How to Measure Manager Effectiveness: 12 Metrics That Predict Team Outcomes

Someone resigns in October. The disengagement started in July. By the time attrition shows up on a dashboard, the decision that caused it is 90 days old and the manager who could have changed it has moved on.

Manager effectiveness measurement is the practice of tracking a manager's observable behaviors alongside their team's outcomes, so you can predict team performance rather than report it after the fact. Behaviors are the leading half. Outcomes are the lagging half. Most companies measure only the second one, then wonder why their manager development spend never seems to land.

Below are 12 metrics: six that move early, six that confirm later. For each: what it is, how to collect it, what it predicts, and how it breaks when people play to the number.

Best for: companies with 10 or more managers who already run some form of engagement measurement and want to know which managers need support before the quarter ends. Fewer than 10 managers? Skip to "Where to start" near the end.

Why measuring management effectiveness is worth the effort

Gallup researchers Randall Beck and Jim Harter, analyzing engagement data from 27 million employees across more than 2.5 million work units, found that managers account for at least 70% of the variance in team engagement scores across business units. Note the word variance. It does not mean managers cause 70% of engagement. It means that when engagement differs between two teams in the same company, roughly 70% of that difference traces back to the manager. (We wrote a full explainer on what the 70% variance figure actually means, because the misquote is everywhere.)

That is the argument for measuring managers specifically rather than the company in aggregate. Company-level scores average away the exact signal you need. And most manager effectiveness evaluation still runs on annual cycles, which is roughly like checking the oil once a year and calling it maintenance.

Leading indicators: six metrics that move first

Leading indicators are manager behaviors. Observable, weekly, and they change before outcomes do. Treat them as the early-warning half of your manager effectiveness metrics.

1. Recognition give-rate

What it is: How often a manager recognizes someone on their team, measured as recognitions given per manager per week, plus the spread across the team.

How to collect it: Pull it from whatever platform recognition already flows through. In Happily.ai, recognition carries gems, a redeemable currency, so every act leaves a timestamped record with a giver, a receiver, and a reason tied to a company value.

What it predicts: Trust and discretionary effort. Happily.ai platform data across 10 million-plus interactions shows people who give recognition are trusted 9x more by colleagues than those who do not.

How it fails when gamed: A manager chasing a weekly number recognizes the same two people repeatedly, or writes "great work this week" with no specifics. Track spread (distinct team members recognized in 30 days) and specificity, not volume.

2. Feedback response time

What it is: The median number of days between a team member submitting feedback or raising a concern and the manager visibly responding to it.

How to collect it: Timestamp both events. Any feedback tool with a status field does this. Manual version: log the date a concern is raised in your 1:1 notes and the date it closes.

What it predicts: Whether people keep speaking up. Response time is the price signal on candor. Once it stretches past two weeks, submission volume drops, and you lose the input long before you lose the person.

How it fails when gamed: Managers close tickets fast without resolving anything. Pair timing with one follow-up question to the person who raised it, asked 30 days later: was this actually addressed?

[IN-ARTICLE IMAGE: A single horizontal line with six warm dots clustered on the left and six cool dots on the right, a vertical divider between them]

3. 1:1 consistency

What it is: The percentage of scheduled 1:1s actually held, over a rolling quarter. Not whether they exist on the calendar. Whether they happened.

How to collect it: Calendar data plus a completion check in your 1:1 tool. Count reschedules separately from cancellations. A meeting moved twice and then held is a different signal from one that vanished.

What it predicts: Engagement, more directly than almost any other behavior. Gallup found that employees whose managers hold regular meetings with them are almost three times as likely to be engaged as employees whose managers do not.

How it fails when gamed: Fifteen-minute status readouts logged as 1:1s. Consistency without substance moves the metric and nothing else. When 1:1 consistency is low across many managers at once, look at span of control before discipline. We covered the arithmetic in how many direct reports a manager should really have.

