borghei/Claude-Skills / project-management/career/pm-interview-prep
pm-interview-prep
VerifiedPlans PM interview prep across product sense, execution, strategy, behavioral and technical rounds, calibrated to level. Use when a PM loop is scheduled, you are running mock interviews, or you are moving up a level.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 8.4 KB |
| assets/mock_feedback_form.md | markdown | 2.3 KB |
| assets/self_assessment.md | markdown | 3.3 KB |
| assets/story_worksheet.md | markdown | 3.8 KB |
| examples/stripe-circles-new-market-product.md | markdown | 14.5 KB |
| references/framework-deep-dive.md | markdown | 11.1 KB |
| references/question-bank.md | markdown | 7.2 KB |
| references/red-flags.md | markdown | 13.1 KB |
| references/round-frameworks-and-prep.md | markdown | 10.6 KB |
references/round-frameworks-and-prep.md
View on GitHub# Round Types, Frameworks, Level Calibration & Prep Playbook
Read this when you need the five round-type reference, the full CIRCLES / AARM / STAR / Estimation frameworks, the per-level calibration tables, the prep workflow, sample question excerpts, troubleshooting, or success criteria.
## The Five Round Types
| Round | Tests | Typical Length | Primary Framework |
|-------|-------|----------------|-------------------|
| Product Sense (Design) | Customer empathy, problem framing, creative solutioning, prioritization | 45-60 min | CIRCLES (Lin) |
| Product Improvement | Metric diagnosis, root-cause analysis, solution generation | 45-60 min | AARM |
| Strategy | Market analysis, competitive positioning, go-to-market thinking | 45-60 min | Porter / 3Cs / Ansoff |
| Behavioral | Leadership, influence, conflict, impact storytelling | 45-60 min | STAR |
| Technical / Analytical | Systems thinking, data fluency, SQL, A/B test design, estimation | 45-60 min | Estimation 5-step / metric trees |
Not every loop covers all five. APM loops emphasize product sense and behavioral. Senior PM loops emphasize strategy and execution. Group PM loops emphasize cross-team strategy and people leadership.
## Framework 1: CIRCLES (Product Sense)
Lewis Lin's CIRCLES Method from *Decode and Conquer* is the canonical product design framework.
| Step | Question | Time |
|------|----------|------|
| **C**omprehend | Clarify the scope, constraints, and goals of the question. Ask 2-3 clarifying questions. | 3-5 min |
| **I**dentify the customer | Pick a specific user segment. Justify why. | 2-3 min |
| **R**eport customer needs | List 3-5 needs of the chosen segment. Categorize by JTBD or pain/gain. | 5-7 min |
| **C**ut through prioritization | Pick 1-2 top needs to solve. Justify with a criterion (impact, frequency, severity, strategic fit). | 3-5 min |
| **L**ist solutions | Generate 3-5 distinct solutions for the prioritized need. Include at least one creative/non-obvious idea. | 8-10 min |
| **E**valuate trade-offs | Compare solutions on impact, effort, risk, and strategic fit. | 5-7 min |
| **S**ummarize | State your recommendation, the why, and the next step. | 2 min |
**Common failure modes:**
- Skipping Comprehend (jumping to solutions before scope is clear)
- Picking "all users" as the segment (no segmentation = no signal of customer empathy)
- Listing 10 generic solutions instead of 3-5 distinct ones
- Forgetting to summarize (interviewer leaves without a clear answer)
## Framework 2: AARM (Product Improvement / Metric Diagnosis)
For "How would you improve X?" or "Why is metric Y down?" questions:
| Step | Action |
|------|--------|
| **A**lign on the goal | Confirm what "improve" means: revenue? engagement? retention? satisfaction? |
| **A**nalyze the current state | Map the funnel or user journey. Identify drop-off points or weak metrics. |
| **R**ecommend interventions | Propose 2-3 interventions targeting the weakest step. Tie each to a measurable outcome. |
| **M**easure success | Define the success metric, guardrail metrics, and how you would A/B test. |
AARM forces the candidate to ground recommendations in data, not opinion. It also signals senior-level instincts (guardrails, measurement) that distinguish PM from Senior PM.
## Framework 3: STAR (Behavioral)
Standard but worth re-stating because most candidates do it poorly.
| Component | Length | Common Mistake |
|-----------|--------|----------------|
| **S**ituation | 30s | Spending 3 minutes on context |
| **T**ask | 30s | Burying the "what was your job" in narrative |
| **A**ction | 2-3 min | Saying "we" instead of "I" -- interviewers cannot grade the team |
| **R**esult | 1 min | No quantified outcome; no "what I learned" |
**Senior PM and above:** Add a 5th element -- **Reflection.** What would you do differently? What did this teach you about your leadership style? This signals self-awareness.
**Prepare 8-12 stories** that cover the canonical behavioral themes:
1. Time you influenced without authority
2. Time you disagreed with your manager
3. Time you killed a feature/project
4. Time you handled a major mistake
5. Time you turned around a struggling project
6. Time you said no to a stakeholder
7. Time you missed a deadline
8. Most ambitious project / biggest impact
9. Time you worked with a difficult cross-functional partner
10. Time you mentored or developed someone (Senior+)
11. Time you built or evolved team processes (Senior+)
12. Time you set strategy that others adopted (Group+)
## Framework 4: Estimation (Market Sizing)
Five-step framework for "How many X are sold in Y per year?" questions:
1. **Clarify scope** -- Geographic boundary, time period, what counts as X.
2. **Pick an approach** -- Top-down (start from population) or bottom-up (start from units). State which and why.
