LeadershipIntermediate6 min read

Skip-Level One-on-Ones

A skip-level one-on-one is a structured private meeting between a leader and an employee two or more levels below them in the org chart, with the direct manager intentionally NOT present. The purpose is not to evaluate the manager — it is to (1) get unfiltered ground-truth about the work, (2) build relationships with rising talent the leader would otherwise never meet, and (3) detect organizational dysfunction (priorities lost in translation, manager weakness, team morale collapse) before it surfaces in attrition or missed quarters. Done well, skip-levels are the single highest-signal leadership instrument for a leader running 50-500 people. Done poorly, they become anxiety-inducing performance theater that destroys trust with both the IC and the manager.

Also known asSkip-Level 1:1sSkip MeetingsTwo-Down ConversationsCross-Layer 1:1s

Challenge a friend Browse library

The Trap

The trap is using skip-levels as either (a) information warfare against a manager you distrust, or (b) a vague 'how's it going?' chat that wastes the IC's time and surfaces nothing. Both failure modes share the same root cause: no declared purpose. The IC walks in confused — am I being interviewed for promotion? Am I being asked to inform on my boss? — and gives the safest possible answers. The leader walks out with nothing and concludes 'skip-levels don't work.' The second trap: leaders who run skip-levels then act on the information without protecting the source. The IC said 'roadmap priorities feel unclear' on Tuesday; on Wednesday the leader emails the manager 'roadmap priorities feel unclear' — the IC is now toxic in the manager's eyes and will never speak truth again.

What to Do

Run skip-levels on a deliberate cadence with a declared purpose: (1) Frame on calendar invite: 'No agenda from you needed. I want to learn about your work, hear what's getting in the way, and answer any questions you have about company strategy. This is not a performance conversation.' (2) Open with: 'Three questions — what are you working on, what's the biggest thing slowing you down, and what would you change if you ran the company?' (3) Take notes for yourself; never share the source of any feedback with the manager. (4) Aggregate themes across 5-8 skip-levels before acting on anything organizational. (5) Tell the manager in advance: 'I'll run skip-levels with your team quarterly. It's about ME staying in touch, not evaluating you.' (6) Cap at 30 minutes, run 6-10 per quarter, target high-potential ICs and engineers central to the business.

Formula

Skip-Level Coverage = Quarterly Skip-Levels Run ÷ High-Potential ICs in Org — target 60-100% annually

In Practice

Andy Grove at Intel formalized skip-level meetings as a core management discipline in 'High Output Management' (1983) — he argued the CEO who only hears from direct reports is operating with a 3-layer information distortion and has no idea what's actually happening in the company. Grove ran skip-levels with engineers across Intel and credited the practice with catching the memory-business decline (employees knew it was over before management did) earlier than the executive team would have on its own. The pattern was later embedded at Netflix, Stripe, and Airbnb, where Brian Chesky has publicly described his skip-level cadence as a foundational operating practice.

Pro Tips

01
The single best opening question is 'What would you change if you ran this company?' It bypasses the safe-answer instinct and gets to opinion. The worst opening question is 'How's it going?' which produces the answer 'good' 95% of the time.
02
Tell the IC's manager BEFORE the meeting, not after. Surprising a manager that their report is meeting with their boss creates paranoia and damages your relationship with both. The frame 'I'm running quarterly skip-levels across your team — here's the schedule' normalizes the practice.
03
Aggregate before you act. If one IC says 'planning is a mess,' it's noise. If five ICs across three teams say 'planning is a mess,' it's signal. Acting on N=1 skip-level data is how leaders blow up trust and make wrong decisions.

Myth vs Reality

Myth

“Skip-levels undermine the manager's authority”

Reality

Done with explicit framing ('this isn't about evaluating your manager'), skip-levels strengthen the org. The leader gets ground-truth, the IC feels heard, and the manager benefits from a leader who actually understands the team's reality. Skip-levels only undermine managers when leaders use them to second-guess decisions or share sourced criticism — that's a leadership failure, not a skip-level failure.

Myth

“Skip-levels should be optional and ad-hoc”

Reality

Ad-hoc skip-levels signal crisis ('why is the CEO suddenly asking to meet with me?'). Scheduled, normalized skip-levels are routine and low-anxiety. The cadence — predictable quarterly cycles, openly communicated — is what makes the practice safe. Treating them as 'when I have time' makes every meeting feel like an investigation.

Try it

Run the numbers.

Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.

