Skip to main content

Release Types

The framework encodes delivery methodology as twelve release types, each defining a different ordered set of in-scope artifacts and the commands that apply to them. When you run /wire:new and select a release type, the framework instantiates that process definition into the release's status.md file.

Typerelease_typeScopeTypical Duration
Discovery (Shape Up)discoveryProblem definition → pitch → release brief → sprint plan1–2 weeks
Discovery (SOP / Canonical)sop_discoveryWide-ranging structured discovery leading to Findings Playback3–6 weeks
Full Platformfull_platformSOW → production dashboards + trained users2–3 weeks
Dashboard-Firstdashboard_firstInteractive mocks drive data model; seed data enables immediate dbt1–2 weeks
Pipeline + dbtpipeline_onlyNew data pipeline + dbt transformation layer1–2 weeks
dbt Developmentdbt_developmentAnalytics engineering on existing infrastructure1 week
Dashboard Extensiondashboard_extensionNew dashboards on an existing semantic layer3–5 days
EnablementenablementTraining and documentation for an existing platform2–3 days
Agentic Commerceagentic_commerceAI-powered ecommerce storefront1–4 weeks
Platform Migrationplatform_migrationFull lifecycle migration from one warehouse stack to another4–16 weeks
Agentic Data Stackagentic_data_stackAI analytics overlay for an existing data platform4–6 weeks
DroughtydroughtySchema introspection and base-layer generation1–3 days
CustomcustomBespoke scope derived from SoW or project documentsVaries

Choosing the right release type

  • New engagement, scope can be shaped in 1–2 weeksDiscovery (Shape Up)
  • New engagement, scope genuinely unknown, requires structured discoveryDiscovery (SOP / Canonical)
  • Client needs a new data source connected end-to-end to a dashboardFull Platform
  • Early stakeholder feedback via interactive mocks before building the data layerDashboard-First
  • Client has a BI tool and just needs new data flowing inPipeline + dbt
  • Data is already in the warehouse; need to build the transformation layerdbt Development
  • Semantic layer already has the data; adding new dashboardsDashboard Extension
  • Platform exists; engaged to train and document itEnablement
  • Building an AI-powered ecommerce storefrontAgentic Commerce
  • Migrating an existing data platform between warehousesPlatform Migration
  • Client wants an AI that answers business questions reliably from their warehouseAgentic Data Stack
  • Need to map an existing warehouse quickly before starting design workDroughty (discovery mode)
  • Bespoke deliverables that don't fit any standard typeCustom

Key distinctions

Discovery (Shape Up) vs Discovery (SOP / Canonical): Use Shape Up when the problem domain is understood and you can shape a solution in a week or two. Use SOP / Canonical when you genuinely do not yet know what to build, stakeholder alignment is low, or this is the first analytics engagement at the client.

Full Platform vs Dashboard-First: Both produce the same end result. Full Platform follows the traditional flow: requirements → conceptual model → pipeline design → data model → dbt → dashboards. Dashboard-First inverts this: requirements → interactive dashboard mocks → visualization catalog → data model → seed data → dbt → dashboards → data refactor.

When to start with Platform Migration vs a discovery release: A discovery release is strongly recommended before starting a migration if the scope is not yet confirmed — migration is irreversible once Fivetran connectors are cut over.