3. **Decompose** -- Break the calculation into 3-5 multiplied factors. Each factor is something you can estimate within an order of magnitude.
4. **Estimate each factor** -- State your assumption and justify it briefly. Use round numbers.
5. **Sanity check** -- Compare the result to a known benchmark. State whether the answer feels high, low, or right and why.
**Example:** Yearly coffee cups sold in San Francisco.
- Scope: SF city (not Bay Area), 12 months, includes home + cafe.
- Approach: Bottom-up via population x cups-per-day.
- Decompose: Population (850K) x % coffee drinkers (60%) x cups/day (2) x 365 = ~370M cups/year.
- Sanity check: SF has ~1,500 coffee shops; if each sells 500/day = 275M shop cups/year. Plus home consumption (~30% of total). Total feels right.
## Level Calibration
Same question, different bar by level:
### Product sense ("Design a product for X")
| Level | Bar |
|-------|-----|
| APM | Picks a clear segment, lists 3 reasonable solutions, summarizes a pick |
| PM | + Articulates trade-offs, considers measurement, surfaces a non-obvious solution |
| Senior PM | + Connects to a business goal, considers downstream eng/design effort, names a v2/v3 path |
| Group PM | + Frames the opportunity strategically, identifies platform / cross-team implications, designs the operating cadence to learn fast |
### Behavioral ("Tell me about a time you...")
| Level | Bar |
|-------|-----|
| APM | Clean STAR, quantified result, owns their part |
| PM | + Shows judgment under uncertainty, surfaces what they learned |
| Senior PM | + Demonstrates influence at the org level, names trade-offs and counterfactuals |
| Group PM | + Demonstrates org-design / people leadership, story spans quarters not weeks |
### Strategy ("How would you grow product X?")
| Level | Bar |
|-------|-----|
| APM | Usually not asked this round |
| PM | Maps growth levers using a known framework (AARRR, Ansoff) |
| Senior PM | + Identifies the highest-leverage lever with reasoning, considers competitor response |
| Group PM | + Defines the operating system to run the strategy (teams, metrics, cadence, decision rights) |
## Workflow
1. **Diagnose your starting point.** Use `assets/self_assessment.md` to rate yourself across the five round types. Identify your weakest two.
2. **Build a question bank.** Use `references/question-bank.md` to pull 5-10 questions per round type targeted at your level.
3. **Practice solo first.** For each question, write a structured answer using the appropriate framework. Time yourself.
4. **Mock with a peer.** Schedule 2-3 mocks per round type. Use `assets/mock_feedback_form.md` for structured feedback.
5. **Refine stories.** Rewrite the 8-12 behavioral stories until each is under 5 minutes and quantified. Practice the opening line of each.
6. **Final week prep.** Re-read the company's product, recent press, and any teardown. Prepare 2-3 thoughtful questions to ask each interviewer.
## Question Banks (Excerpt -- see references/question-bank.md for full set)
### Product Sense
- "Design a product for a deaf person to navigate a city."
- "How would you redesign the airport experience for elderly travelers?"
- "Design a new feature for Spotify that drives daily engagement."
### Product Improvement
- "Why are Instagram Stories views down 5%?"
- "How would you improve Google Maps for commuters?"
- "How would you increase host signups on Airbnb?"
### Strategy
- "Should Netflix enter gaming?"
- "How should Shopify defend against Amazon's small-business push?"
- "What is the most important metric for Slack and why?"
### Behavioral
- See the 12-story prompt list above.
### Technical / Analytical
- "Estimate the storage cost of YouTube."
- "Design an A/B test for a new checkout flow. What is the minimum sample size?"
- "Walk me through how you would diagnose a 20% spike in API errors."
## Troubleshooting
| Symptom | Likely Cause | Resolution |
|---------|-------------|------------|
| You run out of time in product sense rounds | Spending too long in Comprehend or List Solutions | Allocate time per CIRCLES step and self-monitor; aim for solutions in 25 min, summary in 5 min |
| Interviewer keeps asking "and why?" follow-ups | Answers lack reasoning chains | Every answer should include the rationale -- "I picked X because Y, with trade-off Z"; explicitly state trade-offs unprompted |
| Behavioral stories sound generic | STAR is too high-level; "we" instead of "I"; no quantification | Rewrite to be specific to one event; replace "we" with "I"; add 1-2 numbers per story |
| You blank on the metric question | No mental model of the product's metric tree | Before the interview, sketch the product's likely metric tree (NSM + input metrics); rehearse 2-3 diagnostic patterns |
| Estimation answers are off by 100x | Skipping the sanity check step | Always compare to a known benchmark (population, GDP, daily users) before stating the final number |
| You feel under-leveled in feedback | Examples scoped too small (project-level for a Senior role) | Pick larger-scope stories that span multiple teams or quarters; for Group PM, lead with stories about org design, not feature design |
| You over-prepared frameworks but bombed | Robotic delivery; framework became scaffolding visible to interviewer | Use frameworks invisibly -- never name CIRCLES out loud; let the structure show through your answer, not your vocabulary |
## Success Criteria
- You have a structured answer template for each of the five round types
- You have 8-12 behavioral stories written, timed under 5 minutes each
- You have completed >=2 mocks per round type with structured feedback
- You can run CIRCLES end-to-end on a novel question in under 45 minutes
- You can do market sizing within 10 minutes including sanity check
- You have 2-3 thoughtful questions prepared for each interviewer
- You know the level-appropriate bar -- you are not under-scoping (sounds like an APM in a Sr PM loop) or over-scoping (sounds like a Director in a PM loop)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |