In 1977, psychologists Lee Ross, David Greene, and Pamela House ran a study at Stanford that product managers should know by heart. They asked students to wear a sandwich board reading “Eat at Joe’s” and walk around campus. Roughly half agreed, half refused. Then both groups estimated what fraction of peers would make the same choice. Both groups believed they were the majority.
Ross and his colleagues called this the false consensus effect: the tendency to overestimate how much other people share your views. Forty-nine years later, this bias runs unchecked in product meetings.
The nod that means nothing
Here is what happens in most product decision meetings. The PM presents a recommendation. A few people ask clarifying questions. Nobody raises a fundamental objection. The meeting ends. The PM writes “Decided: we will do X” in the notes.
Two weeks later, the engineering lead mentions in standup that she thought X was still under evaluation. The design lead says he assumed X was only for Q3 consideration, not Q2. The VP of sales, who sat in the meeting, says he never agreed; he just didn’t see the point of arguing at that stage.
The decision that felt made was never actually made. It evaporated somewhere between the conference room and the Slack channel.
I have watched this pattern repeat dozens of times running infrastructure and operations teams. A group sits together, nobody objects, and the leader walks out believing they have consensus. What they actually have is the absence of visible disagreement. Those are two completely different things.
Why silence fills the room
People stay quiet in product meetings for reasons that have nothing to do with agreement.
They are still processing. Not everyone forms opinions at meeting speed. PMI’s 2025 Pulse of the Profession found that organizations prioritizing interpersonal capabilities (collaborative leadership, communication, empathy) achieve 72% business goal success rates, compared to 65% for organizations that do not. Part of collaborative leadership is recognizing that the person who talks first is not necessarily the person who thought hardest.
They pick their battles. Stakeholders operate under political constraints that product managers sometimes underestimate. The VP of marketing might disagree with a feature direction but choose not to surface that disagreement in front of the CTO. The disagreement surfaces later, in a Slack DM, in a “quick aside” with the CEO, or in passive resistance during execution. The dynamics of conflicting stakeholder agendas make this even more common when multiple senior leaders have competing priorities.
They don’t realize a decision is being made. This one is the most common and the most preventable. If the PM does not explicitly say “I am asking this group to commit to Direction X today,” many people in the room will interpret the meeting as informational, not decisional.
The cost of the phantom decision
When undocumented, uncommitted decisions move to execution, the bill comes due fast. Research by Barry Boehm estimates that 30 to 50 percent of engineering effort goes to avoidable rework caused by misunderstood or misaligned requirements. On a 50-person engineering team at $150K average fully loaded cost, 30% waste translates to $2.25 million per year building things that nobody actually agreed to build.
A Gartner 2024 study found that 67% of B2B projects fail primarily due to poor stakeholder alignment, not technical or financial obstacles. The technical work is usually fine. The failure lives in the gap between what people thought they agreed on and what they actually committed to.
Three ways to close the gap
The fix is not more meetings. It is more precision inside the meetings you already have.
Name the decision before the meeting starts. Every meeting invitation should state whether this meeting is informational, exploratory, or decisional. If it is decisional, state the specific decision: “We will leave this meeting with a committed direction on X.” This single change eliminates the most common source of phantom decisions. Writing a structured decision brief before the meeting forces this clarity.
Poll the room, do not scan it. Instead of asking “Any concerns?” (which invites silence), go around the table: “I need each person to state whether they can commit to this direction or whether they have a blocking concern.” The difference is enormous. The first question treats silence as consent. The second treats silence as nothing, because each person has to actively speak.
Marty Cagan makes a related point about stakeholder relationships in Transformed: the most effective product teams make limited, high-integrity commitments and then deliver on them. You cannot make a high-integrity commitment to something your stakeholders never actually committed to.
Write the decision, not the discussion. Meeting notes that capture who said what are filing cabinets. Meeting notes that capture the decision, the alternatives considered, and who committed are decision records. The Architecture Decision Record (ADR) format that engineering teams use for technical choices works just as well for product decisions. The core fields: context, decision, alternatives considered, consequences, and the names of the people who agreed.
When I ran infrastructure teams, we adopted a simple rule: if it is not in the decision log, it was not decided. The initial resistance was predictable (“this is bureaucratic”). Within two months, the team stopped relitigating decisions that had already been made. That time savings alone justified the practice.
The conversation nobody wants to have
The hardest version of this problem is when a senior stakeholder stays quiet in a group setting and then reverses the decision in a private conversation with the CEO. Product managers treat this as politics. It is, partially. But it is also a process failure.
If someone in the room did not speak during the explicit commit round, the PM has standing to follow up directly: “You did not raise a concern in the meeting. Is there something you want to discuss before we move forward?” That follow-up is uncomfortable. It is also far less expensive than discovering the objection three sprints into execution.
The goal is not unanimous enthusiasm. Disagreement is healthy and expected. The goal is that everyone in the room leaves knowing exactly what was decided, who committed to it, and what they are expected to do next. No phantom consensus. No retroactive reinterpretation. No “I never agreed to that.”
Nodding heads are not decisions. Decisions are explicit, named, and documented. If yours are not, you are not managing stakeholders. you are hoping they agree.