4. Check-in participation

What it is: The share of a manager's team that responds to regular pulse check-ins, tracked as a trend rather than a snapshot.

How to collect it: Automatically, from your check-in platform. Happily.ai runs daily check-ins and reports participation per team, which feeds DEBI, the Dynamic Engagement Behavior Index, a 0 to 100 team engagement score.

What it predicts: Silence is the cheapest early warning available. A team that stops answering has usually decided that answering changes nothing, and participation drops typically precede sentiment drops by several weeks.

How it fails when gamed: Managers pressure the team to respond, producing high participation and flattened, uniformly positive answers. Watch for participation and variance moving in opposite directions. Real participation has spread in it.

5. Team sentiment trend

What it is: The direction of a team's engagement score over eight weeks, not the absolute level.

How to collect it: Rolling pulse data, aggregated weekly. Direction is the metric. A team at 62 and climbing is in better shape than a team at 74 and sliding.

What it predicts: Near-term attrition risk and performance drag. Sustained decline over four or more weeks is the point where a conversation is still cheaper than a backfill.

How it fails when gamed: Managers time interventions right before measurement windows. Continuous measurement removes the window. Weekly data has no run-up.

6. Focus alignment

What it is: The percentage of a team's work that maps to a stated company priority, as reported by the team rather than by the manager.

How to collect it: Ask team members what they worked on and which priority it served, then aggregate by team. Gaps show up as work with no priority attached, or priorities with no work attached.

What it predicts: Goal completion, one to two quarters out. It also predicts frustration. Gallup's Q2 2025 data found only 47% of US employees strongly agree they know what is expected of them at work, and resolving that ambiguity lands on the manager.

How it fails when gamed: Everything gets mapped retroactively, which makes alignment look like 100% and means nothing. Sample the mappings. Perfect alignment every week means the categories are too loose.

Lagging indicators: six metrics that confirm

Lagging indicators are outcomes. Slower, harder to fake, and the reason the leading indicators matter. Use them to validate that your early signals are actually predicting something.

7. Regretted attrition

What it is: Voluntary departures of people you wanted to keep, as a rate per manager per year. Separate it from total attrition, which mixes in exits you were fine with.

How to collect it: Have the manager and one skip-level classify each voluntary exit as regretted or not, within a week of resignation and before the backfill conversation starts, so the answer is not shaped by hiring urgency.

What it predicts: It confirms rather than predicts. Gallup estimates the cost of replacing an individual employee at one-half to two times that employee's annual salary. Happily.ai customers see up to 40% turnover reduction, which is where roughly $480K per year in avoided replacement cost comes from at mid-size headcount.

How it fails when gamed: Managers reclassify regretted exits as non-regretted after the fact. Lock classification at the two-week mark and audit a sample quarterly.

8. Internal mobility

What it is: How many of a manager's people move into other roles in the company, whether laterally or upward, over a rolling 12 months.

How to collect it: HRIS transfer records, tagged to the sending manager. Count moves out, not moves in.

What it predicts: Retention across the whole company, not just the team. LinkedIn's 2022 Workplace Learning Report found companies excelling at internal mobility retain employees an average of 5.4 years, against 2.9 years at companies that struggle with it.

How it fails when gamed: Managers block moves to protect headcount, suppressing the metric in exactly the teams where it matters. Count blocked or discouraged transfer requests too, which you will only find by asking employees.

9. eNPS by team

What it is: Employee Net Promoter Score, calculated at team level rather than company level, with a minimum team size of five to protect anonymity.

How to collect it: One question, quarterly: how likely are you to recommend working here. Report by team, never by individual.

What it predicts: Advocacy and referral volume. It is a slow, blunt instrument, and that is fine. Happily.ai customers have recorded a 48-point eNPS improvement, but gains that size take quarters rather than weeks, which is why you cannot manage from this number alone.

