What is a composable CDP in retail?

A composable CDP builds unified customer profiles directly in the enterprise data warehouse or lakehouse, using modular, best-of-breed tools. Unlike a traditional packaged CDP, it avoids copying data into a proprietary store, preserving governance and making the lakehouse the single system of data gravity. In retail, it is the preferred architectural foundation for customer data management.

Why is a composable CDP not enough for personalisation?

A composable CDP unifies customer data but does not natively model the other entities that retail decisions depend on: product availability, store context, promotion mechanics, margin constraints, and consent status. Without these connections, personalisation recommendations may be irrelevant, unprofitable, or non-compliant.

What is a Retail Semantic Data Model?

A Retail Semantic Data Model is a business-language layer that sits on top of the composable lakehouse and gives consistent, queryable meaning to customer, product, store, inventory, promotion, loyalty, margin, and consent entities. It is the shared operating contract between analytics, marketing, merchandising, e-commerce, and AI systems.

How does store context change personalisation?

Store context influences availability, assortment, and local demand patterns. Without incorporating store-level signals, personalisation can recommend products that are unavailable or irrelevant in a shopper’s context, reducing effectiveness and customer experience.

Why should promotion rules live at category level instead of SKU level?

Many retail promotions, especially brand-funded trade promotions, are structured at the category or subcategory level and cascade to qualifying SKUs. Storing rules only at the SKU level loses the promotion’s commercial structure, making it impossible to calculate cannibalisation, halo effects, or whether the trade investment achieved its objective.

How does a semantic layer improve next-best-action?

Next-best-action in retail requires reading across multiple entity types simultaneously: shopper context, basket composition, product taxonomy, store format, inventory, promotion eligibility, and consent. The semantic layer is what enables this cross-entity inference, turning fragmented data into a coherent, actionable signal.

What does DPDP change for Indian retailers?

India’s Digital Personal Data Protection Rules require itemised consent notices, purpose-based retention, access and erasure mechanisms, and breach notifications. For retailers, this means consent and governance must be embedded in the activation and decisioning layer, not managed in a separate compliance system.

Is this relevant outside India?

Yes. While India’s omnichannel complexity, WhatsApp commerce, and DPDP regulations make the semantic gap especially visible, the underlying argument applies globally. Any retail market with omnichannel activation, format diversity, promotion complexity, and data-protection requirements faces the same structural need for a semantic layer on top of the composable CDP.