Team effectiveness
·
10 min read
Key findings
A decision right has five parts: the class of decision, the recommender, the decider, the limit and the informing rule.
Hackman divides authority into four functions, and a team given authority but detailed procedures cannot exercise it.
In a West Monroe survey of 1,214 leaders, managers named approval layers and unclear decision rights, while executives named technology or skills gaps.
Bain guidance dated 7 October 2026 gives each decision one decider and one recommender, and says the Agree role carries no veto.
A decision counts as settled when it is not reopened within four weeks without new information, and that share can be tracked.
Richard Hackman divides team authority into four functions: executing the work, managing the work process, designing the team and its supports, and setting direction.
A decision rights framework for a team names, for each recurring decision, who recommends, who decides, who is consulted and who is told afterwards. It also sets the limit within which the decider acts without asking. The team builds it from its own last quarter, and the leader signs it in front of them.
A team meets weekly, discusses everything, and settles almost nothing without the leader present.
The leader sees the constraint and delegates more.
Six weeks later the same decisions return to the inbox with more context attached.
The work moved and the authority stayed, and a framework changes that only if the leader gives something away.
What A Decision Right Contains
A decision right names who can settle a question without asking, and within what limits.
It has five parts, and most teams have agreed none of them.
The class of decision, described concretely
The role that prepares the recommendation
The one role that decides
The limit, in money, reversibility or visibility
Who is consulted before and told after, with a deadline
The last part makes the limit safe.
Leaders who delegate verbally often skip it.
RACI, DACI And RAPID Compared
Three published models cover most of what the leading guides describe, and each answers a different question.
RACI assigns Responsible, Accountable, Consulted and Informed across the tasks in a piece of work. It describes who does what.
DACI assigns Driver, Approver, Contributors and Informed for a single decision. It has one Approver and a Driver who runs the process and recommends.
RAPID, from Bain, assigns Recommend, Agree, Perform, Input and Decide. It has one decider and one recommender for each decision.
Bain guidance published on 7 October 2026 says the recommender typically does 80 per cent of the work in a decision.
The guidance gives no base for that figure, so it is best read as a practitioner estimate.
It also says the Agree role carries no veto.
An earlier Bain page leaves the decider to settle any disagreement between Agree and Recommend.
A role that can block a decision works as a veto whatever it is called.
The team should therefore decide deliberately whether any role holds one.
Umbrex describes DACI as simpler and faster for everyday cross-functional decisions.
It describes RAPID as the fuller model for complex governance.
Bain says RAPID is not needed for every decision and suggests starting with decisions of high value or high frequency.
For a leader and six direct reports, DACI-style roles on ten to fifteen recurring decisions are usually enough.
RAPID earns its weight where legal, finance or risk must confirm a requirement before a decision can stand.
None of the three models covers the leader who signs the chart and then overrides it.
That is where most friction in a real team begins.
Why Teams Escalate
People escalate for reasons that are rational from where they sit, and three account for most of it.
The limit was never stated, so any decision might be the one that was too big.
A previous decision was reversed in public, which taught the team what happens next.
The leader answers questions readily, which makes asking cheaper than deciding.
The third is the most common and the hardest for a leader to hear.
The behaviour producing it looks like helpfulness.
Evidence On Decision Speed
West Monroe published a survey on 27 January 2026, fielded in November 2025.
It covered 214 C-suite executives and 1,000 managers at US companies with annual revenue of at least $250 million.
Managers pointed to excessive approval layers, unclear decision rights and risk-averse leadership as barriers to moving faster.
Executives more often named technology or skills gaps.
The release comes from a consultancy that sells related services.
It states neither a sampling method nor a margin of error.
Even so, the split matters to an adviser.
The leader asking for help may hold a view of the cause that the team does not share.
A separate read of who decided what, taken from the team itself, gives the adviser a second source before agreeing to the leader's diagnosis.
What Hackman Says About Authority
Richard Hackman lists clearly specified authority to manage the work as one of four features of a real team.
The other three are a team task, clear boundaries and stable membership.
He divides authority into four functions.
They are executing the work, monitoring and managing the work process, designing the team and its supports, and setting direction.
A manager-led team holds authority over execution only.
A self-managing team adds management of its own work process.
A self-designing team adds the design of the team and its context.
A self-governing team sets its own direction as well.
Most teams an HR business partner meets are meant to be self-managing.
The leader keeps design and direction, and the team holds the other two.
The failure Hackman describes is a manager who tells a team it holds authority and accountability.
The same manager then specifies procedures in such detail that members cannot exercise either.
His example is the Skylab 3 crew, whose time was scheduled minute by minute from Mission Control.
