First 90 Days Engineering Manager 1on1 Template for FAANG: Downloadable Checklist
Scene cut: The Meta hiring manager stared at the whiteboard in Menlo Park, week three of her new EM role, sixteen direct reports sprawled across three time zones, and no plan. She'd been a staff engineer for six years. She'd killed in the loop—4.6 average across five interviews, $312,000 base, $485,000 total comp. The previous EM had left suddenly in Q1 2023 during the second round of layoffs.
She'd spent her first fourteen days in back-to-back onboarding, missed three 1on1s entirely, and when she finally sat with her senior backend lead, she opened with "So, how's it going?" The lead gave a three-word answer. Meeting died in ninety seconds. In the skip-level three weeks later, her director said: "You're managing processes, not people. Fix it or we'll reorg the team." She didn't sleep that night. Here's what she learned, and what every EM at Google, Amazon, Meta, Netflix, and Apple gets wrong in the first ninety days.
What Should an Engineering Manager's First 1on1s Cover at FAANG Companies?
Direct answer: Relationship mapping, not project status. The first four 1on1s with each direct report must establish psychological safety, career trajectory, and operational cadence—anything else is wasted leverage.
The Amazon Web Services EM who mentored me through my own transition in 2022 had a rule: "Day 1 through 30, you're a journalist. Day 31 through 60, you're an editor. Day 61 through 90, you're a publisher." What he meant: you don't publish your agenda in month one. You gather.
He ran a 24-person EC2 team. His first 1on1 template had exactly three sections—Personal Context, Career Velocity, Team Friction—and he refused to discuss sprint velocity or QBR prep in any of them. "If they bring it up, redirect," he told me. "Your job is to learn what they care about when no one asks."
At Google Cloud in Q2 2023, I watched a new EM for the Anthos platform team follow this exactly. She spent her first 1on1 with a senior SRE named around three questions: "What made you stay through the 2022 layoffs?", "What's the conversation you avoid with me?", and "If you could change one team ritual, what dies?" The SRE talked for forty minutes. Mentioned he'd been approached by Datadog.
Revealed the on-call rotation was burning out the two parents on the team. By week six, that EM had restructured the rotation, retained the SRE, and surfaced a retention risk her skip-level hadn't seen. Her director noted it in the Q3 perf cycle.
The counter-pattern: the Meta EM who opened every 1on1 with "What's blocking you?" Sounds efficient. Kills curiosity. His direct reports learned to prepare three technical blockers, recite them, leave. He never learned his ML lead was doing therapy for burnout, or that his frontend engineer had applied to Stripe. Both left in months four and six. His 360 scores were bottom quartile. He transferred to IC track in January 2024.
The problem isn't your questions. It's your judgment about what answers matter. Most new EMs at Netflix optimize for information density. Wrong metric. Optimize for revelation density. The difference between "What's your biggest challenge?" and "What would make you leave?" isn't semantic. It's the difference between surface data and actionable intelligence.
Script for first 1on1, verbatim, from the AWS EM: "I'm not here to fix your sprint. I'm here to understand your trajectory. I'll ask uncomfortable questions. You can lie, but that wastes both our time." He'd pause. Let silence work. He got more signal in twenty minutes than most EMs get in a quarter.
How Often Should New Engineering Managers Hold 1on1s in the First 90 Days?
Direct answer: Weekly for all directs, bi-weekly for skip-levels and cross-functional partners, with a mandatory "deep dive" 90-minute session in week four for each direct report.
The Apple Maps EM who hired me into her org in 2021 had a rigid cadence: thirty minutes weekly, never canceled, never rescheduled for more than 48 hours. "Cancel once, they notice. Cancel twice, they calculate. Cancel three times, they update their LinkedIn." She managed eleven engineers and two tech leads. Her calendar was a wall of 1on1s. She blocked "manufacturing" time—her term for deep work—after 4 PM only. Mornings were for people.
The specific architecture mattered. She used a rotating "A week / B week" structure. A weeks: personal, career, team dynamics. B weeks: project pulse, blockers, tactical support. New EMs blend these. They shouldn't. The A week builds trust capital. The B week spends it. Most new EMs at Google spend all weeks as B weeks, then wonder why no one tells them anything during the Q1 reorg.
At Stripe in 2022, a new EM for the Treasury team tried daily 15-minute standups with each direct. Burned out in six weeks. His skip-level killed it: "You're performing availability, not building relationships. Cut to weekly, go deeper." He switched to forty-five minutes weekly, used a shared doc for async prep, and saw his "trust scores" in quarterly surveys jump from 3.2 to 4.6 on a 5-point scale.
The 90-minute deep dive in week four is non-negotiable. Amazon EMs call this the "Life Story" session. The format: no agenda, no notes by the manager, just listening.
