Decision making is a team sport, but it’s hard to make a good decision if team members don’t know which position they are playing. That’s why clearly defining roles and who will fill them is central to effective decision making. Determining who makes the decision (“who has the D”) is necessary but not sufficient. That person can only do their job with the help of others who can offer helpful recommendations, insightful input, and the right reviews. And a decision has no effect unless someone is accountable for executing it.

That’s why our approach to decision roles, RAPID, is grounded in those five distinct functions, clarifying who will:

• Recommend the decision that should be made
• Agree that the recommendation meets requirements, such as legal or regulatory mandates
• Perform the decision once it’s made
• Give input, critical expertise, experience, or information that will help shape the recommendation
• Decide and commit the organization to action.

Assigning these roles, then training people to understand their responsibilities and play their part, requires concerted effort. But these roles hold the key to high performance, to delivering the outcomes of quality, speed, effort, and follow-through that companies so badly need.

The Recommend role

What it is. The person in the R role typically does 80% of the work involved in a decision. The recommender gathers relevant input and proposes a course of action—and sometimes alternative courses, complete with pros and cons. Rs are the quarterbacks of the decision process, coordinating the other roles so that the decision maker’s choices are as clear, simple, and timely as possible. If they do their job well, Rs will usually see their recommendation accepted by the D, the decision maker.

Getting the R right. Some organizations fail to formally identify who is acting as the R, stalling decisions. When one mining company, for example, expanded into a new region, corporate headquarters pushed to create an external relations organization to support the new mine. Headquarters assumed the regional vice president would develop a recommendation for the new structure. The regional VP expected headquarters to do it—after all, it was their idea. The upshot: a big delay, until the executives responsible eventually clarified the R role.

Other organizations have too many recommenders. A media company gave all of its editorial units an R on content, leading to recurrent bottlenecks. It solved the problem when it assigned one person to integrate the inputs, set priorities, and recommend trade-offs. There should be only one R.

The R can start by sitting down with the decision maker to establish the parameters for the decision, the form of the recommendation, and the level of rigor required. They can then clarify decision roles and processes and map out a series of meetings to gather the necessary input and sign-offs. A large technology company tries to capture these best practices by arming every R with a checklist of issues to consider, such as “Who should be consulted to develop a complete recommendation?” and “What is the right trade-off between rigor and speed for the recommendation?”

The R may or may not report to the D, and they will sometimes play other roles as well, including the D, I, or P.

The Input role

What it is. People in the I role offer input, providing critical expertise, experience, data, and perspectives. They help the recommender assess whatever trade-offs the decision may require. Increasingly, artificial intelligence can play the role of the I, synthesizing and framing relevant data and expertise and handling data processing and pattern recognition, while humans exercise judgment, safeguard ethics, and take accountability.

AI could, for example, forecast demand so buyers can determine how much inventory to order. It’s up to the buyer to decide whether to trust the AI forecast or perhaps adjust it based on factors the model might not have accounted for, such as an upcoming competitor promotion.

Another job of the I is helping to think through the implications of the decision, such as risks and implementation challenges. Note, however, that the I role is strictly advisory. Recommenders should consider all input, but they don’t have to reflect every point of view in the final recommendation.

Getting the I right. Ideally, a suitable candidate is someone who can provide relevant data, expertise, or experience. The I is also assigned to people who will be responsible for implementation, along with individuals in parts of the organization affected by the decision. Including them ensures that downstream issues are considered, thus improving decision quality and speeding buy-in. Serving as an I is a big responsibility, and people in most high-performing organizations earn the right to influence a decision by providing credible, high-quality analytics and logic in a timely manner.

Companies with consensus-driven cultures often put too many people in an I role. Dozens of unnecessary people in meetings clog the process without improving decision quality. In particular, decisions that can easily be adjusted or undone if the outcome is unsatisfactory (“two-way door” decisions) have limited downside and should be made with appropriate—but not excessive—input to ensure speed and efficiency.

And sometimes senior executives get involved in topics that are best delegated. A consumer products company developed a process to ensure that top leaders focused only on major innovations. The company intended for senior leaders to work on high-priority projects but failed to define which decisions needed their input. Soon the executives found themselves arguing about the height of the letters on product packages when they should have been focused on more important issues.

Other companies, however, don’t have enough people in an I role and don’t get the specific input they need. At a pharmaceutical company, the drug-development decision process included input on scientific and medical factors but lacked input about payers’ willingness to cover the drug. As a result, the company found itself investing in pharmaceuticals that insurers were unlikely to pay for.

The I may or may not also be a P.

The Agree role

What it is. Unlike the I, the A offers a form of input that can’t be ignored. If the person in the A role doesn’t agree, the recommender and the “agreer” must work together to find a mutually acceptable solution. In the rare cases when that’s not possible, the recommender needs to share the A’s dissent with the D, who makes the decision with all perspectives in mind.

