Who can say yes, and does everyone know it?
Why most prioritisation failures are ownership failures in disguise.
Your team has prioritisation criteria. It has a planning rhythm. It might even have a Conflict Wall and a shared evidence base. And priorities still shift without warning, mid-cycle, with no visible decision trail.
The criteria didn’t fail. The ownership underneath them did.
Most prioritisation breakdowns, when you trace them back far enough, land on the same structural gap: nobody can cleanly answer who has authority over what. Who can bring new work in? Who can approve it? Who can re-sequence existing commitments to make room? Who can accept an urgent request mid-cycle? And who can say “stop” or “not yet”?
In most teams, the answers to these five questions are different depending on who you ask. That gap is where prioritisation collapses. The question gets sharper once AI makes redundant work visible: who has the right to say “ditch it”?
Five questions, five gaps
We put five questions on a wall and asked each table for names, not roles. Not an org chart. Not a RACI. A one-page answer to how prioritisation decisions actually get made.
Who can request work? In most organisations, anyone can. A senior leader sends a message. A peer team raises a dependency. A campaign brief arrives. A stakeholder asks a favour. There is no filter on who can place demand into the system, which means the system absorbs everything that arrives and the team’s capacity becomes a shared resource with no gatekeeper.
Who can approve work? This sounds obvious until you ask three people on the same team and get three different answers. Is it the team lead? The brand manager? The chapter lead? The person who raised the request? In one session, the room discovered that the senior brand manager was the de facto approver for day-to-day work, but nobody had formally named that. Decisions were being made by convention rather than by design, which meant the convention broke every time someone more senior weighed in.
Who can re-sequence work? This is the question that reveals the real power dynamics. Re-sequencing means moving something down the list to make room for something else. In theory, the team lead does this. In practice, one leader pulls rank on another and the team’s priorities shift without the team being part of the conversation. The re-sequencing happens to them, not with them.
Who can accept urgent mid-cycle work? Every organisation has urgent requests. The question is not whether they arrive but who has authority to say yes, and what happens to the existing commitments when they do. If the answer is “whoever received the message,” then every urgent request enters the system without a trade-off, and the team’s agreed priorities erode one message at a time.
Who can stop or defer work? This is consistently the hardest question. Once something has started, teams find it remarkably difficult to stop it. Partly because stopping feels like wasting the effort already invested. Partly because nobody is confident they have authority to make that call. The result is that deprioritised work continues to consume effort long after the decision to deprioritise was supposedly made.
Why the gaps matter more than the criteria
You can have the sharpest prioritisation criteria in the organisation and they will still collapse if ownership is unclear.
Criteria tell you what matters. Ownership tells you who decides. Without ownership, criteria become advisory: the team knows what should happen, but nobody is empowered to make it happen. When a senior stakeholder sends a request that doesn’t meet the entry criteria, someone needs to have the authority and the standing to say “this needs clarification before it enters.” If nobody has that authority, the entry criteria are decoration.
The five questions map to three levels where ownership either holds or breaks.
Upstream, where brand demand meets the planning cycle, is where questions one and two live. Who can request? Who can approve? If those answers are unclear at this level, too many initiatives enter the system in parallel and the team absorbs everything because nobody has standing to filter.
At the interface between teams and the people who set priorities is where question three sits. Who can re-sequence? This is where one leader pulls rank on another and the team’s commitments shift without a visible trade-off. The re-sequencing happens to teams, not with them.
Within teams is where questions four and five live. Who can accept urgent mid-cycle work? Who can stop or defer? If these are unclear, agreed priorities erode one message at a time, and deprioritised work continues consuming effort because nobody is confident they can make the call to stop.
Fix the criteria without fixing ownership at all three levels and you have a system that knows what to prioritise but has no mechanism to enforce it.
What a working ownership map looks like
Not a RACI. RACIs are too abstract and too large to be useful in the moment a decision needs to be made. By the time you’ve found the right row and column, the decision has already been made by whoever was loudest.
A working ownership map is a single page with five rows and clear names. Not roles. Names. “The brand manager for this portfolio” is too vague. “Claire” is specific enough to act on.
The map gets built at table level first: each cross-functional group answers the five questions for their own working context. Then the room compares maps. That comparison is where the real value lives. One team has the brand manager as the approver. Another has the chapter lead. One team allows any senior stakeholder to accept urgent work. Another requires it to go through the team lead.
Those inconsistencies are not errors. They are the system’s actual operating rules made visible. Once visible, they can be discussed and agreed. Invisible, they produce the friction the room has been carrying for months.
The connection to entry criteria
Ownership and entry criteria are two sides of the same structure. Entry criteria define what must be true before work enters the system. Ownership defines who enforces those criteria.
Without ownership, entry criteria are a wish list. “Is there a named owner?” is a good entry criterion, but it only works if someone in the system has the standing to send a request back when the answer is no. “What gets displaced if this enters?” is a powerful question, but it only gets asked if someone has the authority to insist on an answer before the work is accepted.
Build them in the same session, with the same people, looking at the same work. They are one structure, not two.
What to do if your team can’t answer the five questions cleanly
Start with one table. One team. One working context. Put the five questions on a wall and ask for names, not roles.
If the room produces different answers to the same question, that is the diagnosis. The disagreement is the finding. Don’t smooth it over. Surface it, discuss it, and agree on one version.
Then test it. Take a real case from the last planning cycle where a priority shifted unexpectedly and run it against the ownership map. Would the map have caught it? Would the right person have been asked? Would the trade-off have been visible?
If the answer is no, adjust the map. If the answer is yes, you have built something the team can use starting this week.
Criteria tell you what to prioritise. Ownership tells you who protects the decision. Without both, you have a system that knows what to do but has no one empowered to do it.
This is the fifth in a series on the structures underneath good prioritisation. Previously: Prioritisation is a scaffolding problem, What your team knows about prioritisation, The wall where the hidden rules live, and The criterion you’re already using. Next: scaffolding is not an event, and the follow-up rhythm that holds decisions in place.