How to Decide Who to Promote (Without Second-Guessing at 11pm)

Two candidates. One slot. Both technically qualified.

You’ve been staring at their performance reviews for an hour, and you still can’t make the call. Your gut keeps tipping toward one of them — but your gut has been wrong before, and someone in the calibration meeting tomorrow is going to ask you to defend it.

Every article about promotion is written for the person trying to get picked. Almost nobody writes for you — the manager who actually has to decide who to promote, document the reasoning, and live with the call for the next decade of someone’s career.

This is the framework I built as a VP of Operations, and the one I now teach the women I coach who are tired of second-guessing themselves at 11pm. Let me walk you through it.

Why Most Promotion Decisions Quietly Go Wrong

The wrong person gets the role in three predictable ways — and they’re the same manager promotion decision mistakes I see repeated across every company I’ve worked with and every woman I’ve coached.

The visible candidate wins. The one who talks most in meetings, who copies you on everything, whose work is most legible to you. Not because they’re best — because they’re easiest to see.

The comfortable candidate wins. The one most like you. Same communication style, similar background, the person you’ve already imagined succeeding because the imagining requires no stretch.

The overdue candidate wins. The one you feel quietly guilty about. They’ve been waiting long enough that fairness feels like it tips toward them, regardless of whether the next role actually fits.

None of these are character flaws. They’re predictable cognitive shortcuts under time pressure — and you’re under more pressure than the decision deserves. Gallup has found that 42% of employee turnover is preventable and tied directly to the quality of management decisions. You’re making a 10-year call in the time most people would take to pick a hotel. Yet most companies still don’t give managers real promotion criteria — just a form and a deadline.

For women managers, the margin is even thinner. Stanford’s analysis of performance reviews found women were 22% more likely to receive personality-based feedback than men, and a 2025 People Management survey found 75% of women were labeled “emotional” in their reviews. The same assertive behavior that earns men a “confident” label gets women labeled “abrasive” or “aggressive.” If you’re a woman evaluating other women, you’re working from data that’s already distorted before you start scoring. If you’re evaluating men too, you have to interrupt your own pattern recognition — because that pattern was trained on biased reviews. If you want to fix the data problem at the source, writing performance reviews that are specific enough to be useful and evidence-based enough to be defensible is the prerequisite — because every promotion decision is only as good as the reviews feeding into it.

Promote the wrong person and you can lose two people. ADP’s analysis of 1.2 million employee records found 29% of newly promoted employees left their employer within one month — and for first and third-level managers, promotions nearly doubled the risk of attrition. So when the call goes wrong, you don’t just lose the runner-up. You lose the person you promoted, too.

Gut isn’t the enemy here. Unstructured gut is the enemy.

The Shift: Promote for the Next Job, Not the Last One

The single biggest mistake managers make is treating promotion as a reward for past work. Past performance gets the person on the shortlist. It shouldn’t be what decides between two people on it — but without a structured way of deciding between team members for promotion, that’s exactly what happens.

Promotion is a hiring decision for the next role. You wouldn’t hire someone externally based purely on what they did at their last company — you’d assess fit for what they need to do next. Same principle applies internally, and almost nobody applies it.

The research is brutal on this point. A study of 214 firms and 1,500-plus promotions in the Quarterly Journal of Economics found that doubling pre-promotion sales performance was associated with a 7.5% decline in the performance of that employee’s subordinates after promotion. The better someone was at the current job, the worse their team did after they were promoted — because the skills that make a top individual contributor are not the skills that make an effective manager. The same researchers found employees with collaboration experience produced better managerial results than top solo performers, even when the solo performers had stronger numbers.

What changes between levels isn’t “more of the same.” An IC promoted to manager stops being measured on individual output and starts being measured on team output. A manager promoted to director stops being measured on team output and starts being measured on cross-team leverage. Each transition is a different job — not a more advanced version of the current one.

This single reframe collapses the visibility bias problem. The loudest IC is not necessarily the best manager. The one who closes the most deals may be exactly the wrong person to lead the team that closes the deals. Your job is to evaluate evidence of the next job, not excellence at the current one.

If you’re reading this from the candidate side instead — the one who didn’t get picked — the playbook for what to do in the next 48 hours is over there. This one is for the manager making the call.

So you’re evaluating for the next role. But how do you assess what someone hasn’t done yet?

