CDP strategy is not just a tool decision. It’s a data strategy decision. Companies that run a modern data warehouse and want to drive marketing activation from that warehouse have made a clear architectural choice: data stays where it is. Activation comes to the data, not the other way around.
This has implications for collaboration between marketing and data teams, for the selection of MarTech partners, and for the long-term AI data strategy. In our CDP consulting work, we support this process from use case definition through technical audit, architecture design, implementation, and governance, as a certified Hightouch partner for the DACH region.
If you want to assess whether your data foundation is ready for a Composable CDP rollout, get in touch. The first step is usually shorter than expected.
What our customers often want to know.
Webinar FAQ on Composable CDP Strategy
What is the difference between a Composable CDP and a classical CDP?
A classical (packaged) CDP imports customer data into its own system and manages it there. A Composable CDP connects directly to the existing cloud data warehouse and activates data from there, without creating copies. The warehouse remains the single source of truth. Capabilities like identity resolution, audience building, and data activation are provided via specialized tools such as Hightouch.
Is a Composable CDP only for marketing, or does it benefit other teams too?
Packaged CDPs are typically driven by marketing teams, which can create friction with data teams who feel bypassed. Composable CDPs tend to bring marketing and data teams together rather than push them apart, because data teams keep working in their existing warehouse environment while marketers gain access to richer data. In practice, this cross-functional alignment is one of the strongest adoption arguments for the composable approach.
How long does a Composable CDP implementation take?
A proof of concept is achievable in two to three weeks. A full rollout with the first productive use cases typically takes around six weeks, provided that relevant data is already in the warehouse. Compared to packaged CDPs, which can take 6–12 months to implement, that’s a significant difference.
Is a Composable CDP GDPR-compliant?
Yes, and in many cases it’s more advantageous from a privacy perspective than a classical CDP. Since no customer data is copied into an external system, there are fewer locations where data needs to be stored and managed. Right-to-erasure requests can be handled through the central warehouse without having to clean up multiple systems.
Can a Composable CDP handle real-time personalization?
Yes. A common concern is that warehouse-based architectures are batch-oriented and therefore unsuitable for real-time use cases. In practice, Hightouch addresses this by combining cached warehouse audiences with real-time event streaming, enabling same-session personalization without copying data into a separate system. Adam Greco addressed this directly in the webinar Q&A.
How does Hopmann support Composable CDP implementations?
As a certified Hightouch partner for the DACH region, we support companies from the initial use case workshop through technical audit, architecture design, implementation, and governance. We speak both the language of marketing teams and that of data teams, and we create the connection between the two that a successful CDP rollout requires.
Do we need a mature data team before we can start with a Composable CDP?
Not necessarily. A common concern is that the data warehouse needs to be “complete” before a Composable CDP makes sense. In practice, it never is. The more useful question is: do you have the data in your warehouse that your first use case needs? If yes, you can start. Adam Greco made another observation in the webinar worth noting: organizations that start using a Composable CDP tend to put more data into their warehouse over time, not less. Marketing teams see what becomes possible and ask for more. Data teams see their work being used and get more motivated to expand it. The CDP and the warehouse reinforce each other.