Skip to main content

Tenants, workspaces, projects, and targets

Hierarchy​

A Tenant is the GEO product boundary beneath an Organization. A Workspace isolates operational configuration and data. API keys, providers, projects, dashboards, tracking sites and external connections must never cross Workspace ownership.

An Organization can contain root Tenants. A root Tenant can optionally contain one level of Child Tenants for regional or operational separation. Users who can access the root Tenant can discover its Child Tenants, while each Child Tenant owns its own isolated Workspaces. Nested Child Tenants are not supported.

Projects​

A project groups targets, markets, languages, topics, prompts, schedules and business objectives. Use separate projects when ownership, measurement cadence, target portfolio, market, or reporting audience differs.

Target Graph​

Supported targets include organization, brand, merchant, website, page, content, product line, product, service, campaign, message, claim, person, topic, market, audience and channel asset.

Targets carry canonical identity, aliases, classification (owned, competitor, or contextual classification), markets, languages, metadata, owner, access policy, status and immutable version history. Relations are typed, weighted and time-bounded. Cycle protection prevents invalid graph structures.

  • Model the parent brand once; relate products, campaigns, pages and channels instead of duplicating them.
  • Classify direct competitors explicitly before interpreting share of voice.
  • Use aliases for spelling, local-language and legal-name variants.
  • Give campaigns and products their own targets when they need independent prompts, measurement or attribution.
  • Retire obsolete targets rather than rewriting historical evidence.

Import and merge​

The Target Graph API supports idempotent import and canonical target merge. Use a unique Idempotency-Key for each mutation. A merge preserves traceability; it must not silently move data across workspaces.

See Data model and lifecycle for graph fields, relation rules, observation lineage and resource state machines.