Template: Handling an Underperformer – Feedback Script for New Managers
How should a new manager start a feedback conversation with an underperforming direct report?
The opening line must state the performance gap, reference concrete metrics, and set a collaborative tone; don’t begin with empathy‑only language, but with data‑driven urgency.
In a Q2 debrief for a mid‑size AI team, the hiring manager interrupted the new senior PM because the manager’s opening was “I know you’re struggling, let me know how I can help.” The panel rejected the script immediately: the manager signaled uncertainty instead of authority.
The senior PM’s performance had slipped from a 95 % sprint success rate to 71 % over three sprints, a clear numeric drop. The correct opening was: “Your delivery rate fell from 95 % to 71 % over the last three sprints; let’s identify the two blockers that caused this regression and agree on a 10 % improvement plan for the next sprint.”
Insight 1 – The authority anchor: New managers must anchor the conversation with a single, verifiable metric before any soft language. This forces the employee to view the issue through the lens of outcomes, not feelings.
Script:
“[Name], over the past 21 days your average cycle time increased from 4 days to 7 days, and defect leakage rose from 0.3 % to 1.2 %. We need to reverse that trend. What two factors are most responsible?”
What concrete steps should a new manager outline in the feedback script to drive improvement?
List three actionable commitments, each tied to a measurable target and a deadline; don’t give vague “work harder” advice, but prescribe a “test‑and‑learn” loop with explicit checkpoints.
During a hiring committee for a cloud‑services product lead, the panel heard a candidate describe a “general improvement plan” that included “better communication.” The HC rejected the plan because it lacked cadence and metrics. The winning candidate presented a three‑step loop: (1) define a weekly KPI, (2) run a 48‑hour experiment, (3) review results in a 30‑minute stand‑up. That structure convinced the committee the candidate could enforce accountability.
Insight 2 – The loop of measurable iteration: Underperformance is a hypothesis; the script must turn it into a testable experiment with a defined success threshold.
Script:
- Goal definition (Day 1): Reduce average cycle time to ≤5 days by Friday, Sep 15.
- Experiment (Days 2‑4): Implement a “blocked‑ticket” triage meeting each morning; record time saved.
- Review (Day 5): In a 15‑minute one‑on‑one, compare actual cycle time to the 5‑day target and decide whether to adopt, adjust, or discard the triage.
How can a new manager handle defensive reactions without derailing the feedback session?
Interrupt the defense within the first 30 seconds, restate the metric, and pivot to a problem‑solving question; don’t let the employee dominate the agenda, but keep the conversation on the pre‑agreed agenda.
In a senior director debrief for a fintech product, the new manager’s first‑time conversation was derailed when the engineer replied, “You don’t understand the legacy code.” The director stopped the interview, noting the manager’s failure to reassert the agenda. The manager who succeeded later said, “I hear the legacy constraints; let’s isolate the impact on the 12 % defect increase you’re seeing.” The manager’s quick pivot reclaimed control and turned the defense into data.
Insight 3 – The agenda‑re‑claim: Defensive statements are signals that the employee is protecting a narrative; the manager must immediately re‑anchor to the original data point.
Script:
Employee: “That metric ignores the new API rollout.”
Manager: “The metric includes all tickets, including the API work; the defect rate rose from 0.3 % to 1.2 % after the rollout. Which two steps can we take to bring it back under 0.5 %?”
> 📖 Related: Coffee Chat Strategy for PM Transition from Consulting to Product at Meta: Scripts That Work
When should a new manager schedule follow‑up check‑ins, and how often?
Set the first follow‑up at 48 hours, a second at one week, and a final at 30 days; don’t wait for the next quarterly review, but embed rapid cadence to signal seriousness.
In a hiring committee for a data‑platform PM, the panel asked the candidate how often they would meet with a low‑performing data engineer. The candidate said, “We’ll review quarterly.” The HC voted the candidate out. The winning candidate described a “48‑hour, 1‑week, 30‑day” cadence, citing a case where a data engineer’s query latency dropped from 2.3 s to 0.9 s after a 48‑hour check‑in and a 7‑day plan.
