Across tech, about 14% of employees get promoted in a given year. That number comes from Pave’s analysis of roughly 245,000 people across more than 1,000 companies. In high-growth companies it climbs to 18.3%; in companies with shrinking headcount it drops to 11.5%. Whichever bucket you sit in, the same quiet fact holds: your promotion is not decided by your manager, and it is not decided in a room you are in.
It is decided in a calibration meeting. Eight to twelve managers, two or three hours, a director or VP in the corner, and a fixed number of slots that has to cover everyone those managers collectively want to move up. Your manager walks in with a case for you and a case for two or three other people. So does everyone else. Then they argue.
I have been in that room. Twenty-plus years running IT operations and, more recently, fractional COO work meant sitting on the deciding side of talent reviews more times than I can count. What I learned there changed how I coach anyone chasing the next level, product managers included. The people who get promoted are rarely the ones doing the best work. They are the ones whose work is easiest to defend when they are not there to defend it.
The person arguing for you is outnumbered and out of time
Start with the arithmetic, because it is unforgiving. If there are ten managers in the room and each has three names, that is thirty people being discussed. Nobody gets a leisurely hearing. The working number I have seen quoted, and the one that matches my own experience, is roughly 90 seconds of real advocacy per candidate before the pushback starts.
Ninety seconds. That is the entire window in which your two years of work gets converted into a decision by people who, in most cases, have never watched you run a single meeting.
And your manager is not a neutral narrator in that window. They are competing. Every slot they win for you is a slot they did not win for their other report. Harvard Business Review research on calibration found that speaking order and sheer verbal persuasiveness measurably affect who comes out ahead. If your manager burns their best, most articulate pitch on the person discussed before you, yours arrives to a tired room. That is not fairness. That is how a time-boxed meeting with limited slots actually behaves, and pretending otherwise is how good PMs get blindsided in March.
“We couldn’t see the impact”
Here is the failure mode I watched play out more than any other. A manager makes an earnest case. The candidate is clearly good. The room nods along. And then someone who does not manage that person asks a simple question: what did they actually do that a strong performer at the current level wouldn’t have done? The manager reaches for specifics and comes up with adjectives. Strategic. Trusted. High-judgment. The energy drains out of the room, and the name quietly moves to “next cycle.”
Michael Lynch, a former Google engineer, has written publicly about being rejected for promotion despite a top-5% performance rating and his manager’s support. The committee’s stated reason: they could not see the impact. That phrase should be tattooed on the inside of every ambitious PM’s eyelids. A top rating and an advocate in the room were not enough, because the impact was legible to his manager and to nobody else.
Product managers are especially exposed here, more than engineers. An engineer can sometimes point to a system that exists because of them. A PM’s contribution is diffuse by design. You aligned three teams. You killed a doomed initiative before it shipped. You reframed a problem so the org stopped arguing. Real work, all of it, and almost none of it leaves fingerprints that a stranger in a calibration room can find on their own.
Make your work legible to people who don’t manage you
The single most useful reframe I can give you is this. Stop optimizing for your manager’s opinion of you. Your manager is already sold. Start optimizing for the ability of a peer manager, one who barely knows you, to repeat your case in a sentence and have it survive a challenge.
That is a different project than doing good work, and most PMs never realize the two are separate.
A few things that actually move the needle in the room, drawn from what survived challenges and what died:
Impact that reaches beyond your manager’s org. If every strong example you have involves people who already report up through your manager, the committee reads it as current-level work done at higher volume. The senior signal they are hunting for is evidence you walked into an ambiguous space with no clear direction, defined the right problem, and set a framework that other teams, teams your manager does not control, actually adopted. Lenny Rachitsky’s own jump from mid-level to senior PM at Airbnb rode on a mission-critical project that spanned departments and reported to the executive staff. The scope was the case.
A number a stranger can hold. “Improved the onboarding experience” dies in calibration. “Cut activation time from eleven days to four, which the growth team now cites as the reason Q3 signups converted at nearly double the prior rate” survives, because the person repeating it does not need to trust you. They need to remember one figure.
Named witnesses outside your reporting line. When a director who does not manage you says, unprompted, “the reason that launch didn’t blow up was the PM,” your case is no longer your manager’s word against a skeptical room. It is a pattern other leaders have noticed. This is why building genuine sponsors across the org does more for your trajectory than any amount of polish on your self-review. Sponsors are the people whose secondhand endorsement your manager can quote.
A written trail your manager can lift. The promotion packet that works reads like a legal brief, not a performance review: a claim, then a specific piece of evidence, for each dimension of the next level. If you hand your manager that document, you have effectively scripted their 90 seconds. If you make them assemble it from memory, you are trusting recall under time pressure, and recall under time pressure is where careers stall.
Start a year before you think you need to
Google runs what it calls lagging promotions: you are expected to demonstrate next-level work for roughly six months before the promotion is granted, and the industry benchmark for time-in-level tends to land somewhere between eighteen and twenty-four months. The practical consequence is that the evidence deciding your next promotion is being created, or not created, right now, long before the meeting.
This is the part I wish someone had told the PMs I coached years earlier than they heard it. The calibration room does not evaluate your potential. It evaluates a paper trail. If you start assembling that trail the quarter before review season, you are documenting whatever happened to be in front of you, which is rarely the cross-org, ambiguous, direction-setting work the committee is actually looking for. If you start a year out, you can choose the work. You can volunteer for the messy initiative that spans three teams precisely because it will be legible to strangers later.
I built a weekly impact log into my own routine and pushed it on everyone I mentored, for exactly this reason. Not because reflection is nice, though it is, but because the person arguing for you in that room can only use what has been written down.
The uncomfortable version of the truth
None of this means the work does not matter. It means the work is necessary and nowhere near sufficient. A promotion is two separate achievements stapled together: doing the job at the next level, and making that fact provable to people who never saw you do it. PMs are trained relentlessly on the first and left to discover the second by getting passed over.
If you want the honest test of your own readiness, it is not “am I operating at the next level.” It is this. Sit down and imagine a manager who barely knows you, standing in a room full of peers with a fixed number of slots, defending your promotion in one sentence against someone determined to poke a hole in it. If you cannot write that sentence for them, and back it with a number and a witness, the room will not write it for you.
The good news is that the room is winnable. It just rewards a skill most PMs are never told they need. Not the work. The evidence of the work, in a form that travels without you.
For the mechanics of assembling that evidence, see how to build a promotion case the committee can actually use. And if you want your skip-level to walk into calibration already on your side, prep them deliberately rather than hoping proximity does the work.
Sources
- Lenny Rachitsky, “How to get promoted” and “Becoming a senior Product Manager”
- The Pragmatic Engineer, “Performance Calibrations at Tech Companies”
- CareerClimb, “Why Your Manager Can’t Promote You Alone” (Pave promotion-rate data; Michael Lynch example; HBR calibration research)