RACI vs. RAPID vs. DACI: Which Decision Model, When

Four people marked A

Picture a steering committee, a composite drawn from common practice. Slide three is the RACI chart, built with care at kickoff. On the row that matters, “recommend go or no-go on the private-label exit,” four people are marked A: the VP Strategy, the CFO, the general counsel and the head of sales. Each of them reads A as “I can stop this.” The meeting runs long, ends with “let’s take it offline,” and nobody can say who decides. The chart was followed perfectly. It just answered a different question.

RACI, RAPID and DACI all hand out letters to people, which is why they get mixed up. They solve different problems. RACI assigns responsibility for tasks and deliverables. RAPID assigns roles in a decision that crosses units and has people who must sign off. DACI assigns roles in a team or project decision, with one driver pushing it to one approver.

That is the thesis of this page. RACI charts rarely fail because people ignore them. They fail because they assign tasks when the fight is over a decision. For the basics of RACI itself, see the RACI matrix page. This one is about choosing.

• Choose RACI when the decision is already made and the question is who does which part of the work.
• Choose RAPID when the decision is contested, crosses business units, functions or partners, and some people (legal, finance, compliance) must agree before it goes forward.
• Choose DACI when a team needs a decision soon, one leader can approve it, and the main risk is drift rather than conflict.

Pick the model by the fight you are having, not by the letters you already know.

The roles side by side

The RAPID entries follow Bain’s own definitions ³; the DACI entries follow Atlassian’s Team Playbook ⁵. The rows that matter are the second and third. RAPID is the only one of the three with a named gatekeeper role. RACI is the only one in which the person who “makes the call” is not clearly defined as a decider.

What each model is built for

RAPID is a proprietary tool created by Bain & Company ³. Paul Rogers and Marcia Blenko set it out in “Who Has the D?” in the January 2006 Harvard Business Review, as a way to clear decisions that stall at boundaries such as global versus local or center versus business unit ¹². Three points are often missed. The letters are not a sequence; the roles are not carried out lockstep in that order ², and the Bridgespan Group’s adaptation, made with Bain’s permission, says the same ⁴. The Agree role is not a late veto; Bridgespan describes it as helping the recommender reach a viable recommendation, not vetoing a decision after it has been made ⁴. And performers should be named early; Bain suggests giving them an input role too, so implementation problems shape the recommendation ³.

DACI is published by Atlassian as a play in its Team Playbook. The Driver is the person responsible for corralling stakeholders, collating the information, setting the scope of the decision and getting it made by the agreed date. The Approver is “the one person (yes: one!) who makes the decision” ⁵. Contributors have “a voice, but not a vote”; the Informed get “no vote, no voice” ⁵. Atlassian sizes the play at 15 minutes of preparation and 60 minutes to run, for three to six people ⁵.

Nobody knows for certain where DACI started. Many secondary sources credit Intuit, some placing it in the 1980s, but we found no Intuit document or other primary source that confirms it, and Atlassian's page gives no origin. Treat the Intuit story as unverified. Atlassian's own page also cites a third-party survey claiming a specific performance benefit for DACI projects; we could not trace that figure to any original publication, so we are not repeating the number here. Treat it as unverified until someone produces the source.

RACI is a responsibility assignment chart for work, and no single origin for it can be traced. Its weakness for decisions is the A. Even Atlassian’s own playbook, describing RACI as a variation, defines Accountable as “the final approver of the work” ⁵. Approving work and choosing between options sound alike. In a contested decision they are different jobs, and the gap between them is where the four-A row comes from.

How to choose

1. Has the decision already been made? If yes, use RACI for the work that follows.
2. Does it cross units, functions or an outside partner, or must legal, finance or compliance sign off? If yes, use RAPID.
3. Is it a team or project decision that one leader can approve, where the risk is drift? Use DACI.
4. Is it cheap to reverse? Skip the framework. Name an owner and a date.

