Skip to scenario
Corpus-wide refactoring

Restructure the source without losing what depends on it

Years of migrations can leave documentation nested too deeply, inconsistently organized, and linked in circles. A refactoring workflow needs the authoritative source and its dependencies before it moves files. The reader benefit must survive technical validation.

See the comparison
The task in this example

Restructure the docs site without breaking reuse, links, or metadata.

Business goal
Increase efficiency
Business metric
Content cost

A flatter site must preserve its source relationships

The same corpus needs a clearer structure. The comparison changes whether the transformation starts from published HTML or the authoritative structured source.

The website is reconstructed as source

Published HTML is converted back into files before restructuring.

Metadata, reuse, and source references are lost or misinterpreted during reconstruction.

Input
Published output
Dependencies
Incomplete reconstruction
Validation
Breakage discovered after transformation
The team must repair the corpus before judging the navigation

The source and dependencies guide the change

A reviewed plan drives deterministic transformations on a copy of the source.

Links, reuse, metadata, and builds are checked; writers resolve explicit exceptions before release.

Input
Authoritative structured source
Dependencies
Mapped before changes
Validation
Build and relationship checks before release
The team can assess a working new structure

Measure both transformation effort and reader success

Valid builds establish that the new corpus can be published. They do not establish that readers find tasks more easily. The efficiency case includes exception cleanup and the user checks needed to judge the new structure.

  1. Plan with dependencies

    Review the proposed hierarchy and affected references.

    • Time to finish
  2. Transform and validate

    Preserve the corpus’s technical contracts.

    • Nothing broken
    • Valid builds
  3. Test navigation

    Observe readers finding representative topics.

    • Search success
    • Clicks to a topic
    • Navigation complaints
Business metric

Content cost

Compare total refactoring and subsequent maintenance effort, including planning, validation, cleanup, and reader testing.

What to measure and how to interpret it

Save the source baseline and reader-task baseline before changing structure. Track intended URL changes separately from broken links.

Time to finish
Elapsed time from an approved scope to validated, reviewed output, including exception cleanup.
Nothing broken
Unresolved failures in links, reuse, metadata, and publishing dependencies; define the checked contracts explicitly.
Valid builds
Successful required output builds with validation errors resolved or explicitly dispositioned.
Search success
Share of representative findability tasks ending at the correct topic.
Clicks to a topic
Steps to the intended topic for defined navigation tasks; fewer clicks alone does not establish better understanding.
Navigation complaints
Reader reports attributed to structure or findability, classified and compared with usage.

Check the business result

Include information-architecture review, deterministic transformation work, validation, redirects, and manual cleanup in content cost. Self-service is a possible downstream benefit to test after findability improves.

Keep a transformation manifest reviewers can inspect

Stable IDs and dependency relationships let code change locations while preserving meaning and reuse. The plan should identify intended moves and the checks that must remain true.

Corpus source / Approved restructuring plan
Refactoring manifest

A source move with explicit contracts

Move the approved topic group while preserving identity, references, reuse, and applicability metadata.

Baseline
Versioned authoritative source snapshot
Proposed move
Old location → reviewed new location
Dependencies
Incoming links, reuse, maps, publishing rules
Contracts
Stable IDs and required metadata preserved
Release checks
Builds, redirects, findability tasks, exception disposition

Intelligent content

Use the authoritative structured source so topic identities, metadata, and reuse survive the transformation.

Taxonomy

Preserve and validate classification values while changing organization; a new folder is not a reason to erase applicability.

Ontology and knowledge graph

Map links, reuse, and publishing dependencies before moving content. Use the dependency map to validate the transformation and identify exceptions.

A practical starting point

Refactor one bounded branch on a copy. Establish technical contracts and reader tasks, approve the plan, and inspect every exception before broadening the transformation.

The model proposes structure. Information architects and writers approve it; deterministic code applies approved transformations, and people resolve semantic exceptions.

See how this scenario works

Make rollback part of the plan

Retain the source baseline, transformation manifest, redirect map, and prior publishing configuration. Define who can restore them and what failure triggers rollback. A clean build does not remove the need for an operational recovery path.

Lief
Lief
Amber
Amber
Dan
Dan
Explore the work with Intuitive Stack

What must survive your next documentation restructure?

Intuitive Stack can help define the source dependencies, reader tasks, and reviewable transformation plan for a bounded refactoring pilot.

  • Authoritative source and dependencies
  • Deterministic transformations
  • Technical and reader verification