Understand
A typed graph for product knowledge.
UPG models the knowledge a product team produces as a typed, directed graph. The entities and relationships that every framework describes informally become one explicit, queryable structure. This page is the conceptual overview; the specification is the full definition.
“A framework, a PRD, a Jira board, a research synthesis: every piece of product work captures a partial view of the same structure. Named once as a typed graph, they become projections of one source rather than separate documents.”
One structure, many vocabularies.
Frameworks, methods, and tools each describe product work in their own vocabulary. Underneath, they refer to a small, recurring set of objects: people with goals, the jobs they pursue, the opportunities and solutions that follow, and the outcomes that measure them.
UPG names those objects and the relationships among them once, as types. The structure becomes explicit rather than implied by prose, so the same knowledge can be read by a person, a tool, or a model without re-interpretation.
Everything reads and writes the same graph.
The specification is built from four constructs.
Entities
The nouns of product work: persona, job, outcome, feature, metric. Each gets a stable id, properties, and a maturity tier, so the same persona means the same thing in every tool.
Browse entity types →Edges
The verbs. A persona pursues a job; a feature addresses it; a hypothesis targets an outcome. Each edge has a direction, a meaning, and rules about what it can connect.
See edge catalog →Regions
The graph groups into neighbourhoods: Discovery, Strategy, Execution, Validation. Each has its own anchor type and the domains it spans.
Tour the regions →Frameworks
Business Model Canvas, OST, RICE, Jobs-to-be-Done: each is a view of the same graph. Switch views without rewriting what is underneath.
See frameworks →The four constructs above are the core. The specification also defines, in full:
- Lifecycles
- Phase and state machines for entities that progress through stages.
- Playbooks
- Region-anchored sequences for building a graph.
- Approaches
- Ways to work through a graph: plan, inspect, prioritise, trace, reflect.
- Scales and assessments
- How judgments and measures are encoded.
- Validation
- Anti-patterns and the rules that keep a graph sound.