The metric says the launch is working. The customer calls say something else.
The metric is very calm about this. It has been promoted to the meeting.
The sample excludes the people who abandoned the flow. They are not in the room, but they are in the denominator problem.
So the HiPPO is not a person today. It is the metric, the deadline, and whoever got to define success first.
I will add the missing customer outcome before we call this a win.
The classic product warning is easy to picture: the highest-paid person shares an opinion, everyone’s evidence develops stage fright, and the roadmap changes before lunch.
Hence the HiPPO.
It is a useful caricature because seniority does distort decisions. But caricatures become dangerous when they teach teams to look for one obvious villain. Authority can enter through a title, and it can also enter through a metric definition, a deadline, a research sample, a customer escalation or the person who owns the presentation.
Sometimes the HiPPO is a spreadsheet nobody feels permitted to question.
Opinion is not the enemy
Product decisions require judgment. Evidence is always incomplete, futures are not observable and somebody eventually has to choose.
A senior leader may also know something the room does not: a financing constraint, partner risk, legal exposure or strategic commitment. Pretending every decision can be settled by an experiment is its own form of theatre.
The problem is not that an influential person has an opinion. The problem is when the source, assumptions and decision rights remain hidden.
“We are doing this because the CEO wants it” is crude but legible. “The data says so” can be harder to challenge even when the data reflects a narrower choice made by someone else.
The quiet forms of authority
Watch for these less photogenic HiPPOs:
- The dashboard owner: decides which behaviour appears real.
- The deadline: makes one solution feel inevitable before alternatives are discussed.
- The loud customer: arrives through an escalation and temporarily impersonates the market.
- The research gatekeeper: controls whose voices enter the synthesis.
- The sunk-cost plan: has existed so long that changing it feels irresponsible.
None of these is automatically wrong. Each has disproportionate influence that should be made visible.
Experiments do not abolish judgment
Large-scale experimentation has helped product organisations challenge confident opinions with observed behaviour. Work shared by practitioners from companies including Microsoft, Google, LinkedIn and Airbnb shows how controlled experiments can improve decisions at scale.
But an experiment still needs a hypothesis, population, metric, duration and interpretation. It can estimate what happened under particular conditions. It cannot decide whether the measured outcome is ethically acceptable, strategically important or worth the trade-off.
“Data wins” is not a decision process. It is a slogan that hides the choices made before and after measurement.
Make authority explicit
Before a consequential decision, write down four roles:
- Who provides evidence?
- Who recommends?
- Who decides?
- Who must live with or operate the consequence?
They may be different people. That is normal. Confusion appears when the room believes it is collaborating while one person believes everyone else is advising.
Then ask the decision-maker to state what would change their mind. If no plausible evidence could do so, stop calling the activity discovery. Discuss the constraint or commitment directly.
Experiments can challenge authority, not replace it
The online experimentation summit paper brings together experience from thirteen organisations. Controlled experiments can make it harder for rank alone to settle empirical questions, because predicted behaviour can be compared with observed outcomes.
Yet experimentation creates new authority structures. Someone selects the metric, chooses the population, sets stopping rules and decides whether a statistically detectable effect is strategically meaningful. Mature experimentation makes those choices reviewable. Immature experimentation hides them behind the phrase “the test won.”
Participation is not decision rights
Intercom’s account of introducing a decision-making framework distinguishes levels of input and involvement. That distinction helps teams avoid fake consensus. Not every decision needs everyone to agree, but people should know whether they are being informed, consulted or invited to decide.
Clarity also improves dissent. A person can document a concern without believing one more argument will secretly change who owns the call. The decision-maker can acknowledge the trade-off and remain accountable for the result.
Record the decision before memory edits it
For consequential calls, write the owner, evidence, assumptions, dissent, expected outcome and review date. When results arrive, compare them with the original record.
Without this, organisations reward confident retrospective stories. The person with the most authority gets to explain that the outcome was expected all along. A decision log gives learning a chance against status.
Power becomes less dangerous when it leaves a trace.
Questions to take back to your team
- Who truly decides the product question currently described as collaborative?
- Which metric, deadline, customer or research sample has more authority than the room admits?
- What evidence could realistically change the decision-maker’s mind?
- Who chose what counts as evidence, and who is missing from that choice?
- When the result arrives, can you compare it with a written prediction and recorded dissent?
Think about the last decision introduced with “the data says.” Who selected the question, metric and interpretation? Was data informing authority, or simply lending it a lab coat?
My take
Do not try to remove power from product decisions. You cannot. Decisions are exercises of power because they allocate time, money and customer consequences.
Make power inspectable.
Name when a leader is making a call beyond the available evidence. Name when a metric encodes a value judgment. Name when a deadline is choosing on the team’s behalf. Record the dissent and the expected outcome so the organisation can learn rather than rewrite history.
The loudest person may be the HiPPO. The most dangerous HiPPO is the one the room has mistaken for gravity.