RACI Model: Who Does the Work and Who Owns the Outcome
The RACI Model helps you assign four letters on named activities: Responsible (does the work), Accountable (owns the outcome), Consulted (gives input before), and Informed (is updated after). Every activity needs exactly one Accountable. It is a responsibility grid, not a promise that conflict will…
Framework Card
- Name:
- RACI Model
- Goal:
- Assign Responsible, Accountable, Consulted, and Informed on named activities so each activity has one outcome owner and a clear doer.
- Flow:
- Responsible → Accountable → Consulted → Informed
- Best For:
- Cross-functional work with blurry ownership; Unblocking a decision that has no clear owner; Mapping roles when a new person joins existing tasks
Why it matters
Cross-functional work fails in predictable ways. Two people think they own the decision. Nobody writes the spec. A long Consulted list turns every task into a meeting. Updates go to no one, or to everyone.
Talking about “better communication” does not fix that. The missing object is a shared picture of who does the work and who owns the outcome on each activity.
RACI is a habit of putting letters on a grid so those gaps and overlaps are visible. It does not staff the work or teach the skill. It names the roles.
What it is
RACI is a responsibility assignment matrix. One axis is activities (tasks or decisions that could stall). The other axis is people or roles. Cells hold R, A, C, or I.
- Responsible executes the work. More than one R is possible; too many R’s usually means the work is still too coarse.
- Accountable owns the outcome and the yes or no. Exactly one A per activity.
- Consulted is two-way input before action.
- Informed is a one-way update after action.
The same person may be both R and A, especially on a small team. That is allowed. What is not allowed is two Accountable owners for one activity.
RACI is not a job description, not an org chart, and not a format for a speech.
How it works
A useful RACI is letters on real activities, not a poster of the four words.
1. List the activities and the people
Name tasks or decisions at the level where ownership could actually blur. “Ship the campaign” may be too coarse. “Open the design file” is too fine. Name the people or roles who might hold a letter.
2. Assign R, A, C, I
For each activity, mark who does the work, who owns the outcome, who must be consulted before, and who only needs an update after. Leave cells empty when a person has no role on that activity.
3. Check one-A, gaps, and pile-ups
Every activity needs one A and at least one path to getting the work done (usually an R, which may be the same person as A). Scan for two A’s, zero A’s, and C lists that will stall the work.
4. Confirm the matrix with the named people
The grid is a draft until the people on it can live with the letters. When the work or the roster changes, update the letters. A stale matrix is a snapshot of last month’s politics.
How it compares
When another lens fits better, or when you need a complementary view, these frameworks do different jobs. They are not interchangeable labels for the same question.
| Framework | What it helps you see | How it differs from RACI |
|---|---|---|
| Stakeholder Saliency Model | Power, legitimacy, urgency | Ranks people. RACI assigns letters to tasks. |
| Work Breakdown Structure | How work is decomposed | Breaks scope. RACI needs those activities already named. |
| DACI | Driver, Approver, Contributor, Informed | A decision-rights cousin. Different letters, often used for decisions. |
| RAPID | Recommend, Agree, Perform, Input, Decide | A decision process. RACI is a standing activity matrix. |
| PREP Framework | Point, reason, example, restated point | A talk structure. Not a role grid. |
The RACI Model is the lens for who does the work and who owns the outcome. Other methods help when the question is influence, scope, or a spoken recommendation.
When to Use This Framework
- Kickoff of a cross-functional project. Several teams will touch the same deliverables.
- A decision with no owner. The work sits because nobody is Accountable.
- A new person on existing tasks. Letters need to move when the people change.
Example
A concrete example makes the structure easier to reuse when you are under uncertainty.
Example: A campaign with two “owners”
A marketing lead and a product lead both think they approve launch copy. Legal is copied on everything. Sales hears about the date in Slack.
Activities might include: brief, draft copy, legal review, final approve, sales enablement.
A possible picture (illustrative):
- Draft copy: Designer (R), Marketing lead (A), Product (C), Sales (I).
- Final approve: Marketing lead (R and A), Legal (C), Sales (I).
- Sales enablement: Sales lead (R and A), Marketing (C), Product (I).
Implication: Final approve now has one A. Legal is Consulted on approval, not Informed after the tweet goes out. The matrix did not write the copy. It stopped two people from both believing they were Accountable.
Example: An incident process with no owner
A startup has “someone should write the postmortem.” After an outage, the document never appears.
Activity: postmortem write-up. Before RACI: three engineers feel Responsible, nobody is Accountable. After: on-call engineer (R), engineering manager (A), support lead (C), company (I).
Implication: Combining R and A on a three-person team would also have been valid. The missing piece was the single A, not a ban on wearing two letters.
Takeaway
What the RACI Model can help with
- Making doer and outcome owner visible on named activities
- Separating two-way Consulted from one-way Informed
- Catching two A’s or zero A’s
- Updating letters when people or work change
What the RACI Model cannot replace
- Stakeholder power mapping. Stakeholder saliency ranks power, legitimacy, and urgency. It does not assign task letters.
- Scope decomposition. A work-breakdown structure breaks the work. RACI assigns roles on work you already named.
- Other decision-rights grids. DACI and RAPID use different letters and are often aimed at decisions rather than standing task ownership.
- Job design and staffing. A perfect matrix cannot create a skill or a headcount that is not there.
- A spoken message structure. The production URL mentions stakeholder communication. PREP and SCQA structure talk. They do not assign roles.
Honest scope: RACI structures role assignment on a defined activity list. It does not replace the methods above, and it should not be sold as conflict elimination.
Frequently asked questions
Yes. On a small team that is common. The hard rule is exactly one Accountable per activity, not a ban on combining R and A.
Two A’s create a fight over the yes or no. Zero A’s create a gap. The work may still have several Responsible people.
Break them to the level where ownership could reasonably blur. If two people would both claim the same letter, the activity is a candidate. If nobody would argue, you are too fine.
No. Consulted is input before action (two-way). Informed is an update after (one-way). A long Consulted list is a common way to stall work.
Not by itself. It makes role confusion visible. Skill gaps, missing budget, and unclear activities remain separate problems.