The crew eventually turned off the radio.
A decision rights table is the working version of the authority matrix.
It says which of the four functions the team holds, one decision at a time.
Four Levels Of Delegation
Most confusion resolves when the leader states which of four levels applies to a given question.
You decide, and I do not need to know
You decide, and tell me afterwards
You recommend, and I decide
I decide, and I want your input
All four are legitimate, and the damage comes from leaving the level unsaid.
A team that believes it sits at level one while it operates at level three stops bringing recommendations.
It usually stops after the second reversal.
The level belongs at the point the work is assigned, because a disagreement is too late.
Setting The Limit
Leaders find levels easy and limits uncomfortable, because a limit is a number they can be held to.
Three kinds of limit cover most work.
Money is simple and usually set too low to be useful.
Reversibility is more useful, because a decision that can be undone within a week sits at a different level from one that cannot.
Visibility asks whether anyone outside the function will notice the decision.
Reversibility tracks the actual risk, and the money limit only approximates it.
A decision counts as settled when it was taken at the first meeting where it was listed and not reopened in the following four weeks without new information.
A Worked Sequence
The sequence below assumes a regional operations director, an HR business partner and the director's seven direct reports.
Week one, contracting. The business partner asks the director which three decisions she would like to see less often, and which one she would never hand over
Week one, evidence. The business partner and the director's chief of staff list last quarter's decisions from minutes and email, marking each as taken, escalated or reversed
Week two, individual read. Each direct report marks who decided each item before seeing the director's marks, and the differences become the findings
Week two, working session. The team spends ninety minutes assigning recommender, decider, consulted parties, informed parties and limit to each item
Same session, levels. The director states a level for each decision aloud, and anything she cannot state is marked unresolved
Week three, record. One page is signed and kept where the team can find it, and meeting agendas are sorted into operating, decision and review items
Day 30 and day 90, review. The team counts what moved and adjusts no more than three levels per review
The list usually runs to ten or fifteen decisions.
It comes from last quarter's evidence rather than from job descriptions.
The argument in the working session is the useful part.
It shows where the director and the team held different assumptions.
The individual read is the evidence the director cannot easily dispute, because it is her own team's account of who decided.
Applied Frameworks, a consultancy that built a framework for its own business, offers two further rules.
It advises keeping true consensus decisions to about 10 per cent of the list.
It also advises revisiting the document at least twice a year.
The Conversation With The Leader
The leader holds the decision rights, and the adviser makes each choice visible so that the leader makes it knowingly.
A limit is a number the leader can be held to, which explains most of the reluctance.
The leader also remains accountable upward for every decision on the table.
Two questions open the conversation without asking for a concession.
Which decisions has this team escalated that you would have been content to see taken without you?
Which decision of theirs did you reverse in the last year, and what did the team learn from it?
The second question usually produces the leader's own explanation for the escalation pattern.
A leader who wants to keep nearly everything can be offered a staged version.
Level three for a quarter, a count of what moved and a review date is a smaller request than permanent delegation.
The wider conversation is covered in advising a leader you disagree with.
Where Rights Break
Five failures account for most framework charts that sit unused.
The veto. The Agree or Contributor role becomes a block, and Umbrex lists hidden veto players among the common pitfalls of DACI
The override. The leader signs the table and then settles a listed decision personally, so the team treats the table as decoration
The unlisted decision. The dispute concerns something the first list missed, which is why a review at day 30 matters
Rights pushed too far down. The decision needs information the team lacks, trades off against another team's objectives, or has consequences beyond its horizon
No group rule. A decision that belongs to the whole team has no agreed way of being closed
Reversal is sometimes necessary and always expensive, and its cost depends on how it is done.
Reverse it in private, and say what information the leader held that the team did not.
Then adjust the level or limit explicitly.
A reversal without an adjustment teaches the team to escalate everything.
That is the outcome the leader was trying to avoid.
Group Decision Rules
Where a decision belongs to the team rather than to one person, the closing rule has to be agreed beforehand.
Sam Kaner and colleagues name six common decision rules.
They are unanimous agreement, majority vote, the person in charge deciding after discussion, the person in charge deciding without discussion, delegation, and flipping a coin.
They define a decision rule as the mechanism that answers how a group knows it has made a decision.
Groups without one produce a recognisable pattern.
One member acts on a decision that others believe is still open.
The member who acted is called impulsive, and the member who did not is called disloyal.
The same guide notes a problem with the person in charge deciding after discussion.
Some members offer the advice they think the boss wants to hear.
Kaner's gradients of agreement replace yes and no with a scale.
It runs from whole-hearted endorsement through support with reservations to veto.