How it fails when gamed: Managers ask for good scores in the meeting before the survey, which small teams make easy. Keep the reporting threshold at five or more and watch for teams whose eNPS is far out of line with their weekly sentiment trend.

10. Goal completion

What it is: The percentage of goals a team committed to at the start of a quarter that were completed by the end of it.

How to collect it: Whatever goal system you already run, with one rule: freeze the goal list at the start of the quarter. Goals added mid-quarter get counted separately.

What it predicts: It validates focus alignment. High alignment with low completion points to capacity or scoping. Low alignment with high completion means the team is busy on the wrong things.

How it fails when gamed: Sandbagged goals. Completion near 100% every quarter is a scoping signal, not a performance signal. Pair the rate with a difficulty rating set at commitment time.

11. Promotion rate of direct reports

What it is: The share of a manager's team promoted over a rolling 24 months, benchmarked against the company average for comparable roles.

How to collect it: HRIS promotion records. Use 24 months, because 12 is too noisy for teams under 10 people.

What it predicts: Whether a manager develops people or just deploys them. Managers who consistently produce promotable people are the ones worth putting in charge of other managers.

How it fails when gamed: Title inflation with no scope change. Check that promotions came with expanded responsibility, and track how promoted people perform 12 months later.

12. Time-to-productivity for new hires

What it is: Days from start date to the point a new hire meets the agreed definition of fully ramped, defined per role before the hire starts.

How to collect it: Manager and new hire mark the date independently at 30, 60, and 90 days. Use the later of the two.

What it predicts: Onboarding quality, which is largely a manager behavior. Gallup found only 12% of employees strongly agree their organization does a great job onboarding new employees, and the gap usually sits with the direct manager rather than with HR.

How it fails when gamed: Managers declare readiness early to look efficient. Anchor it to a role-specific output, and let the new hire's own assessment override an optimistic manager.

The 12 manager effectiveness metrics at a glance

Turning 12 metrics into one manager scorecard

Twelve numbers scattered across five systems is a research project nobody has time for. The mechanism that makes it operational is a manager scorecard: one view per manager, leading indicators on top, lagging indicators underneath, refreshed weekly.

DEBI sits at the top of that scorecard as the team-level engagement score. It works like a bathroom scale. It gives you the number, reliably and often, and the number is genuinely useful. It does not give you the diet. A DEBI of 58 tells a manager something is off. The leading indicators underneath tell them which behavior to change this week.

That distinction is the reason to run continuous measurement alongside an annual engagement survey. The survey gives a precise reading once a year plus a set of themes, which is a good annual instrument and a poor weekly one. A scorecard gives a manager one thing to do differently on Monday.

Across an organization, the same data rolls into a hotspot map: which teams are trending down, how fast, and for how long. That is the view that tells a COO where to send help before a resignation letter does. For a ready-made structure, our manager effectiveness evaluation template lays out the scoring format, and our 15Five comparison on manager effectiveness covers how platforms differ.

Honest tradeoffs: every one of these breaks under pressure

All 12 metrics are gameable when tied directly to compensation. That is Goodhart's law, given its familiar phrasing by anthropologist Marilyn Strathern in a 1997 paper: when a measure becomes a target, it stops being a good measure.

Recognition give-rate is the clearest case. Attach a bonus and within a quarter you have high volume and empty content. Feedback response time is next: fast closes, unresolved issues. eNPS in a team of six is one conversation away from being whatever the manager wants it to be.

Measurement changes behavior. That is the point of it, and also the risk. Three guardrails hold up in practice:

• Development first, evaluation second. Managers who believe the scorecard is a support tool report honestly. Managers who believe it is a performance case start managing the number.
• Never let one metric stand alone. Every leading indicator needs a paired quality check, and every lagging indicator needs a leading one that should have predicted it.
• Keep compensation one step removed. Tie pay to outcomes over a year or more, not to weekly behavior counts.

