Compass Instruments · CHANGE & TRANSFORM
Most projects do not fail at go-live. They fail quietly, months earlier, in conditions everyone could see and no one rated honestly. This is the interim read — taken on evidence, at any point in the value stream.
Green status reports are the most expensive document in a transformation.
The Project Charter sets the contract. The Change Readiness Assessment reads the ground before the work starts. The Reinforcement & Institutionalization Plan closes it. This instrument sits between them: a fast, evidence-based read on whether a project already underway is actually healthy.
Twelve conditions in four domains — Direction, Design, Delivery and Adoption. Each carries a RAG rating, an owner by role, the evidence seen and the corrective action. A condition rated on feeling is not rated. A red without a decision is a forecast.
Instrument · Project Health Check
A project is rarely lost in one decision. It is lost in twelve small conditions nobody rated honestly. Rate each condition on evidence you could show a board — not on how the project feels this week.
Project health
0% · overall Not rated · 0 red · 0 amber · 0 green · 12 unrated
No conditions rated yet. An unrated project is not a healthy project — it is an unread one.
Direction
0% Not rated
Is the project still solving the problem it was chartered to solve?
FOUNDATION
Design
0% Not rated
Has the future state been designed, or only described?
SIMPLIFY
Delivery
0% Not rated
Is the work being governed, or only reported?
LEAD
Adoption
0% Not rated
Will the organization actually work the new way?
GROW · ARC
Is the project still solving the problem it was chartered to solve?
Not rated
A current charter states the problem, scope, measures and decision rights — and matches the work being done.
Evidence required: Signed charter, dated within the current phase, with change history.
Failure mode: The team is delivering something no one re-approved.
Not rated. No owner named.
Not rated
The sponsor makes decisions, removes barriers and is visible to the affected organization.
Evidence required: Decisions made by the sponsor in the last cadence, and where they appeared.
Failure mode: Sponsorship is a name on a slide; escalations die in the middle.
Not rated. No owner named.
Not rated
The benefit is quantified, owned by the business, and still true at today's assumptions.
Evidence required: Benefit model with a named business owner and a revalidation date.
Failure mode: The project completes and no one can prove it was worth doing.
Not rated. No owner named.
Not rated
Scope changes go through one decision path with cost, time and benefit impact stated.
Evidence required: Change log showing decisions, impacts and who approved them.
Failure mode: Scope grows quietly and the date is defended by cutting adoption work.
Not rated. No owner named.
Has the future state been designed, or only described?
Not rated
The future way of working is designed to role, handoff and decision level — not described in principles.
Evidence required: Future-state process design with roles, handoffs and decision points.
Failure mode: Go-live becomes the design workshop.
Not rated. No owner named.
Not rated
Process, data and system configuration are aligned; data quality is measured, not assumed.
Evidence required: Data quality measures with thresholds, and open defects by severity.
Failure mode: The system works and the data makes it unusable.
Not rated. No owner named.
Is the work being governed, or only reported?
Not rated
Every open decision has an owner, a date and the evidence required to make it.
Evidence required: Open decision register, aging shown.
Failure mode: The plan slips one week at a time while decisions wait.
Not rated. No owner named.
Not rated
The plan is resourced, sequenced and has been re-baselined honestly when reality changed.
Evidence required: Current baseline, variance, and the last re-baseline decision.
Failure mode: Green status reports until the month before go-live.
Not rated. No owner named.
Not rated
The people doing the work have protected time; backfill is real, not intended.
Evidence required: Committed hours by role against demand, and what was stopped to free them.
Failure mode: The project runs on the goodwill of six exhausted people.
Not rated. No owner named.
Not rated
Risks and issues are surfaced early, rated consistently, and unwelcome news travels upward safely.
Evidence required: Risk register movement and at least one escalation the sponsor acted on.
Failure mode: Everyone knows the date is wrong and no one says it in the room.
Not rated. No owner named.
Will the organization actually work the new way?
Not rated
Stakeholders are mapped, ADKAR® barriers are diagnosed, and managers are prepared to lead the change.
Evidence required: Stakeholder map, barrier diagnostic and manager readiness measures.
Failure mode: Training is mistaken for adoption; usage collapses after week three.
Not rated. No owner named.
Not rated
Standard work, measures, ownership and ARC review triggers are designed before go-live, not after.
Evidence required: Reinforcement & Institutionalization Plan started, with owners named.
Failure mode: The project closes and the old way returns within two quarters.
Not rated. No owner named.
This is a management instrument, not audit, legal or personnel advice. Record conditions, owners by role and evidence — never third-party personal data. Entries stay in the browser that created them until you export the check.
Where this leads
A red condition becomes an ARC cycle: analyze the cause, refine the design, commit the correction. Adoption reds route to the barrier diagnostic; sustain reds route to the institutionalization plan.
Not sure where to start?
Under a minute. From this page, most leaders begin with The organization.
