Insight · Practice
Decisions in a Leadership Tandem: Act Without Permanent Consensus
A leadership tandem does not need two yes-votes for every question. It needs a shared logic for who decides, who is consulted, when to escalate and when a genuine veto applies.
In short:
A leadership tandem does not need two yes-votes for every question. It needs a shared logic for who decides, who is consulted, when to escalate and when a genuine veto applies.
Which decisions really need to be joint — and how can the tandem remain able to act when the two leaders disagree?
- Not every decision belongs to both leaders
- Use a repertoire of decision modes instead of permanent consensus
- Treat deadlock as a design problem, not a character judgement
- The decision architecture must make sense to other people
- Decision rules must remain learnable
- Separate research evidence from design practice
Decision logic
A practical way to structure the question
Define the decision space
Decide which issues genuinely belong to shared accountability.
Explore →
Choose the mode
Autonomous, consultative, joint, delegated or emergency.
Explore →
Design deadlock rules
Agree the next move before conflict occurs.
Explore →
Clarify the interface
The team must understand the decision logic.
Explore →
Review
Test rules against real work.
Explore →
01 · DECISION SPACE
Not every decision belongs to both leaders
The practical problem is often not too little coordination but an unclear expectation about when coordination is required. If every operational issue becomes a joint decision, equality turns into a bottleneck. If almost everything is decided independently, the shared mandate becomes little more than a title. The aim is therefore not maximum togetherness, but the right form of involvement for different kinds of decisions.
A useful starting point is to distinguish reversible from hard-to-reverse choices, local from systemic risk, and decisions where one leader has clearly stronger proximity or expertise from those that alter the shared mandate. These criteria are not universal laws. Their value is that they force an implicit decision logic into the open.
For the team, predictability matters more than the elegance of a matrix: who may decide, when must the other leader be consulted, which issues require a joint commitment, and what happens when time pressure makes extended alignment impossible?
Take ten real decisions from the last four weeks. Classify them not by who happened to decide, but by which decision mode should apply next time.
02 · MODES
Use a repertoire of decision modes instead of permanent consensus
A tandem needs more than a binary choice between ‘alone’ and ‘together’. In practice, five modes are useful: decide autonomously and inform; consult before deciding; decide jointly; delegate to a person or role; and use a pre-agreed emergency rule. The terminology matters less than both leaders sharing the same expectation about involvement and accountability.
Consensus can make sense for identity-defining, strategic or difficult-to-reverse choices. It should not become the default for everything. Otherwise equality is confused with simultaneity. One person can lead a domain without dissolving shared accountability for the whole.
It is especially important to separate disagreement from veto. Disagreement must be possible without damaging the relationship. A veto, by contrast, should be tied to a small number of pre-defined red lines. Otherwise it becomes an informal reserve of power that can block almost any decision.
| Mode | Useful when | Expectation |
|---|---|---|
| Autonomous | reversible, clear domain | decide + inform |
| Consultative | the other leader has relevant expertise | hear input, lead decides |
| Joint | high impact / shared mandate | both own the decision |
| Emergency | time critical | pre-defined lead |
03 · DEADLOCK
Treat deadlock as a design problem, not a character judgement
When both leaders reach different conclusions after serious examination, the system needs a next move. Without a rule, task disagreement easily becomes relationship tension: giving way can feel like losing status, escalation can feel disloyal, and postponement merely moves the problem. A deadlock rule removes some of that personal meaning.
Depending on context, the rule might be a pre-defined lead domain, a time-boxed experiment, external expertise, a sponsor decision or a formal escalation path. What matters is that the rule was legitimate before the specific conflict. Writing a new constitution in the middle of a dispute is much harder.
After the decision, ‘one voice’ is not a ban on independent thought. The leader whose proposal was not chosen may document concerns and risks. What they do not do is undermine a legitimately reached decision through the team, stakeholders or informal side channels.
- Which decisions must never remain unresolved?
- Who decides under time pressure?
- Which red lines justify a genuine veto?
- How is dissent recorded and reviewed later?
04 · INTERFACE
The decision architecture must make sense to other people
Many failures occur not between the two leaders but at the interface with the organisation. Team members ask the second leader after receiving an unwelcome answer. Stakeholders search for the ‘real’ decision maker. A sponsor addresses only the more familiar person. This may begin as uncertainty, but over time it can erode the architecture.
The tandem should therefore communicate its decision logic in simple terms. The team does not need to memorise every exception. It should understand which domains have a clear lead, which issues belong to both leaders and how representation works. Different answers are not automatically a problem; contradictory authority is.
Visibility is part of the same design. If the same person always announces important decisions, an informal hierarchy can emerge despite formal equality. For major joint decisions, the tandem should deliberately show how both leaders carry responsibility.
05 · REVIEW
Decision rules must remain learnable
A decision map is not a one-off workshop output. Roles change, expertise develops, new stakeholders appear and some rules prove too slow or too broad. The review question is therefore not simply ‘Did we follow the matrix?’ but ‘Which decisions became unnecessarily slow, duplicated or contradictory?’
Useful signals include recurring postponement, duplicate approvals, surprising unilateral moves, repeated renegotiation or a team that systematically approaches only one leader. None of these signals proves that the tandem has failed. They show where a rule may no longer fit the real work.
A strong architecture does not remove all conflict. It makes conflict more workable because it becomes clearer whether the disagreement is about the substance, authority, risk or the decision rule itself.
Review the decision map after roughly 30, 60 and 90 days, then at a rhythm that matches how quickly the role and context change.
06 · EVIDENCE
Separate research evidence from design practice
Research on shared leadership treats leadership as a distributed influence process and reports, on average, positive relationships with team effectiveness. That matters because shared leadership is not created by two titles alone: leadership contributions have to be coordinated and usable by the team. Those findings do not, however, establish one universal decision matrix for formal leadership tandems.
The distinction between decision modes, deadlock rules and review rhythms on this page should therefore be read as design practice. It translates the need for coordination, clarity and shared accountability into agreements that can be tested. Research provides orientation; the actual architecture has to prove itself in context.
Research indicates: shared leadership can relate positively to team effectiveness. We infer: distributed leadership contributions require coordination. We recommend: making that coordination explicit through decision modes and review rules.
- Define the decision space: Decide which issues genuinely belong to shared accountability.
- Choose the mode: Autonomous, consultative, joint, delegated or emergency.
- Design deadlock rules: Agree the next move before conflict occurs.
- Clarify the interface: The team must understand the decision logic.
- Review: Test rules against real work.
FAQ
Frequently asked questions
Answers to the most common questions about this topic.
Which decisions genuinely need to be joint?
Usually those that materially affect the shared mandate, hard-to-reverse risks or both lead domains. Reversible operational decisions can often sit with a clear lead.
When is a veto appropriate?
Only for a small number of pre-defined red lines such as serious legal, ethical or existential risk. A broad veto can become a permanent blocking mechanism.
How can a tandem resolve deadlock without abandoning equality?
Use a pre-legitimised rule: lead domain, time-boxed experiment, external expertise or escalation. Equality does not require identical decision power in every situation.
Understand first. Then design.
Discuss your decision architecture
If decisions repeatedly stall or authority remains unclear, a structured review of decision modes and deadlock rules may be the most useful next step.
Next step
Discuss your decision architecture
If decisions repeatedly stall or authority remains unclear, a structured review of decision modes and deadlock rules may be the most useful next step.