There is also a real cost. Collecting 12 metrics by hand across 30 managers is a part-time job, and data quality degrades the moment someone gets busy. Which leads to the sizing question.

[IN-ARTICLE IMAGE: Three nested rectangles of increasing size, the smallest solid, the middle dashed, the largest dotted]

Where to start, based on how many managers you have

Fewer than 10 managers: start with three leading indicators and nothing else. Recognition give-rate, 1:1 consistency, and team sentiment trend. All three fit in a shared spreadsheet at under 30 minutes a week, and at this size the data checks your judgment rather than replacing it. Add lagging indicators after two full quarters.

10 to 50 managers: run all six leading indicators and pick three lagging ones. Regretted attrition, goal completion, and time-to-productivity give the widest coverage for the least effort. This is the band where a scorecard starts paying for itself, because you can no longer hold 30 managers in your head.

More than 50 managers: you need automated collection or the data decays. Manual entry at this scale produces stale numbers that people stop trusting within two quarters, and a distrusted scorecard burns the credibility of the next attempt. Automate the six leading indicators first. They need weekly frequency, and they are the ones humans are worst at maintaining by hand.

Frequently asked questions

How do you measure manager effectiveness objectively? Combine behavioral data that does not depend on anyone's opinion (1:1 completion rates, recognition frequency, feedback response times) with anonymous team-reported outcomes (sentiment, eNPS, whether feedback was actually addressed). Neither source is objective alone. Behavioral data misses quality, and self-reported data carries bias. When they agree, you have a defensible read on management effectiveness. When they disagree, you have found something worth investigating.

What is a good manager effectiveness score? There is no universal number, and any vendor offering one is selling a benchmark you cannot audit. Use internal comparison: score each manager against your own company median and against their own trend over the previous two quarters. A manager below the median but improving for three straight months is usually a better bet than one above the median and sliding. On a 0 to 100 team engagement scale like DEBI, most healthy teams cluster between 65 and 80, and direction of travel matters more than position.

What is the difference between leading and lagging manager metrics? Leading metrics measure what a manager does (recognition, 1:1s, feedback response). Lagging metrics measure what happened to the team (attrition, promotions, goal completion). Leading metrics change within weeks and let you intervene. Lagging metrics take one to four quarters and let you verify. Leading indicators without lagging validation are activity tracking.

How often should you measure manager effectiveness? Leading indicators weekly, lagging indicators quarterly. Weekly catches a four-week decline while it is still reversible. Quarterly is the shortest window in which outcome data means anything. Annual-only measurement produces a report, not a change.

Can you measure manager effectiveness without a survey? Partially. Calendar data, recognition logs, promotion records, and attrition classification all come from systems you already run. What you cannot get without asking people is sentiment, feedback quality, and whether a manager's response resolved anything. A weekly pulse of two or three questions covers that gap at a fraction of the fatigue cost of an annual survey.

Start with three, not twelve

Building the whole scorecard at once usually ends with a spreadsheet nobody opens in March. Pick three leading indicators. Run them for a quarter. Add a lagging indicator that should confirm what they told you, and see whether it does. That loop, run four times, teaches you more than any annual assessment cycle, because it catches the behaviors while they are still changeable.

See how Happily.ai builds manager scorecards from daily behavioral data

Sources:

• Managers Account for 70% of Variance in Employee Engagement - Randall Beck and Jim Harter, Gallup (2015)
• Employees Want a Lot More From Their Managers - Jim Harter and Amy Adkins, Gallup (2015)
• Anemic Employee Engagement Points to Leadership Challenges - Gallup (2025)
• This Fixable Problem Costs U.S. Businesses $1 Trillion - Shane McFeely and Ben Wigert, Gallup (2019)
• Why the Onboarding Experience Is Key for Retention - Gallup
• 2022 Workplace Learning Report - LinkedIn Learning (2022)
• "Improving ratings": audit in the British University system - Marilyn Strathern, European Review 5(3) (1997)