For examples of the A role in action, consider how a safety executive might sign off on a change in work processes or a brand manager agree that a given decision won’t hurt the brand. Risk management and finance functions often play an A role to ensure that decisions fit the company’s overall risk profile and budget constraints. People in the A role don’t have veto power to block, undo, or reject a decision. Rather, they work with the R to ensure feasible alternatives are considered and to shape a recommendation.

Getting the A right. Many decisions require no A at all. Others may need only one or two people in the role. Agreement should always be part of developing the recommendation, and it should come before the decision, not after it. Ideally, the recommender and the A work things out between themselves, with the R amending the proposal until the A’s concerns are addressed.

It’s important to specify not just the A role but the scope to which the A applies. At a medical device company, the group responsible for regulatory compliance had to sign off on the company’s marketing brochures. In time, they were exercising their A on every aspect of the brochures, down to the colors, slowing down production. When the company reassigned decision rights, it narrowed the scope of the regulatory team’s A to focus on ensuring the text complied with federal regulations.

Some recommendations may be missing an essential A, leading to the potential rework of a decision. Some people in the A role may wait for the decision rather than weigh in on the recommendation; when they slap down what they wrongly believe is a veto card, they undermine the decision maker’s authority, cause further delay, and miss the opportunity to ensure their perspectives contribute to a robust recommendation. Occasionally a company will assign an A to a senior executive just because that individual “should have a chance to weigh in.” But that confuses the I role with the A.

Leading companies typically reduce the need for constant sign-offs by providing guidelines to their businesses. Only if a decision goes outside the guidelines does it require approval from someone in an A role. Take a retail bank that for years had run a cumbersome process to create direct mail campaigns, with every mailing requiring approvals from finance (on the financial assumptions) and risk (on the mailing’s loss rate projections). Even identical campaigns required a reexamination. When the company finally provided guidelines, marketing managers could make their own decisions within them. A new offer or a mailing to a new population that fell outside of those guidelines continued to require approval from finance and risk.

There are times, of course, when a recommender feels that the A is constraining the recommendation too much and proceeds, highlighting the A’s concerns to the D. In effect, the decision maker is prepared to take the risks that the A flagged.

When a company begins to specify decision roles, many people who thought they had A responsibilities will be redefined into the I role. This may feel like a demotion, so it’s important that everyone understand both the merits of an A role and the importance of an I role.

The Decide role

What it is. The person in the D role makes the decision and commits the organization to action. The D is the single point of accountability for bringing the decision to closure, including resolving any impasse between the R and an A when they cannot reconcile their views.

The D’s job begins well before the final call. Early in the process, the D and the R should align on the decision’s parameters, timing, criteria, required inputs, and the form and level of rigor expected in the recommendation. Once the decision is made, the D communicates both the decision and its rationale to everyone involved and makes sure the organization has what it needs to execute.

Getting the D right. As a rule, each decision should have only one D. When two or more people believe they have the final say, decisions can become a tug-of-war; when no one does, they can simply stall. If the D must sit with a group, the group should agree in advance how it will reach a decision—for example, by majority vote or by giving the chair the final say.

The D must sit at the right level. They should have enough perspective to understand the relevant trade-offs and commit the necessary resources, but decision authority should remain as close to implementation as is practical. Pushing too many decisions upward can overwhelm senior executives and separate decisions from the people with the most operational knowledge. The strongest decision makers combine good judgment and a grasp of the trade-offs with a bias for action and a clear understanding of the organization that will carry out the decision.

The Perform role

What it is. Companies need to assign a P promptly once a decision is made to ensure a timely transition to the execution phase, give the P the opportunity to flag implementation issues, and encourage their buy-in to the decision they will be executing.

Getting the P right. Sometimes the P is never defined, so a decision is never implemented. A beverage company, for instance, decided to relocate its IT center of excellence to a European city. No P was assigned, so no one looked for office space, figured out how to consolidate current IT operations with the new center, and so on. When a new CIO came on board, she reopened the entire decision, requiring the company to, in effect, make the same decision twice.

Once a major decision is made and moved into execution, it will likely involve a set of significant follow-on decisions. Implementation teams need to apply the same rigor and RAPID-style analysis to these execution-related decisions that they applied to the original one.

The P may or may not also be an I.

Clear decision roles are an essential foundation, but they are only part of making and executing effective decisions. Companies also need clarity on which decisions matter, how and when they should be made, and the behaviors expected of everyone involved. Bringing those elements together—the what, who, how, when, and behaviors—creates a decision-making ecosystem in which RAPID can deliver its full potential.

The authors would like to acknowledge Marcia Blenko for pioneering Bain’s research into decision-making roles.