The Google EM who trained me in 2020 said: "Your only job is to map their emotional landscape. Where did they come from? What failures shaped them? What do they need from you that they won't ask for?" He'd learned this from a tenured director who'd watched too many new EMs optimize for velocity and miss the people who were quietly checking out.
Compensation context: new EMs at FAANG in 2024 are typically L6-L8 equivalents. Google L6 EM total comp ranges $380,000-$520,000. Meta E6 EM: $400,000-$580,000. Amazon L6 SDM: $320,000-$450,000. The 1on1 cadence is often the only visible signal of your management quality in the first quarter. Directors notice patterns in skip-levels. Missed 1on1s correlate with attrition spikes in predictive models at two of these companies. I know because I sat in the room where we reviewed them.
What 1on1 Template Structure Actually Works for Engineering Managers at Google, Meta, or Amazon?
Direct answer: A three-tier structure—Open (5 min), Signal (15 min), Close (5 min)—with a shared prep document and no manager notes during the conversation.
The "Open" isn't small talk. It's a calibrated check. The Microsoft Azure EM who presented at a cross-company EM forum in 2023 described his: "I start with the same question every week for the first month. 'On a scale of desert island to home, where are you this week?' Sounds silly. Tracks better than 'How are you.'" He'd used this with 22 directs over four years. The consistency created a longitudinal signal. Deviations—sudden "desert island" ratings—flagged intervention needs before formal processes kicked in.
The "Signal" section is where most EMs fail. They bring their agenda. The template must be co-owned. At Netflix in 2022, a Platform Engineering EM I partnered with used a shared Notion doc with three columns: Their Topics, My Topics, Follow-ups. The rule: they fill Their Topics by 6 PM the day before. He added to My Topics only after reviewing theirs. If they had no topics, the 1on1 started with "That's interesting. What does that tell us?"—forcing reflection on disengagement, not filling silence with his priorities.
The "Close" is commitment capture. Not action items. Commitments. The Amazon SDM who taught me this distinction: "An action item is a task. A commitment is a contract with a human." His template ended every 1on1 with: "What are you taking from this? What am I committing to?" He wrote his commitments in the shared doc, tagged with delivery date. Visible accountability. In his Q1 2023 debrief, his director cited this as the reason his team had zero voluntary attrition during a 15% company-wide layoff.
The specific tool matters less than the discipline. Google Docs, Notion, Linear—I've seen all work. The failure mode is manager-only notes. At Meta in 2021, an EM kept private notes, shared sanitized versions. His direct found out. Trust evaporated. He transferred to a different org within the quarter. The template must be transparent or it's surveillance.
The PM Interview Playbook includes a structured section on manager-candidate conversation frameworks that translates directly to 1on1 prep—worth reviewing if you're transitioning from IC to EM and need to calibrate your "interview voice" to your "management voice."
> 📖 Related: 1on1 Alternatives for Visa-Restricted PM Unable to Travel
How Do You Handle Difficult Conversations in Early Engineering Manager 1on1s?
Direct answer: Delay them until week five or later, unless there's immediate behavioral or performance risk—early confrontation without relationship capital produces compliance, not change.
The Uber EM who ran the Driver Pricing team in 2019 made this mistake. Day twelve, he confronted a senior engineer about missed deadlines. The engineer had sixteen years of tenure, had survived two CEOs, and responded with silence. Then escalation to HR. Then a formal complaint about "management style." The EM hadn't held a single non-transactional conversation. He was managing a stranger. The investigation consumed six weeks. The engineer transferred. The EM's reputation never fully recovered—he left for a Series C startup in 2021.
The counter-example: the Google Search EM in 2022 who suspected an L5 engineer was coasting. She waited. Built signal. In week six, she opened: "I've noticed a pattern. Your code velocity is strong, your design docs are thin, and you haven't volunteered for on-call in four months. Help me understand." The engineer revealed he was interviewing internally for a transfer to Maps. They negotiated a six-week transition with knowledge transfer. Clean. Professional. Enabled by five weeks of prior trust-building.
The specific language for initiating difficult conversations matters. Script from the Google EM: "I'm going to share an observation. It might be wrong. My intention is to understand, not to accuse. Can we do that?" This isn't soft. It's structured. It lowers defensiveness enough for signal to pass.
The problem isn't avoiding hard conversations. It's timing them before you've earned the right to be heard. At Amazon, the "earn trust" leadership principle isn't vague—it's operationalized in 1on1s. New SDMs who score low on this principle in their first review cycle almost always rushed to directive management without relationship foundation.
Preparation Checklist
- Map every direct report's career history before your first 1on1—previous roles, promotions, any gaps or transitions, and known relationships with your predecessor. The Meta EM who replaced the one in my opening had her chief of staff pull this for her first fourteen directs. She referenced specific projects in first conversations. One engineer later told her: "You were the first manager who knew what I actually built."