The Four-Lens Framework for Deciding Who to Promote

Four lenses, applied to every candidate, scored on observable evidence only. If you can’t cite a specific instance, the score is zero — not “I think so.” The framework’s whole job is to force the conversation about the decision to be honest. Vague impressions don’t get a score.

Lens 1 — Scope Stretch

Has this person already been operating one step above their current role on at least two or three occasions? You’re looking for evidence they’ve already done parts of the next job before you formally give it to them.

Concrete evidence looks like:

• A project where they were the de facto lead, not the named one
• A customer escalation they owned end-to-end without escalating back up
• A peer they coached without being asked
• A meeting they ran when you were out, and the team still made the decision

If the answer is “they would do well at it” instead of “here are three things they’ve already done at that level,” score it low. Potential isn’t evidence. A track record at the next level — even informally — is.

Lens 2 — Judgment Under Ambiguity

When the path wasn’t clear, what did they actually do? This lens separates people who can be told what to do from people who can decide what should be done.

The signal you want: a defensible call made with incomplete information, followed by adjustment when it turned out to be partially wrong. You’re not looking for people who were right. You’re looking for people who reasoned out loud, picked a direction, and corrected when reality pushed back.

The signals to discount: people who waited for instructions when they should have moved. People who barreled through when feedback was telling them to stop. Both are failure modes of judgment, in opposite directions.

Lens 3 — Influence Without Authority

Can they get things done through people who don’t report to them? This is the single biggest predictor of who succeeds at the next level — and it’s the lens most managers underweight because it’s the hardest to see in the org chart.

Evidence to gather:

• Who consults them outside their team?
• Whose minds have they changed in the last six months?
• Where have they negotiated a trade-off between two competing priorities without going up the chain to break the tie?

If you want to evaluate this lens more rigorously — or develop it in yourself — influence without authority is its own discipline. The Cohen and Bradford model, Wharton’s executive education research, and the Center for Creative Leadership all converge on the same finding: cross-team influence is the strongest predictor of senior-level effectiveness.

Lens 4 — Self-Awareness

Do they know what they’re bad at? A candidate who can name two specific weaknesses and what they’re doing about them is dramatically lower-risk than the candidate with the better resume but no admitted gaps.

The research from organizational psychologist Tasha Eurich is stark: only 10–15% of people are truly self-aware, despite 95% believing they are. So a candidate who can articulate their gaps is already in the top sliver of the workforce. Korn Ferry’s research has further connected leader self-awareness to actual financial performance — this isn’t a soft-skill checkbox, it’s a predictor of outcomes.

Score Lens 4 on the specificity of the answer, not its existence. “I should delegate more” scores a 2. “I default to doing the work myself when I’m anxious about quality, and I’m fixing it by handing off the next three release-launch checklists to my senior IC” scores a 5.

The Scoring Grid

Each lens scores 1–5 based on cited evidence. Total below 14/20 means the person isn’t ready, regardless of how much you like them. The number isn’t the decision — it’s what makes the conversation about the decision honest.

Fill this out for each finalist. Do it before you walk into calibration. The version you’d scratch out at 11pm the night before is not the version that survives someone else asking “why her?”

You have a framework now. But what about the bias you won’t see in yourself when you’re scoring?

The Bias Check: Five Questions to Ask Before You Finalize

Run these after you’ve scored, before you’ve decided. They’re not philosophical questions — they’re interrupts, designed to catch the parts of bias that operate before you’re aware of them. If you’ve ever wondered how to evaluate employees for promotion without bias creeping in, this is the checkpoint.

Q1: Have I given my top candidate’s evidence the same scrutiny I gave the others?

This is the confirmation bias check. The cognitive psychology research is consistent — we examine the candidates we doubt more rigorously than the candidates we like. The person you prefer gets a lighter vetting by default. Go back to your top pick and apply the scrutiny you applied to the candidate you almost ruled out. If the case still holds, you’re good. If it cracks, you weren’t actually evaluating them — you were defending them.

Q2: If I swapped names and demographics on the scoring sheet, would my ranking change?

This is the affinity bias check. Decades of resume-callback research have documented persistent gaps favoring traditionally male and majority-group names, and interviewers form decisions within the first 30 seconds of meeting a candidate. You’ve formed an impression before you’ve started evaluating. The swap test forces you to re-examine whether your ranking is about evidence or identity. Be brutally honest. No one’s watching.

Q3: Am I penalizing one candidate for the same thing I’m tolerating in another?

