Identara Try Axis ↗

Open Framework · v0.1 Public Draft

The Open IGA Operating Framework

The instruction manual for running an identity governance program. Six layers in build order, forty-seven numbered decisions, and four metrics that read whether the program responds. Open, tool-agnostic, and free to adapt under CC BY 4.0.

Status · v0.1 public draft · nothing ratified · review invited
The operating core carries normative language for review. Two metric mappings, Trust Gradient on Prioritization and Justification Half-Life on Process, are proposed and unconfirmed. The change process and the review panel required for ratification are defined in GOVERNANCE.md; every version is recorded in the changelog.
Maintainer
Vidyaa Ganesh
Publisher
Identara
Version
v0.1 draft ·
License
CC BY 4.0 · MIT (code)
Suggested Citation
Ganesh, V. (2026). “Open IGA Operating Framework” (v0.1 draft). Identara. https://doi.org/10.5281/zenodo.21251831
Style Identara house DOI 10.5281/zenodo.21251831 This exact draft 10.5281/zenodo.21251832

The DOI above covers every version, so citations keep resolving as the framework matures. To cite a single decision, use its identifier: Open IGA Operating Framework, M2.

What This Is

The layer under the tools.

The tools handle the mechanics of granting and removing access. This framework covers everything around them: why the program exists, who owns the access decisions, what the program covers, what gets attention first, how the day-to-day processes run, and how often everything gets checked.

The identity field has controls catalogues, bodies of knowledge, and maturity models. What it has not had is a published account of that operating layer. The material exists, but it is held inside organizations or delivered through consulting engagements, so every program that stands up ends up rebuilding the same operating model alone, without a baseline to measure against. This framework publishes an open, adaptable version so the work stops being repeated in isolation.

Every normative statement in the core carries an identifier, M1 through C8, and every failure mode a number, so a program can gap-assess against the framework chapter by chapter, and review feedback can cite the exact statement it disputes.

Fig. 1: The operating core, six layers in build order M1–C8
why who what which how when 01 02 03 04 05 06 MANDATE OWNERSHIP SCOPE PRIORITIZATION PROCESS CADENCE M1–M9 O1–O8 S1–S7 P1–P6 PS1–PS9 C1–C8 ENTITLEMENT DRIFT TRUST GRADIENT JUSTIFICATION H-L GOVERNANCE LAG CONFIRMED PROPOSED, UNCONFIRMED why who what which how when MANDATE OWNERSHIP SCOPE PRIORITIZATION PROCESS CADENCE M1–M9 O1–O8 S1–S7 P1–P6 PS1–PS9 C1–C8 EDR TG JHL GL
Each layer consumes the output of the layers above it. The four metrics attach where they instrument: solid connections are confirmed mappings, dashed are proposed and carry no normative weight until confirmed against pilot data.

The Operating Core

Six layers. Each answers one question.

An organization starting from nothing builds in this order. An organization already running a program holds each layer up as a checklist, and the failure modes show what is broken.

01 · Mandatewhy
Why the program exists
Fixes the program’s purpose in writing, with authority behind it. The charter is the first artifact any program produces, and it carries its own review cycle.
M1–M9 · 9 decisions5 failure modes
02 · Ownershipwho
Who owns the decisions
Distributes authority into named roles, decision rights, and escalation routes, and keeps the people who build the program separate from the people who run and govern it.
O1–O8 · 8 decisions5 failure modes
03 · Scopewhat
What the program covers
Bounds what is governed: populations, resources, and entitlement granularity, with an onboarding gate so the scope register grows deliberately from day one.
S1–S7 · 7 decisions5 failure modes
04 · Prioritizationwhich
What gets attention first
Sequences governance across the estate by risk, and hosts the on-ramps for greenfield, bluefield, and brownfield starting states.
P1–P6 · 6 decisions4 failure modes
05 · Processhow
How the day-to-day runs
Defines the mechanics: joiner-mover-leaver lifecycle, requests, certification, revocation, and exceptions that expire instead of accumulating.
PS1–PS9 · 9 decisions9 failure modes
06 · Cadencewhen
How often everything gets checked
Sets the clock: review frequencies, expiry, decision time bounds, and measurement windows, including the cadence table that reviews itself.
C1–C8 · 8 decisions4 failure modes

Instrumentation

Four metrics read whether it responds.

The Operational Metrics Specification gives each one a formal definition and calculation method. No benchmark thresholds; the values are the program’s own.

EDR
Entitlement Drift Rate
The rate at which entitlements change across a governed population, decomposed into gross velocity, net drift, and unreviewed drift.
Instruments Scope · confirmed
GL
Governance Lag
The elapsed time between an access state becoming inappropriate and the program detecting it. The true risk window.
Instruments Cadence · confirmed
JHL
Justification Half-Life
The estimated time before a grant’s business justification loses half its original relevance.
Proposed for Process · unconfirmed
TG
Trust Gradient
A continuously updated confidence score for standing access, blending validation age, usage recency, context stability, and peer alignment.
Proposed for Prioritization · unconfirmed

The formal definitions derive from the published research: Measuring What Moves →

Reading Paths

Three ways in.

New to identity governance
Read it as a manual.
Start with the terminology, then each chapter’s purpose and failure modes. The observable signals work as a symptom checklist for any organization.
Running a program already
Read it as an assessment.
Gap-assess against the numbered decisions chapter by chapter. Every unmet statement is a backlog item, and the framework’s own tier logic sequences the backlog.
Reviewing the framework
Read it to argue.
The contestable stances are flagged in the text, and every statement has an identifier so feedback can cite it directly. The review section below is the front door.

And one organization does all of it in the worked example: Fernway Software, fictional, every layer decided and every worksheet filled in. Read the Fernway example →

Companions

Ten companions. Eight you fill in.

Non-normative templates and worksheets, in Markdown and Word. Once filled in, they stop being templates and become the program’s operating documents. The other two companions: the explorer, hosted here, and a presentation kit in the repository.

  • MandateProgram charter templatemddocx
  • OwnershipDecision-rights and RACI startermddocx
  • ScopeScope register templatemddocx
  • PrioritizationTier criteria and on-ramp worksheetmddocx
  • ProcessLifecycle checklistmddocx
  • ProcessCertification checklistmddocx
  • CadenceCadence table templatemddocx
  • Cross-cuttingPlatform capability checklistmddocx

Review

Published to be argued with.

Nothing in the draft is ratified, and the stances most worth contesting are flagged in the text. Each pill below opens a pre-titled GitHub issue; cite the identifier and make the case.

  1. Read one chapter, file one issue. Every decision has an identifier, so cite it: GitHub issues →

  2. Run an internal IGA operating model? A profile documents how the core adapts to your organizational shape, and it lets others reuse the adaptation: the profile template →

  3. Prefer email? vidyaa@identara.ca works too. Name the decision identifier if you can.

On the roadmap, in intended order and not as commitments to a date: a pilot reference dataset to ground the metrics, a standards crosswalk annex, practitioner review through IDPro and the conference circuit, standards-community review at IIW XLIII in November 2026, and ratification of the core after review.

Sits Alongside

Where it fits.

ISO and NIST specify controls. The IDPro Body of Knowledge explains concepts. Maturity models score programs. This framework addresses the layer beneath them, and it connects to the rest of the work here.

Follow the draft

One email when v0.2 ships and one when the core is ratified. Nothing else.