One wrinkle. Atlassian itself suggests DACI for high-stakes and cross-functional decisions, including strategic market entry ⁵. It can work there, as long as the gatekeepers are few and the Approver really can overrule them. When legal or finance hold a requirement nobody can overrule, DACI has no place to put it, and RAPID’s Agree role exists for exactly that case.

Whichever you choose, write the roles down before the analysis starts. Roles assigned after a recommendation exists invite people to claim whichever role lets them block it. When the decision is made, note who held each role in the decision log, so the next fight about who decides can start from what was agreed last time. The full sequence around the roles, from framing the question to getting commitment, is in our decision-making framework.

One decision, three ways

The last sentence is the practical lesson. The models are not rivals. A big decision often uses RAPID to make the call and RACI to carry it out, and the handoff between them is the moment the chart should change.

Where each one breaks

Why the chart is political, and when ambiguity is a choice

Here is what the comparison tables leave out. Assigning decision rights is not an administrative act. It moves power. Bridgespan’s guide says so plainly: mapping how decisions are made exposes the ways in which power flows through the organization, not everyone loves hierarchy, and some leaders prefer to leave control of certain issues ambiguous ⁴.

That explains the four-A row better than any theory of bad charts. Nobody in that meeting misunderstood RACI. Each of them had a reason to keep the right to stop the decision, and RACI let them keep it without saying so. A RAPID chart would have forced the question into the open: which one of you holds the D? That is a harder conversation, which is exactly why it is worth having before the analysis rather than after.

The fair objection is that ambiguity sometimes helps. Early in an exploration, a leader may not want to hand the D to anyone yet. Keeping roles loose avoids a fight before anyone knows what the options are, and lets the CEO move a decision as the picture changes. That is a legitimate choice, with a cost. The cost lands on the recommender, who has to build consensus among people with undeclared vetoes, and on speed, because decisions stall at exactly the boundaries Rogers and Blenko named ². If you leave roles open on purpose, put a date on it: the day the options are clear is the day the D gets assigned.

One more trap for heads of strategy. Bridgespan warns that a leader who keeps a veto “in their pocket” for decisions they don’t like has not actually ceded decision-making power ⁴. A chart that names a decider while the boss keeps a pocket veto is worse than no chart, because it teaches the organization that the chart is fiction. Before you publish one, map who can actually stop the decision; stakeholder mapping is the tool. And when roles are set for a whole initiative, the strategic initiative charter is where they belong.

The letters are the easy part. Any of the three models works if the people on it agree, out loud, about who gets to choose. None of them works if they do not. Choose the model that forces that agreement for the fight you actually have.

Questions

What is the difference between RACI and RAPID?

RACI assigns responsibility for work: who does it, who answers for it, who is consulted and who is informed. RAPID, created by Bain & Company, assigns roles in a decision: who recommends, who must agree, who gives input, who decides and who performs. RAPID names one decider and a gatekeeper role, which RACI lacks, so it suits contested decisions better.

What does DACI stand for?

DACI stands for Driver, Approver, Contributors and Informed. In Atlassian’s Team Playbook, the Driver gathers information and gets a decision made by an agreed date, the Approver is the one person who makes the decision, Contributors have a voice but not a vote, and the Informed are told the outcome. It suits team and project decisions that need to move quickly.

Who created the DACI framework?

It is not clear. Many secondary sources credit Intuit, some placing it in the 1980s, but we found no Intuit document or other primary source confirming this, and Atlassian’s widely used DACI play gives no origin. RAPID, by contrast, is documented: Bain & Company created it, and Paul Rogers and Marcia Blenko described it in Harvard Business Review in 2006.

Can you use a RACI matrix for decisions?

You can, but it is a weak fit. RACI was built to assign responsibility for tasks, and its Accountable role often means approving or owning a deliverable rather than choosing between options. For contested or cross-functional decisions, use RAPID or DACI to settle who decides, then use RACI to organize the work that carries the decision out.

Where to go next

• Getting a Room to Decide: A Decision-Making Framework
• Stakeholder Mapping for Strategy Work
• Decision Log
• Governance
• Decision Stakes Classifier (free question set)