🧪

Knowledge Check

You're a VP running skip-levels with engineers two levels below you. In one meeting, an IC tells you 'my manager is checked out — barely shows up to 1:1s, no roadmap clarity.' What's the right next move?

Industry benchmarks

Is your number good?

Calibrate against real-world tiers. Use these ranges as targets — not absolutes.

Skip-Level Cadence (Founder/CEO)

CEO/VP of 100-500 person org

Disciplined Operator

10-15 skip-levels/quarter

Solid Practice

6-10 per quarter

Token Effort

2-5 per quarter

Abdicated

0-1 per quarter (or only in crisis)

Source: Hypothetical: Composite of operator interviews and Andy Grove's 'High Output Management'

Real-world cases

Companies that lived this.

Verified narratives with the numbers that prove (or break) the concept.

🟦

Intel (Andy Grove era)

1980s

success

Andy Grove formalized skip-level meetings as core management discipline in 'High Output Management' (1983). Grove argued that a CEO hearing only from direct reports operates with multi-layer information distortion. He ran skip-levels with engineers across Intel and credited the practice with surfacing the strategic reality of the memory-business decline earlier than the executive team would have on its own — engineers knew the business was over before VPs would admit it. Intel's pivot from memory to microprocessors in 1985 was informed in part by ground-truth Grove gathered through skip-levels.

Author

Andy Grove (CEO, Intel)

Source

'High Output Management' (1983)

Practice

Routine skip-level meetings with ICs

Strategic Impact

Earlier memory-business pivot

The CEO who only hears from direct reports has no ground-truth. Skip-levels are not optional in a 100+ person company — they're the only way a leader gets signal that hasn't been filtered through the management chain.

Source ↗

🏢

Hypothetical: 200-person Series C SaaS

Composite case

success

A Series C SaaS CEO (200 people) stopped running skip-levels at the COO's request, citing 'authority confusion.' Six months later, she discovered via an exit interview that two product workstreams had stalled for a quarter without surfacing through the management chain — each layer assumed the next would escalate. She restarted skip-levels with stricter discipline (listen 80%, never editorialize, tell managers in advance) and within three months had caught two emerging issues the chain hadn't surfaced. The COO's underlying concern was real but the solution was discipline, not abandonment.

Headcount

200

Issues Missed (no skip-levels)

2 stalled workstreams

Issues Caught (restored)

2 emerging risks

Cadence Restored

10/quarter

Skip-level dysfunction is usually a discipline problem, not a structural one. When ICs quote the leader back to their manager, the leader was talking too much. Fix the discipline; keep the practice.

Decision scenario

The Sourced Feedback Trap

You're a VP of Engineering. In a skip-level with a senior IC, she tells you 'My manager (one of your directs) hasn't done a 1:1 with me in 6 weeks and the team has no idea what we're shipping next quarter.' She asks you not to share that she said this. Your next 1:1 with that manager is tomorrow.

Skip-Levels Completed This Quarter

8 of 10 planned

Source IC

Senior, well-respected

Manager Tenure

14 months

Team Size

9 engineers

Decision 1

Tomorrow's 1:1 with the manager. You have feedback that's specific (no 1:1s for 6 weeks, no roadmap visibility) and credible — but sourced from one IC who asked for confidentiality.

In the 1:1, say: 'I heard from someone on your team that you haven't done 1:1s in 6 weeks — is that true?' Even without naming the source, the specifics will identify them.Reveal

The manager immediately reverse-engineers who said it (only one IC fits the pattern). She confronts the IC the next day. The IC stops trusting you. Word spreads in the team that skip-levels are a snitch line. Your next 4 skip-levels surface nothing — everyone has learned to give safe answers. You've burned your highest-signal listening tool to address one issue.

Skip-Level Signal: High → ZeroIC Trust: DestroyedIssue Addressed: Yes (but at huge cost)

Don't mention it tomorrow. Run 3-4 more skip-levels across that team in the next 3 weeks to test whether the pattern holds. If it does, address it as your own observation in a structured 1:1 next month.Reveal

Three more skip-levels confirm the pattern: another IC mentions inconsistent 1:1s, two more mention roadmap fog. You now have aggregated signal you can act on without burning any source. You have a structured 1:1 with the manager: 'I've been observing the team — I'm noticing some signs that 1:1 cadence and roadmap clarity have slipped. Walk me through how you're running the team.' She acknowledges the gaps; you build a 30-day improvement plan together. Skip-level signal stays high because no source was burned.

Skip-Level Signal: PreservedIssue Addressed: Yes (with full context)Manager Relationship: Strengthened by direct, sourced-as-yours feedback

Related concepts