I'mBoardDocs
Board Ontology

Board Ontology

Meet Rogue, the I'mBoard Board Ontology — 148 KPIs in the catalog, 27 anchored to third-party standards, all tier-labeled, versioned, citable, open JSON, free MCP for any AI assistant.

Three tools, three definitions of "ARR."

Meet Rogue, the I'mBoard Board Ontology. I'mBoard publishes one catalog you can defend — for every metric on your board agenda. 148 KPIs total: 27 anchored to third-party standards (SMSB, KBCM, Carta, Mercer, SaaS Capital, Retently, NVCA, YC SAFE, SHRM), 121 editorial (I'mBoard-authored, transparently labeled). Every entry ships through the live JSON and MCP — pin a version, filter to the tier you trust. Free to read, free to wire into your AI assistant.

What's here

148 KPIs in the live JSON and MCP. 27 have a dedicated MDX page on this site.

  • 148 KPIs in the live JSON / MCP catalog — full board ontology, tier-labeled. Each entry carries a definitionSource.tier field: published (industry-backed) or editorial (I'mBoard-authored). Pulled live from app.imboard.ai/api/ontology.json.
  • 27 industry-backed KPIs with a dedicated page here — anchored to a third-party standard (SMSB, KBCM, Carta, Mercer, SaaS Capital, Retently, NVCA, YC SAFE, SHRM). Each carries the cited source, formula, benchmarks, and interpretation notes.
  • 121 editorial KPIs ship in the JSON / MCP but don't yet have a dedicated MDX page — they're I'mBoard-authored definitions for KPIs without a public standard. Each release we back-attribute more — track progress in epic #1415.
  • Calculation policy — an optional per-KPI block telling AI agents how to compute the metric from messy source data (inclusion / exclusion rules, required inputs, edge cases, validation checks, common miscomputations). Populated on 25 high-leverage KPIs as of v1.3.0, rolled out in batches (highest-leverage metrics first); more land with each promotion or new addition. See methodology § 5.
  • Metric basis — an optional per-KPI block of enum axes (timeBasis, moneyBasis, cohortBasis, dateBasis, production) you can filter the catalog on deterministically — e.g. every point-in-time metric, every cash-basis metric. See methodology § 6.

Versioning

Every JSON response carries version and releasedAt. The current release is v1.3.0 (2026-05-20). Semver semantics:

  • Patch — typo, description tweak, formula clarification.
  • Minor — KPI added, KPI promoted editorial → published, new optional field.
  • Major — KPI removed, rogueId renamed, required field added.

Immutable snapshots are published at docs.imboard.ai/api/ontology/v/1.3.0/index.json — pin a version for reproducibility. The unversioned URL always returns the latest. See the changelog for what's changed.

Three ways to use it

  • Browse here — start with the domain index below. Industry-backed KPIs have a dedicated page each; editorial-tier KPIs are listed in their domain's page with a link to the live JSON.
  • Fetch the JSON — full live catalog at app.imboard.ai/api/ontology.json (148 entries, tier-labeled, no auth) or the docs-build snapshot at docs.imboard.ai/api/ontology/index.json (same 148, snapshotted at docs build, no auth). Both open in any browser.
  • Wire it into your AI assistant — see the 30-second snippet below.

Plug the ontology into your AI assistant

What you get. Ask Claude, Cursor, Cline, Continue, Goose, or any MCP-compatible client: "what's the SMSB definition of NRR?" or "what's the median CAC payback for early-stage SaaS?" — and get back the cited definition, formula, and benchmarks from this ontology, in-chat, with sources you can defend in a board meeting.

Free signup, token in 30 seconds. Create a free I'mBoard account and a personal access token at Account → API Access. The token is the only thing the MCP server needs to talk to I'mBoard on your behalf. Read-only against the public catalog by default — no data about your company leaves your machine unless you point an agent at your own boards.

