Skip to main content

Essence of Base Product Thinking

1. Foundational Idea​

At its core, a Product is a unit of value offered for use or consumption under explicit or implicit agreements.
Regardless of whether the product is data, AI, software, service, or physical, all products share a common essence:

  • They can be identified and versioned.
  • They have an owner or responsible agent.
  • They expose interfaces through which value is delivered.
  • They are subject to contracts, obligations, or expectations (e.g., SLA, accuracy, fairness).
  • They evolve through a lifecycle from creation to retirement.
  • They are associated with assets and provenance describing their origins.
  • They can be evaluated against measurable criteria (e.g., quality, performance, compliance).

This common kernel transcends any particular domain and provides the basis for a Base Product Specification (BPS).

2. Core Abstractions​

The following abstractions constitute the minimum viable ontology of a product:

  1. Identity
    A product must have a unique identifier and a human-readable name.

  2. Version
    Products evolve. Each release must be traceable and immutable.

  3. Ownership and Accountability
    A responsible agent (individual, team, or organization) must be declared.

  4. Interfaces and Capabilities
    Products expose input and output channels, defining how they interact with the world.

  5. Contracts and Obligations
    Products carry explicit or implicit guarantees, constraints, or commitments.

  6. Assets and Provenance
    Products derive from underlying assets and activities whose lineage must be captured.

  7. Evaluation and Metrics
    Products are assessed against defined metrics for quality, performance, or compliance.

  8. Lifecycle and Status
    Products move through states such as Draft, Published, Deprecated, and Retired.

3. Universality​

These abstractions apply consistently:

  • Data Product: Identified dataset, with versions, ports, freshness SLAs, lineage, and quality metrics.
  • AI Product: Identified model, with versions, input/output interfaces, fairness obligations, training dataset lineage, accuracy metrics.
  • Software or Service Product: Identified API, with versions, input/output endpoints, uptime guarantees, operational metrics.
  • Physical Product: Identified manufactured item, with versions (model numbers), warranty contracts, bill of materials, safety metrics.

Thus, BPS is domain-neutral, ensuring any product can be described as an extension.

4. Distinction from Domain Specifications​

Domain-specific specifications (DPDS/DPROD, APDS/APROD, etc.) extend this base by adding:

  • Specialized semantics (e.g., schema for datasets, metrics for AI bias).
  • Domain-relevant lifecycle states (e.g., AI model retraining, dataset curation).
  • Compliance requirements (e.g., EU AI Act risk classification).

The BPS does not replace domain specifications. Instead, it anchors them in a shared ontology.

5. Essence Statement​

The essence of Base Product thinking is that:

  • Every product, regardless of domain, can be represented as an identity with accountability,
  • exposing interfaces that deliver value,
  • governed by contracts and obligations,
  • linked to provenance and assets,
  • measured through evaluations and metrics,
  • and evolving through a lifecycle of versions.

This essence is the unifying principle of the Base Product Specification.