Most stalled product decisions are not hard decisions. They are decisions nobody was ever told they owned. Someone recommends, six people weigh in, everyone treats their opinion as a stop button, and the thing sits. The problem is almost never the quality of the thinking. It is that the room never agreed on what each person’s voice was allowed to do.
I learned this years before I worked in product, running network operations at a large telecom. We had a change-advisory board that reviewed maintenance windows. Fifteen people sat in that meeting. Every one of them believed that being invited meant they could veto the change. So a routine window that should have been approved in ten minutes turned into a negotiation, because a manager from a team that would not even be touched by the work decided he was uncomfortable and treated his discomfort as a blocking vote. Nobody had told him he only had input. Nobody had told anyone anything. We had confused being consulted with being in charge, and it cost us a full week on work that carried almost no risk.
That confusion is the single most common communication failure I see product managers make with stakeholders. It is not that they communicate too little. It is that they never separate the two roles that look identical from the outside and are completely different underneath: input and veto.
Input is not a veto, and most stakeholders assume it is
When you ask a stakeholder what they think, you are making an implicit promise. The question is which promise. “I want your perspective and I will weigh it” is one promise. “You can stop this if you disagree” is a completely different one. If you never say which, the stakeholder defaults to the stronger interpretation. People who are asked for their opinion in a high-stakes setting tend to assume their opinion is load-bearing. Why else would you have asked?
So the sales director who wanted a different feature order believes he was overruled rather than consulted. The senior engineer who flagged a concern believes she was ignored rather than heard. Both of them gave input. Neither of them held a veto. But because you never drew that line out loud, they experienced a normal decision as a personal loss. Now they are quietly less cooperative on the next one, and you have no idea why.
This is why “over-communicating” as generic advice does not fix stakeholder friction. Sending more updates to people who think they have a vote does not remove the confusion. It feeds it. The fix is not volume. It is role clarity.
Consensus feels safe and quietly wrecks the timeline
The instinct, when everyone thinks they have a vote, is to keep talking until everyone is happy. That instinct is expensive. A 2019 McKinsey survey of more than 1,200 managers and executives found that 61 percent said at least half the time they spent making decisions was used ineffectively, and only 37 percent said their organization’s decisions were both high quality and timely. For a typical Fortune 500 company, McKinsey estimated the waste at roughly 530,000 days of managers’ time a year, about 250 million dollars in salary.
The same research names the culprits directly: overreliance on consensus, death by committee, and unclear roles. And it punctures the assumption that slowing down buys you a better answer. Respondents who said their decisions were fast were 1.98 times more likely to also call those decisions high quality. Speed and quality were not a trade-off. They moved together, because the same thing that makes a decision fast, a clear owner, is what makes it good.
Chasing consensus is not the cautious choice. It is the choice that trades a defined week of debate for an undefined month of drift, and calls the drift diligence.
RAPID gives every voice a role and only one of them a decision
The most useful tool I have found for cutting this knot is old and boring, which is usually a good sign. In January 2006, Bain consultants Paul Rogers and Marcia Blenko published a Harvard Business Review article with a blunt title, Who Has the D?, and it remains the clearest thing written on this problem. Their framework, RAPID, assigns five distinct roles for any real decision:
- Recommend. One person drives the process, gathers the facts, and proposes the course of action. For most product decisions, that is you.
- Agree. A small set of people, sometimes nobody, who must formally sign off on narrow, predefined grounds. Legal on a compliance question. Security on a data question. This is the only real veto, and it is deliberately rare and scoped.
- Perform. The people who will execute once the call is made.
- Input. Everyone whose expertise or perspective genuinely shapes the recommendation. They are consulted. They do not decide, and they cannot block.
- Decide. One person who makes the call and commits the resources.
The value is not the acronym. It is the wall the framework builds between Agree and Input. That wall is exactly the one my old change-advisory board never had. Most of your stakeholders are Input. A handful, on specific dimensions, are Agree. Almost none of them are Decide. Once you name which is which, the sales director’s strong opinion stops functioning as a brake, because he now knows, and everyone in the room knows, that his role was to inform the recommendation, not to approve it.
Note the deliberate scarcity built into the model. If everyone can Agree, the veto is meaningless and you are back to consensus. RAPID works because it starves the veto down to the two or three people who actually carry accountability for that dimension.
Name the decider before the meeting, not during it
The mistake is trying to sort roles live, in the room, while the decision is on the table. By then it is too late; people have already anchored on the belief that showing up equals having a say. Assign the roles in writing beforehand, in one or two lines at the top of whatever document or invite you send.
It can be as plain as this: “Decision: which onboarding flow ships in Q4. I am recommending. [VP Product] decides. [Security lead] must agree on the data-handling piece. Everyone else on this thread is input, and I want it by Thursday.” Four sentences. Every person now knows the exact shape of their own voice before they open their mouth. The engineer who reads that knows her concern will be weighed, not obeyed, and she is fine with that, because being heard was what she actually wanted. The sales director knows he is not being handed a veto, so he does not spend the meeting acting like he has one.
This is also what makes disagree and commit actually work. You cannot ask people to commit to a decision they disagree with unless they understood, going in, that they were input and not the decider. Commitment is easy when the roles were honest. It is impossible when someone believed they held a vote they never actually had. The same clarity is what lets a room genuinely pressure-test a plan without the exercise curdling into a standoff, which is the whole point of running something like a pre-mortem before you commit.
Twenty-plus years of watching decisions stall taught me that the meeting is rarely where things go wrong. The rot sets in earlier, in the quiet gap where nobody said who had the D. Close that gap in two sentences before anyone gathers, and half your stakeholder friction disappears before the conversation even starts.