This is the gendered-language check. Stanford’s analysis of performance reviews found the same assertive behavior was described as “confident” and “decisive” in men’s reviews — and as “abrasive” or “aggressive” in women’s. A Fortune analysis of 25,000 review documents corroborated the pattern. Look at your own notes. If the candidate you ranked lower has any version of “abrasive,” “intense,” “too direct,” or “needs to soften their style” — and the candidate you ranked higher has the same behavior described as “decisive” — you have a problem. The behavior didn’t change. The label did.

Q4: What’s the strongest case AGAINST my preferred pick?

If you can’t make that case out loud for sixty seconds, you haven’t actually evaluated them. You’ve decided. Make the case. Make it as if you were the candidate who didn’t get picked, sitting across from you, asking why not. If the case is strong and your scoring still holds, your decision is defensible. If the case lands and your scoring suddenly feels thin, your scoring was a story, not an evaluation.

Q5: Would I make the same decision if I had to defend it publicly to the team?

Not in the calibration meeting, where everyone’s incentives line up to back each other. To the people who weren’t picked. To the team that has to live with this manager for the next two years. If the answer is “I’d hedge,” your evidence isn’t there yet.

A note specifically for women managers: bias doesn’t only distort who you under-pick. It also distorts who you over-pick. University of Michigan Ford School research on the double bind has documented how women leaders can swing toward overweighting women candidates as a correction — which backfires when those women aren’t ready, because the failure then gets coded as proof the original bias was right. The questions above cut both ways.

You’ve got the method and the bias check. But what about the call you make six months from now, when someone asks why this person and not that one?

The Documentation That Saves You Later

Most managers don’t document the decision until they’re forced to explain it — by which point the reasoning has fuzzed and what’s left is a vibe.

Write a one-pager the same day as the decision. This is what turns your evaluation into a fair promotion decision framework you can reuse — not a one-off judgment call. Four bullets, nothing more:

1. Who got the role.
2. The cited evidence per lens — pull it straight from the scoring grid.
3. The strongest case against the pick — the one you made for yourself in Q4.
4. What the runner-up needs to do to be ready next cycle — specific, observable, time-bound.

That last bullet is the most important one, and it’s the one most managers skip. Tell the candidate who didn’t get the role within 48 hours, and give them the runner-up doc verbatim. Not the scoring on the winner — the development path for them. Gallup’s most recent global workplace data has engagement at just 20% worldwide, with managers as the primary driver of whether employees stay or leave. In that environment, how you handle the “no” determines whether your runner-up becomes your next star or your next resignation.

If you want the script for that conversation — and a way to deliver the “no” that doesn’t burn the relationship — I’ve covered it in detail in how to give feedback that people actually hear. It’s the next move after the one-pager.

Keep the one-pagers in a personal folder. Over three or four cycles, they become your pattern-recognition library. You’ll start seeing your own tendencies — which lenses you over-weight, which kinds of candidates you consistently underrate — and that’s how managers actually get better at this. Not by reading articles. By reviewing their own calls.

One more thing: if you ever face a “why her, not me” conversation (and you will), the one-pager is the difference between defensible and panicked. You don’t have to remember. You wrote it down on the day you knew.

You have the method, the check, and the documentation. But what about when the call is genuinely 50/50?

When It’s Genuinely a Tie — And What to Do With Your Gut Now

Back to where we started. Gut wasn’t the enemy. Unstructured gut was.

When two candidates score within one point of each other across all four lenses and pass the bias check, your gut earns its seat at the table. It’s pattern recognition built from years of seeing who actually succeeds and who flames out at the next level. That’s real expertise — Gary Klein’s recognition-primed decision research is built on exactly this kind of expert pattern matching. But the expertise has to come after the framework, not before it. Lead with the gut and you’re just running the visibility, comfort, and overdue heuristics on autopilot.

The tiebreaker question I use, and the one I teach the women I coach: which of these two is the bet I can defend to myself in six months if it doesn’t work out? Not which one I want to work out — which one I’ll still be proud of having chosen on the bad day. The framework can’t make you infallible. It can make you defensible, repeatable, and honest with yourself. That’s the bar — and it’s the answer to how to decide who to promote that you can actually trust. That’s what makes you the manager people actually want to be evaluated by.

If this is the kind of thing you wish someone had handed you earlier, I write one of these every Sunday — the practical playbook for women managers who’d rather make the call than second-guess it at 11pm. No fluff, no motivation-speak. Just the tools I built as a VP that I now hand to the women I coach.