Copy-paste mcp.json snippet:

{
  "mcpServers": {
    "imboard": {
      "command": "npx",
      "args": ["-y", "@imboard.ai/mcp-server"],
      "env": {
        "IMBOARD_API_BASE_URL": "https://app.imboard.ai",
        "IMBOARD_API_TOKEN": "imb_pat_YOUR_TOKEN"
      }
    }
  }
}

Or in Claude Code, the one-command equivalent:

claude mcp add imboard \
  -e IMBOARD_API_BASE_URL=https://app.imboard.ai \
  -e IMBOARD_API_TOKEN=imb_pat_YOUR_TOKEN \
  -- npx -y @imboard.ai/mcp-server

Pull the plug any time by removing the entry from your config. The ontology tools are browse_rogue_kpis and get_rogue_kpi.

What I'mBoard adds on top. The open ontology is the foundation. The I'mBoard product builds on it — change tracking and versioning of definitions over time, AI agents that run against your actual numbers using these definitions, a UI to browse, customize, and adopt KPIs for your board, and a CLI for programmatic use. For the full MCP setup, the rest of the 135 tools across 28 domains (most of which do need a token to do anything with your board), and the token flow, see /docs/guides/claude-code-and-mcp.

Why this exists

Three tools, three different ARR numbers, same board meeting. Most dashboards render metrics without saying what they mean — "ARR" alone has at least three common definitions in circulation, and a director who challenges a definition mid-meeting can knock a CEO off their footing for the rest of the agenda.

Every KPI here cites its source. Where a third-party standard exists (SMSB, KBCM, Carta, Mercer, SaaS Capital, Retently), we anchor to it. Where one doesn't yet, I'mBoard authors a definition and labels it editorial — visible, defendable, and clearly distinguished from the industry-backed tier. When I'mBoard's calculation diverges from a cited standard, the divergence is documented on the KPI page.

For the longer story behind Rogue and the X-Men metaphor, see imboard.ai/ontology.

Domains

Finance

Cash, revenue, burn, runway, expenses — 19 KPIs (1 anchored to third-party standards, 18 editorial).

Sales

ARR, bookings, pipeline, deals, forecast — 45 KPIs (10 anchored to third-party standards, 35 editorial).

HR

Headcount, hiring, attrition, compensation — 28 KPIs (4 anchored to third-party standards, 24 editorial).

Product

Roadmap, delivery, R&D efficiency, quality — 16 KPIs (1 anchored to third-party standards, 15 editorial).

Customers

Retention, health, NPS, concentration — 16 KPIs (5 anchored to third-party standards, 11 editorial).

Fundraising

Round status, investor pipeline, cap table — 23 KPIs (5 anchored to third-party standards, 18 editorial).

Operations

Cross-functional operational metrics — 1 KPI (1 anchored to third-party standards).

Machine-readable access

EndpointAuthUse
/api/ontology/index.jsonnoneFull catalog (docs build snapshot, latest)
/api/ontology/v/1.3.0/index.jsonnoneImmutable v1.3.0 snapshot
app.imboard.ai/api/ontology.jsonnoneFull live catalog (148 entries)
/api/ontology/{domain}.jsonnoneDomain-filtered subset
/api/ontology/{domain}/{slug}.jsonnoneSingle KPI
/api/ontology/schema.jsonnoneJSON Schema for the response shapes (validate against #/definitions/OntologyResponse)
/llms.txt · /llms-full.txtnoneAgent discovery

All JSON responses are wrapped in a versioned envelope: { version, releasedAt, kpis } (or { ..., kpi } for per-KPI). Strict-mode consumers filter kpis on definitionSource.tier === 'published'.

How attribution works

Every entry in the JSON surfaces a definitionSource block with:

  • tierpublished (industry-backed) or editorial (I'mBoard-authored)
  • sourceName and sourceUrl — the standard (or I'mBoard) and its canonical page
  • publicationDate — when the cited edition was published
  • attributionNotice — verbatim text required by the source publication, surfaced on every page that uses it

Industry-backed (published) KPIs are surfaced as dedicated pages on this site AND in the live JSON / MCP. Editorial-tier KPIs are surfaced in the JSON / MCP with full attribution metadata; they don't have dedicated MDX pages here yet — see methodology for the policy and epic #1415 for the back-attribution workstream.

On this page