A poll with no vetoes and only lukewarm support counts as unanimous, and Kaner treats that as adequate when the stakes are low.
It is not enough when the work needs sustained effort from the people who voted.
A poll with strong support and a few objectors raises a choice between overriding them and searching for a more inclusive proposal.
Kaner says lukewarm compromises will not do where everyone's strong support is needed.
Across Two Teams
Most decision friction sits between teams rather than inside them.
Two questions settle most of it, and neither takes long.
Which decisions require both teams to agree, listed explicitly.
Who arbitrates when they disagree, named before the argument begins.
The second question is the one nobody asks until it is needed.
By then the choice of arbiter has become the dispute.
Evidence After A Quarter
Three counts tell the adviser whether the rights have taken hold.
Decisions taken without escalation
Items on the leader's agenda that could sit lower
Days from a question being raised to it being settled
A decision counts as settled when it was taken at the first meeting where it was listed and not reopened in the following four weeks without new information.
The share of decisions that meet that standard is the figure an operating director will recognise.
It can be reported without any claim about autonomy.
The tally of reopened decisions is also the plainest answer to why a team keeps revisiting what was decided.
Where the counts do not move, the cause usually lies in a condition outside the table, which team diagnosis before team building sets out.
The meetings that follow from clear rights are covered in workshops that end in decisions and in disagreement in a leadership meeting.
Where The Tools Sit
In Trusted Talent Advisor, component 02, The Stakeholder Conversation Toolkit, holds the conversation with the leader who made the request.
Component 05, Facilitating the Talent Forum, covers closing a group decision against an agreed evidence standard.
Related reading: psychological safety team level.
Sources
1. Hackman, J. R. Leading Teams: Setting the Stage for Great Performances. Harvard Business School Press.
2. Kaner, S., Lind, L., Toldi, C. et al. Facilitator's Guide to Participatory Decision-Making. Community At Work, 2007.
3. Cochemé, A. and Mottura, L. Great Decisions, Not a Solo Performance. Bain & Company, 7 October 2026.
4. Bain & Company. RAPID Decision Making. Bain & Company, 13 October 2023.
5. West Monroe. Slow Decisions Are Costing Companies Millions in Lost Revenue, New West Monroe Research Finds. Press release, 27 January 2026.
6. Umbrex. DACI Decision-Making Framework. Umbrex.
7. Tanner, J. Decision Rights Framework: Better Decisions, Faster. Applied Frameworks, 24 March 2021.
From the system
Trusted Talent Advisor sets out the full method: seven components, 51 tools, The Talent Template Pack and The Prompt Library.
See the system
Questions this answers
What are decision rights in an organisation?
Decision rights say who may settle a named class of decision, who prepares the recommendation, who is consulted or informed, and the limit inside which the decider acts without asking. Most teams have agreed none of these parts, so decisions drift upward.
What is the difference between RACI and RAPID?
RACI assigns roles across the tasks in a piece of work. RAPID, from Bain, assigns Recommend, Agree, Perform, Input and Decide for a single decision, with one decider and one recommender. RACI describes who does the work, and RAPID describes who decides.
When should you use DACI instead of RACI?
Use DACI when one decision needs a single Approver and a Driver who runs the process and recommends. Use RACI for the tasks in a workstream. DACI-style roles usually suffice for a team's recurring decisions, and RAPID suits decisions where legal, finance or risk must confirm a requirement.
How do you clarify who makes decisions in a team?
List the decisions from last quarter's minutes and email. Have each member mark separately who decided each one, then compare with the leader's marks. In a ninety-minute session assign a recommender, decider, consulted and informed parties and a limit to each, and review at day 30.
Why do teams keep revisiting decisions that were already made?
Usually no rule tells the group how it knows a decision is closed. Kaner describes members who act on a decision that others believe is still open. Name the decider, agree the closing rule beforehand, and count decisions reopened without new information.
How do you get a leader to delegate decisions?
Offer a staged version: level three, where the team recommends and the leader decides, for one quarter, with a count of what moved and a review date. Ask which decisions the leader would gladly see taken without them, and which one they would never hand over.
Keep reading
Measuring HR impact
Internal Mobility Metrics Without A Platform
Seven measures an HR adviser can build from existing records, each with a stated definition, a base and a way to read it.
7 min read
Workforce and future skills
Job Rotation As A Talent Gap Strategy
A job rotation programme closes a talent gap when the role, the person, the cover and the evidence are agreed with both managers before it starts.
8 min read
Talent reviews and succession
Seeing Talent Below The Top Team
Visibility of leadership potential drops below the top team. A cross-unit forum, run on one page per unit and a written evidence standard, is the method for seeing it.
8 min read