A new manager at Uber should treat underperformance as an evidence problem first, a feedback problem second, and a personnel problem only after the record is clear. The mistake is not being too hard; the mistake is being vague long enough for the situation to turn political.
TL;DR
A new manager at Uber should treat underperformance as an evidence problem first, a feedback problem second, and a personnel problem only after the record is clear. The mistake is not being too hard; the mistake is being vague long enough for the situation to turn political.
In a Q3 debrief, the room never cared that the manager thought someone was “smart but inconsistent.” They cared whether there were dated examples, an explicit standard, and a 30-day reset with weekly checkpoints. Not a pep talk, but a management contract.
If the person is already in a six-figure, high-visibility role, delay compounds fast. A 4-round interview rewards narrative; a live team does not. The standard is simple: write the gap, show the proof, name the expectation, and decide within 4 to 6 weeks whether the behavior moved.
Who This Is For
This is for the first-time manager who inherited a strong-looking IC and discovered, after two skipped deadlines and one tense stakeholder call, that the person is not carrying the role. It is also for the newly promoted manager who still wants to be liked, and is therefore avoiding the one conversation that would restore order.
In an Uber-style environment, the failure mode is rarely cruelty. It is drift. The team starts compensating, the skip-level starts asking questions, and the manager begins narrating intent instead of showing output. The room does not reward interpretation. It rewards clarity.
What should a new manager do in the first 30 days with an underperformer at Uber?
The first 30 days should be a reset on expectations, not a character study. A new manager who opens with personality theories has already lost the room.
In one performance review I sat through, the manager said, “I think she’s disengaged.” The director cut in immediately: “Show me the work.” That was the right correction. In a metric-heavy company, the first job is to pin the problem to dated artifacts, not to moods.
The framework is blunt. Identify 2 or 3 concrete misses. Write the standard in one sentence. Schedule weekly 20-minute check-ins for 4 weeks. End each meeting with one owner, one deliverable, and one date. If the person improves only in tone but not in output, the issue is still active.
Not a morale conversation, but a performance contract. Not “How do you feel about the role?”, but “What will be different by next Friday?” Not a broad coaching arc, but a short sequence of visible commitments. The first month is about building a record that can survive a skip-level conversation.
How do you tell a skill gap from a behavior problem?
The right judgment is usually slower than the manager’s first instinct. Most new managers over-diagnose attitude because attitude feels easier to name than process debt.
In a skip-level conversation, I heard a hiring manager say, “He knows what good looks like; he just doesn’t close loops.” That sentence matters because it separates capacity from execution. The employee may not need inspiration. They may need structure, better prioritization, or a narrower definition of done.
The counter-intuitive point is that many “motivation” problems are actually ambiguity problems. The person is working, but the work is not aligned to the manager’s bar. Or they are competent, but the team’s operating rhythm is sloppy enough that competence never turns into trust. Not a personality flaw, but a system failure until evidence proves otherwise.
Use three buckets: can’t, won’t, and inconsistent. Can’t means the person lacks the skill or judgment. Won’t means they understand the standard and still ignore it. Inconsistent means they can perform, but only under pressure, and only when the manager is hovering. Each bucket demands a different response. Treating them as the same is managerial laziness.
What feedback framework works in a metrics-heavy company?
The only framework that survives Uber is one that ties observation to consequence. Storytelling without timestamps is decoration.
In a Q3 debrief, a manager defended an underperformer by saying, “They’re collaborative.” The room asked for two examples. The manager had one vague anecdote and a lot of adjectives. That is why performance discussions fail. The organization does not trust adjectives. It trusts evidence, sequence, and repetition.
The usable framework is short: observe, standardize, contrast, and confirm. Observe the exact miss. Standardize the expected behavior or result. Contrast what happened with what should have happened. Confirm the next checkpoint in writing. This is not a coaching script. It is a decision trail.
A single miss is data. Two misses are a pattern. Three misses without escalation become a management failure. That is the cold arithmetic inside a high-velocity company. Not a sandwich, but a ledger. Not “I want to help you grow,” but “Here is the bar, here is the gap, here is the timeline.”
The strongest managers do not over-talk. They close the loop. They send a recap after the meeting, they reference the same standard in the next 1:1, and they stop pretending that kindness and clarity are opposites. They are not. Clarity is what keeps kindness from turning into avoidance.
