What If

What-If Simulation

Simulate the impact of a proposed product decision by reasoning over accumulated organizational context — strategy, past decisions, customer feedback, and execution history. This is the "world model" layer: if the system can answer "what if" grounded in real context, it's more than a search index.

When to Use

• Evaluating a proposed prioritization change ("What if we deprioritize mobile?")
• Assessing a strategic pivot ("What if we shift ICP from SMB to enterprise?")
• Analyzing experiment patterns ("We've killed 3 pricing experiments — what does that mean?")
• Trade-off decisions ("Should we invest in feature A or feature B?")
• Risk assessment ("What's the blast radius of cutting this team's scope?")
• When asked "what if", "what would happen", "simulate", "impact analysis", or "trade-off analysis"

Process

1. Clarify the Hypothetical

Ask if not clear from context:

• Proposed action: What change is being considered? (1 sentence)
• Scope: What does it affect? (product, team, segment, timeline)
• Constraint: Is this reversible? What's the time horizon?

2. Load Context

Read the following to ground the simulation in actual organizational context:

• knowledge/ (strategy, segments, metrics, competitive landscape) — the baseline "world model"
• memory/decisions/ — past decision traces for precedent and pattern
• memory/feedback/ — customer signals that would be affected
• memory/weekly-plans/ — current execution state and momentum
• memory/strategy-reviews/ — alignment patterns and recurring gaps

3. Run the Simulation

For the proposed action, analyze:

A. Strategic Impact

• Which strategic pillars does this strengthen? Which does it weaken?
• Does this move toward or away from stated goals?
• Precedent: have similar decisions been made before? What happened?

B. Customer Impact

• Which customer segments are affected? (reference knowledge/customer-segments.md)
• Does customer feedback data support or contradict this direction?
• What customer pain points get addressed vs. neglected?

C. Metric Impact

• Which key metrics move? In which direction?
• Are there second-order effects? (e.g., deprioritizing mobile affects engagement, which affects retention)
• What guardrail metrics might regress?

D. Execution Impact

• What current work gets affected? (reference recent plans)
• What dependencies break or change?
• What new work does this create?

E. Competitive Impact

• Does this open or close gaps with competitors? (reference knowledge/competitive-landscape.md)
• Does this create differentiation or reduce it?

F. Risk Assessment

• What's the worst-case outcome?
• How reversible is this decision?
• What early signals would indicate this was the wrong call?

4. Surface the Trade-offs

Frame the analysis as explicit trade-offs:

• You gain: [what improves]
• You lose: [what degrades]
• You risk: [what could go wrong]
• You assume: [what must be true for this to work]

5. Recommend

Based on the accumulated context, provide:

• A clear recommendation (proceed / modify / reconsider)
• Conditions under which the recommendation changes
• Suggested experiment or validation step if uncertainty is high

Output

## What-If Analysis: [Proposed Action]

**Date:** YYYY-MM-DD
**Proposed action:** [1 sentence]

### Strategic Impact
- [Pillars strengthened/weakened]
- [Goal alignment assessment]
- [Precedent from decision history]

### Customer Impact
- [Segments affected]
- [Feedback alignment]

### Metric Impact
| Metric | Expected Direction | Confidence | Rationale |
|--------|--------------------|------------|-----------|
| [metric] | ↑/↓/→ | High/Med/Low | [why] |

### Execution Impact
- [Current work affected]
- [Dependencies]

### Competitive Impact
- [Gap analysis]

### Trade-offs
| You Gain | You Lose | You Risk | You Assume |
|----------|----------|----------|------------|
| [gain] | [loss] | [risk] | [assumption] |

### Recommendation
**[Proceed / Modify / Reconsider]**
- [Rationale grounded in context]
- [Conditions that would change this recommendation]
- [Suggested validation step]

The strength of this analysis depends on the depth of accumulated context in memory/. With sparse memory, flag low confidence and recommend building more context before committing.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: hardiktiwari
• Source: hardiktiwari/PM-operating-OS
• License: MIT

Install and usage instructions live in the source repository linked above.