ADR 0001: Rename the project to impact-model

Status: Accepted Date: 2026-11-01

Context

The project began as a specification for a Theory of Change, and was named theoryofchange-spec. It now holds more than that. One document records the outcomes and causal links of a program, the assumptions and evidence behind them, the interventions that bring them about, how progress is measured, who needs to see which numbers, and where each claim came from. The same document is drawn as a theory of change, a logic model, an outcome map, and a logframe, and it exports to OKRs.

Two things made the name wrong:

• Scope. The name says the document is a Theory of Change, and the other methods look like parts of it or exports from it. A logic model is not a part of a theory of change. It is a more operational projection of the same underlying thing, and each projection leaves out what another keeps.
• Direction. The project is aimed at more than nonprofits. A product team has the same need to say why an intervention should move a metric, and to keep what it believes apart from what it has observed.

The published releases, v0.1.0 and v0.2.0, had no known users, so renaming cost little.

Decision

The project is the impact model. The theory of change, the logic model, the pathway of change, the outcome map, and the logframe are projections of it, and so are strategy formats: OKRs, which the model already exports, and V2MOM, which is among the follow-ups. "Projection" names a method's rendering of the model. It is not report.View, the data a report is built from.

Was	Is
github.com/ProductBuildersHQ/theoryofchange-spec	github.com/ProductBuildersHQ/impact-model
command theoryofchange	command impact-model
packages pkg/toc, pkg/tocio	pkg/impact, pkg/impactio
type toc.Document	impact.Model
kind: TheoryOfChange, TheoryOfChangeStatus	kind: ImpactModel, ImpactModelStatus
schemas theory-of-change, theory-of-change-status	impact-model, impact-model-status
theory.yaml	impact-model.yaml
fields tocId, childToc	modelId, childModel
cross-document reference toc:<id>#<element>	model:<id>#<element>
lint rule IDs TOC-STR-001	IMP-STR-001
flag --toc	flag --model
MCP tools toc_*, skills toc-*	impact_*, impact-*

What did not change

• The method names. report pathway, logic-model, outcome-map, and logframe are still the names of those projections, and the research brief still describes the Theory of Change method.
• specVersion, which stays 0.1. The reader accepts any 0.x.
• Identifiers of the work. The initiative INIT-THEORYOFCHANGESPEC-001 and the roadmap items RMI-THEORYOFCHANGESPEC-* keep their names, since an identifier is not a label. New roadmap items take the repository's new slug, RMI-IMPACTMODEL-001 onward, so the two series sit side by side in one initiative.
• The quality rubric and the specification template in spec/. They are the theory of change quality checklist and outline, and will become the rubric of that projection.
• The example's own title. The membership network example is a theory of change, and says so.

Consequences

• Releases. v0.1.0 and v0.2.0 were published under the old module path and stay there. A module cannot change its path under an existing tag, so the first release under the new name is v0.3.0.
• Migration. A document written for v0.2.0 needs kind: ImpactModel, the renamed fields, and level: medium_term if it predates that change. Nothing else in the format moves.
• Links. GitHub redirects the old repository address. The documentation site moves to a new path.
• No alias. The old command name is not kept, since nobody uses it.

Alternatives considered

Name	Why not
Keep theoryofchange-spec	Understates the scope and ranks the other methods below one of them
impact-model-spec	Matches the naming of some sibling repositories, but this repository also ships a command, an SDK, and an MCP server
structured-impact-model	Longer, and "structured" adds nothing a schema does not already say
impact-spec	Broad enough to mean almost anything

impact-model has one real drawback: the phrase also means financial and environmental impact assessment. The tagline says what it is, and "impact" is already the name of the top outcome level.

A leadership framework with a similar name exists in another project. The two are unrelated and independent, and nothing here depends on it.

Follow-ups, not decided here

• A theory-of-change report kind, so each method is a named projection.
• context.values, and an export to V2MOM beside the existing export to OKRs, with the OKR and V2MOM documents staying separate and linked to the model by ID.
• A report saying which projections a given model has enough information to draw.
• Relations between indicators, such as one predicting another, with the evidence for each.