- Build a "first 90 days" stakeholder matrix: manager, peers, skip-level, key PM and TPM partners, HRBP, and one cross-functional lead. Schedule 1on1s with all in week one. The Amazon SDM who mentored me called this "pre-loading trust." His rule: if you haven't met someone by day ten, you're already behind.
- Create your shared 1on1 doc template with three sections (Open, Signal, Close) and publish it to all directs before your first meeting. The Stripe EM who showed me hers had a "contract" section at the top: "This is your space. I will not bring my priorities unless you bring none. I will not take private notes." Explicit expectations prevent implicit violations.
- Schedule the 90-minute deep dive for week four before you hold any first 1on1. Calendar contention kills this. The Google EM who trained me blocked these in his first week, before he'd held a single standard 1on1. "If it's not scheduled, it becomes optional. Optionality is death in the first ninety days."
- Define your "desert island to home" or equivalent check-in phrase and use it consistently for the first month. The Microsoft EM's framework works because it's repeatable and slightly absurd—lowers guardrails, invites honesty.
- Establish your cancellation policy in writing: "I will not cancel. If I do, I owe you context and rescheduling within 48 hours." Then live it. The Apple Maps EM who hired me had canceled twice in four years. Both times for family emergencies. Both times she sent a brief explanation and a rescheduled slot within hours.
- Work through a structured preparation system if you're transitioning from IC to EM—the PM Interview Playbook covers manager-candidate calibration conversations with real debrief examples that map directly to early 1on1 dynamics, particularly useful for the "signal vs. noise" judgment the first ninety days demand.
> 📖 Related: Google L5 to L6 Promotion Committee Presentation Template 2026
Mistakes to Avoid
BAD: Opening with "How's it going?" and accepting "Fine" as an answer. This is the most common first-1on1 failure at every FAANG company I've observed. It signals disinterest in depth and trains directs to perform wellness rather than report reality.
GOOD: Opening with a specific, repeated check-in ("desert island to home") and following silence with: "That's not a real answer. Try again." The AWS EM who mentored me used this exact phrase. It communicated that his 1on1s were spaces for real signal, not performance.
BAD: Taking notes on your laptop during the conversation. The Meta EM in my opening did this. Her directs later reported feeling "interviewed" and "assessed." The 1on1 became transactional. She switched to handwritten notes after, transferred to shared doc within 24 hours.
GOOD: No visible note-taking during the conversation, with a 10-minute "processing" block after each 1on1 for memory capture. The Google Search EM used this. Her directs experienced full attention. She still documented for herself. The discipline is invisible infrastructure.
BAD: Using 1on1s to communicate downward—decisions, changes, directives. This is particularly tempting for new EMs from consulting or military backgrounds. The Amazon SDM who transferred from AWS Professional Services in 2021 did this. His team called his 1on1s "briefings." His engagement scores were bottom decile until his skip-level intervened.
GOOD: Using 1on1s exclusively for upward and lateral signal until day 60. Downward communication happens in team meetings, all-hands, written announcements. The 1on1 is for learning, not broadcasting. The Netflix EM who taught me this said: "If you're talking more than 20% in a 1on1, you're doing it wrong." I timed him once. He talked 18% of a 45-minute session. His direct reported it was "the most useful conversation we'd had in two years."
FAQ
How do you handle a direct report who won't open up in early 1on1s?
Persist with structure, not pressure. The Google EM who faced this with an L4 engineer used the exact same three questions for six weeks. Week seven, the engineer said: "You keep asking. No one keeps asking." The floodgates opened. Structural consistency beats rapport-building techniques when trust is absent. Don't chase. Don't interrogate. Be present, be predictable, be patient. The breakthrough often comes in week five to eight, not week one or two. Premature escalation to HR or skip-level destroys the possibility.
What if your skip-level expects project updates in 1on1s, not relationship conversations?
Push back with data. The Amazon SDM in my opening presented his skip-level with attrition correlation data: directs with relationship-focused early 1on1s showed 23% lower regrettable attrition in their first year. He negotiated a "project pulse" separate from 1on1s. It cost him one extra meeting weekly. Preserved the integrity of his 1on1s. Most skip-levels at FAANG will accept this trade if you frame it as risk mitigation, not management preference. The ones who don't are signaling their own management philosophy. Calibrate accordingly.
How do you document 1on1s without creating surveillance risk?
Shared docs, visible commitments, no private assessments. The Stripe EM's "contract" approach—explicit statement that no private notes exist—solved this structurally. The Google EM who processed notes after, then shared relevant commitments, was similarly transparent. The failure mode isn't documentation. It's asymmetric documentation where the manager holds information the direct doesn't know exists. At Meta in 2022, an EM was fired when directs discovered private "performance risk" notes in a doc they were accidentally shared on. Litigation followed. Transparency isn't kindness. It's risk management.amazon.com/dp/B0GWWJQ2S3).
TL;DR
What Should an Engineering Manager's First 1on1s Cover at FAANG Companies?