Avinash Kaushik, then an analytics evangelist at Google, gave the problem its name in his 2007 book: the HiPPO, the Highest Paid Person’s Opinion. The idea stuck because every product manager has watched it happen. A room full of evidence, a team that spent three weeks on research, and then the VP says “I just don’t think customers want that,” and the decision is made. Not debated. Made. Bernard Marr described the pattern plainly in Forbes: we subconsciously hand highly paid people authority they have not earned on the specific question in front of us, because it feels natural to assume they are smarter, savvier, more strategic.
Here is the part most PMs get wrong about the HiPPO. They treat it as a character problem. The executive is arrogant, or dismissive, or does not respect the team. Sometimes that is true. Usually it is not. The HiPPO wins most rooms for a structural reason that has nothing to do with the person’s ego and everything to do with the order in which people talk. Fix the order, and a surprising amount of the problem dissolves. That makes this a communication design problem, which means it is squarely a product manager’s job.
The real mechanism is who goes first, not who earns most
Group decisions anchor on the first opinion voiced. A 2024 analysis in the journal Erkenntnis called anchoring a structural bias of deliberation: the people who speak early exert more pull on the outcome than the people who speak later, and across the various forces shaping a group’s conclusion, that early-speaker effect is often the strongest one. The first argument sets the frame. Everything after it is a reaction to that frame rather than an independent read of the evidence.
Now stack the two effects. The highest-paid person usually speaks first, because the room defers to them to open, or because they are chairing, or because nobody wants to stake out a position before the boss has. So the HiPPO is not just the loudest voice. The HiPPO is frequently the anchor voice. The seniority and the speaking order compound, and the team’s three weeks of research arrive after the frame is already set.
There is good evidence that letting a few voices dominate makes groups measurably worse at the actual task. In a 2010 Science study, Anita Woolley and colleagues found a general collective intelligence factor in groups, and it was not predicted by the average IQ of the members or the smartest person present. It tracked with social sensitivity and with the equality of conversational turn-taking. Groups where a few people dominated the conversation were less collectively intelligent than groups where speaking was spread around (the correlation was negative and significant, r = -0.41). Google reached a similar conclusion studying its own teams in Project Aristotle: the most effective teams had roughly equal conversational turn-taking, where over the course of a discussion everyone spoke about the same amount. A HiPPO-dominated room is the exact opposite of that shape.
The experiment that tells you what to change
If anchoring were destiny, you could stop here and just complain about executives. But one of the cleanest tests of herding in group judgment found that the effect can be designed out, and the design is something a PM controls.
Researchers ran a large randomized controlled trial inside the peer-review process of a major machine learning conference, covering 1,544 papers and 2,797 reviewers. They manipulated who kicked off each paper’s discussion, seeding some discussions with a reviewer who liked the paper and others with a reviewer who disliked it, to see whether the first voice would drag the group toward its verdict. The intervention worked on the thing it was supposed to move: it reliably changed the order in which reviewers entered the discussion. And then the outcomes did not budge. The final decisions converged toward the average of the reviewers’ positions regardless of who spoke first. The authors found no evidence of herding.
Why would this room resist anchoring when so many rooms do not? Because of one feature of the process: every reviewer had written down an independent opinion, with a score and a justification, before any discussion began. The anchor had nothing to grab. People were not forming their views in real time in the presence of a dominant first speaker. They walked in already committed to a position on the record.
That is the whole lesson, and it is not abstract. The difference between a room that herds and a room that does not is whether people formed their opinions before or during the conversation. You cannot change how senior your VP is. You can change whether opinions exist on paper before that VP opens their mouth.
What this cost me once, concretely
Running IT operations at a large telecom years ago, I sat on the change-approval board that signed off on maintenance windows: the overnight slots where we would take systems down to patch, migrate, or upgrade. One night’s plan touched a billing component, and the engineer who had built the rollback procedure flagged a real risk in the sequence. He was right, and he was two levels down from the most senior person in the room.
He never got the sentence out. The senior leader opened the meeting with “this one’s routine, let’s keep it moving,” and the frame was set before anyone with hands on the system had spoken. The board nodded. We ran the window. The rollback behaved exactly the way the junior engineer had worried it would, and we spent the next morning cleaning up a billing-data problem that one unhurried sentence would have prevented. The expertise was in the room. The speaking order buried it.
What I changed afterward was mundane and it worked. Before each board meeting, every system owner submitted a one-line risk call in writing: green, yellow, or red, with a reason for anything not green. We read the reds first, and the people closest to the work spoke before the people with the biggest titles. The senior leaders still had the final say. They just exercised it after hearing independent positions instead of replacing those positions with a frame nobody felt able to push against.
The practice: collect positions before you open the room
If you run meetings where a HiPPO keeps overriding the team’s work, stop trying to win the live debate. You will lose it, because the anchor already has. Move the decision earlier, into a stage the anchor does not dominate.
Make people commit in writing before the meeting. A short pre-read where each stakeholder records a position and a reason is the single highest-leverage move, because it is exactly the condition that neutralized herding in the peer-review study. Opinions formed in private do not anchor on the loudest voice. A one-page brief that forces each person to write down what they would decide and why is worth more than an hour of live discussion. This is the same reason a written decision brief outperforms a meeting for the calls that actually matter.
Invert the speaking order. The person closest to the work speaks first; the most senior person speaks last. This is deliberate and you can say so out loud. “I want to hear from the people who built this before we react to it” is a sentence no reasonable executive argues with, and it quietly removes the anchor’s opening move.
Separate the evidence from the recommendation. Present what the research found before anyone says what they want to do about it. When the data and the verdict arrive fused into one slide, the HiPPO reacts to the verdict. When the data stands alone first, the room reacts to the data.
Ask for the position, not the reaction. “What’s your read?” invites a fast gut take that hardens into an anchor. “Here’s the decision we need and the three options, which would you choose and why?” invites a considered position. Structure the question so the answer has to carry a reason with it.
None of this strips authority from the people who are supposed to have it. The executive still decides. The point is to make sure the thing they are deciding is the team’s actual best thinking, not a frame that got set in the first fifteen seconds and never got tested. A pre-read that surfaces real disagreement is also what makes disagree-and-commit function instead of becoming quiet compliance, and it is why running a structured pre-mortem before a launch pulls out risks that a live status meeting never would.
The HiPPO is not going anywhere. Hierarchy is real and senior people are paid to make calls. What you can change is whether the room hands them a fully formed judgment to rule on, or an empty frame they fill by accident. That is design, and it is yours to run.