Insight 4 – The rapid‑feedback cycle: Early, frequent check‑ins create a feedback loop that compresses the time between problem identification and correction, preventing the issue from becoming systemic.
Script:
- Day 2: Quick status email – “What did you discover in the triage meeting?”
- Day 7: 15‑minute stand‑up – “Did the new triage reduce cycle time to ≤5 days?”
- Day 30: Formal review – “Compare the 30‑day defect rate to the 1.2 % baseline; decide next steps.”
What language should a new manager avoid to prevent the underperformer from feeling threatened?
Never use “you always/never” or “you’re the problem”; instead, use “the data shows” and “we can adjust the process”; the problem isn’t the person’s character, but the observable output.
In a debrief for a mobile‑app product lead, the manager opened with “You never meet your deadlines.” The hiring panel noted the statement as a red flag for cultural toxicity. The manager who got the role said, “The data shows a 22 % missed‑deadline rate over the last six sprints; let’s adjust the planning process.” The panel praised the neutral framing.
Insight 5 – The neutral data frame: Language that attributes performance to data removes personal blame and reduces defensiveness, keeping the conversation solution‑focused.
Script:
Wrong: “You’re always late with your deliverables.”
Right: “The sprint reports show a 22 % missed‑deadline rate for your tickets over the past six sprints; let’s explore two process changes to bring that below 5 %.”
> 📖 Related: Remote PM Job Alternatives for Visa Holders in the US: Sponsorship-Friendly Options
Preparation Checklist
- Review the last 30 days of the direct report’s quantitative output (cycle time, defect rate, delivery success).
- Identify the single metric with the largest deviation from the team norm; note the exact numbers (e.g., 7 days vs. 4 days).
- Draft the opening line using the authority anchor template; rehearse it aloud.
- Outline a three‑step measurable loop, assigning dates and success thresholds (e.g., reduce defect rate to ≤0.5 % by Day 30).
- Schedule calendar invites for Day 2, Day 7, and Day 30 follow‑ups; include agenda items in each invite.
- Anticipate two likely defensive statements and script agenda‑re‑claim responses.
- Work through a structured preparation system (the PM Interview Playbook covers data‑first feedback scripts with real debrief examples).
Mistakes to Avoid
BAD: “I know you’re overwhelmed; just try harder.” – Vague, no metric, no accountability.
GOOD: “Your defect rate rose from 0.3 % to 1.2 % in three weeks; let’s test a daily triage to cut it back to ≤0.5 % by Day 30.”
BAD: “We’ll talk about this in the next quarterly review.” – Delays corrective action, signals low priority.
GOOD: “We’ll reconvene in 48 hours to review the triage results, then again in one week for a progress check.”
BAD: “You never meet deadlines.” – Absolutes trigger defensiveness and can be perceived as personal attack.
GOOD: “The sprint data shows a 22 % missed‑deadline rate for your tickets; let’s adjust the planning process together.”
FAQ
What if the employee disputes the numbers?
The judgment is to stand on the documented metrics; request the employee to share any data that contradicts the reports, but keep the conversation anchored to the original numbers.
How long should the improvement plan run before deciding on termination?
A 30‑day data‑driven loop is the minimum; if after the final review the metric remains above the agreed threshold, the manager should initiate a performance‑action process.
Can I use this script for remote team members in different time zones?
Yes, but adjust the cadence to respect time‑zone overlap; the 48‑hour check‑in can be an asynchronous status update, while the 7‑day and 30‑day reviews should be live video calls to preserve alignment.amazon.com/dp/B0GWWJQ2S3).
TL;DR
In a Q2 debrief for a mid‑size AI team, the hiring manager interrupted the new senior PM because the manager’s opening was “I know you’re struggling, let me know how I can help.” The panel rejected the script immediately: the manager signaled uncertainty instead of authority.
The senior PM’s performance had slipped from a 95 % sprint success rate to 71 % over three sprints, a clear numeric drop. The correct opening was: “Your delivery rate fell from 95 % to 71 % over the last three sprints; let’s identify the two blockers that caused this regression and agree on a 10 % improvement plan for the next sprint.”