When does feedback become escalation?
Escalation should begin when the pattern is visible, not when the manager feels uncomfortable. Waiting for emotional certainty is how teams end up in avoidable PIP discussions.
I have seen HC-style calibration rooms where everyone already knew the outcome before the manager entered the sentence. The reason was simple: the evidence trail was either clean or absent. If the trail was clean, the room moved. If it was absent, the manager got pulled into a debate about whether the issue was “really there.” That is a loss of leverage.
The judgment is straightforward. If you have 2 documented misses, 2 explicit resets, and no durable change after 4 to 6 weeks, escalate. If trust has broken in front of stakeholders, escalate faster. Not because the person is bad, but because the organization cannot run on hope. Not because you want to punish, but because the team needs predictability.
Do not confuse escalation with aggression. Escalation is simply the point at which the manager stops improvising and starts using the organization’s decision system. The new manager who waits too long usually thinks they are protecting the employee. In practice, they are protecting their own discomfort.
How do you run the final conversation without making it emotional theater?
The final conversation should be short, explicit, and closed-ended. Drama is a substitute for judgment.
In one manager conversation I sat in on, the leader tried to soften the blow by speaking for ten minutes about “fit” and “trajectory.” The employee left confused, and the manager left relieved. That is the wrong outcome. A difficult conversation is not successful because it feels humane in the moment. It is successful because the person understands the decision and the next step.
State the facts first. Name the prior feedback. State what changed and what did not. Then say the decision in plain language. Not “This has been a journey,” but “The role is not being met.” Not “We need to revisit alignment,” but “The current pattern is not acceptable.” Not a negotiation, but a conclusion.
The best managers do not perform certainty they do not have. They speak from the record they built. That is why the earlier weeks matter so much. If the record is clean, the final conversation is brief. If the record is sloppy, the final conversation turns into a defense of the manager’s own delay.
Preparation Checklist
Prepare the case, not the apology.
- Pull the last 6 to 8 weeks of artifacts, including missed deadlines, stakeholder feedback, and delivery notes. A memory-based case is weak.
- Write the performance gap in one sentence. If the sentence uses words like “ownership,” “presence,” or “proactivity” without examples, it is not ready.
- Separate output, behavior, and context. A missed launch is not the same as a communication problem, and neither is the same as a workload issue.
- Set a 30-day reset with weekly 20-minute check-ins. End every meeting with one commitment, one owner, and one date.
- Send a written recap after each 1:1. The recap is not bureaucracy. It is the record that survives emotion and memory drift.
- Align your skip-level or HR partner before escalation. A manager who hides until the last minute usually turns a process into a surprise.
- Work through a structured preparation system (the PM Interview Playbook covers feedback loops and debrief examples in a way most new managers never practice), because the judgment muscle is the same even when the context changes.
Mistakes to Avoid
Most new managers fail by hiding inside vague feedback or over-indexing on kindness. Both errors look humane. Both are expensive.
- BAD: “You need to be more proactive.” GOOD: “You missed the Thursday handoff, and the dashboard went out 18 hours late.”
Vague language invites disagreement. Concrete language creates accountability.
- BAD: Waiting until the review cycle to surface the issue. GOOD: Addressing the miss within the same week and resetting expectations for the next deliverable.
Delay turns a fixable problem into a narrative.
- BAD: Turning the issue into a debate about personality or intent. GOOD: Keeping the conversation on evidence, pattern, and next observable change.
People can argue intent forever. They cannot argue timestamps as easily.
FAQ
How fast should a new manager act?
Fast enough that the pattern does not become normal. The correct window is usually days, not months. By the time a 90-day review arrives, the manager has already signaled tolerance, and the team has noticed.
Should I start with empathy or directness?
Start with directness. Empathy without clarity becomes avoidance, and avoidance is what damages trust. The employee needs to know the standard before they can respond to it.
What if the underperformer is well-liked?
Well-liked is not a performance category. A person can be pleasant, respected, and still fail the role. The manager’s job is not to protect social comfort. It is to protect the team’s output and standards.
Ready to build a real interview prep system?
Get the full PM Interview Prep System →
The book is also available on Amazon Kindle.