Purpose
Turn a fuzzy supply-vs-demand decision into a structured tradeoff evaluation — mapping the immediate effect on each side, tracing second-order impacts on liquidity, quality, and trust, and returning a directional recommendation with guardrail metrics and a rollback trigger. Prevents locally-optimal decisions that quietly degrade the other side of the marketplace.
When to Use
Invoke this skill when the user asks to:
- Decide whether to prioritize supply-side or demand-side needs for a specific change
- Evaluate a pricing, take-rate, subsidy, or incentive change and its marketplace-wide effects
- Assess a matching, ranking, or search-relevance change that shifts who wins allocation
- Resolve a conflict where fixing one side's problem appears to harm the other
- Pressure-test a growth lever before rollout to surface second-order health risks
- Compare 2–3 competing options for the same marketplace problem
Inputs Required
Before starting, collect:
- Proposed change: The specific lever or decision being evaluated (one clear sentence)
- Marketplace type: Category, whether transactional or lead-gen, and any structural notes (e.g., managed vs. open, single-homing vs. multi-homing suppliers)
- Which side is constrained: Is supply or demand the current bottleneck, and why
- Primary objective: The metric this change is meant to move (GMV, take rate, retention, match rate, etc.)
- Current health signals: Any available numbers — liquidity/fill rate, supplier churn, buyer repeat rate, NPS, cancellation or dispute rate
- Constraints: Budget for subsidies, regulatory limits, brand commitments, or non-negotiables
- Time horizon: Whether the decision is judged on next-quarter or next-year outcomes
If health signals are missing, proceed using directional assumptions and flag each assumption explicitly.
Instructions
Step 1: Frame the Tradeoff Precisely
Restate the change as an explicit tradeoff, not a neutral feature.
- Name what each side gains and gives up: "This gives demand [X] at the cost of supply [Y]."
- Identify the constrained side and confirm the change either relieves or pressures that constraint.
- Flag if the change is actually a false tradeoff (a Pareto improvement) — if so, say so and skip to Step 6 with a lighter analysis.
Step 2: Map First-Order Effects on Each Side
For supply and demand separately, state the direct, intended effect:
- Behavior change: What each side does differently (list more, browse more, churn, wait longer)
- Economic change: How the payout, price, or fee each side faces moves
- Perceived-fairness change: Whether the change feels better or worse to that side Rate each first-order effect as Strong Positive / Mild Positive / Neutral / Mild Negative / Strong Negative for that side.
Step 3: Trace Downstream Impacts Across the Four Health Vectors
Follow the causal chain from first-order effects into marketplace health. For each vector, state the likely direction and the mechanism:
- Liquidity: Does the change deepen or thin the pool that gets matched? (fill rate, time-to-match, search-to-fill ratio). Watch for the classic trap where a demand win reduces supply willingness to participate.
- Quality: Does it raise or lower the caliber of supply that wins allocation, or the caliber of demand supply is exposed to? (rating distribution, completion rate, repeat quality)
- Trust & Fairness: Does it change who wins and whether the losing side sees the system as fair? (dispute rate, cancellation rate, perceived favoritism, gaming risk)
- Unit Economics: Does contribution margin per transaction improve or erode after accounting for subsidies, take rate, and support/dispute cost?
For each vector, mark the impact as ↑ Improves / → Neutral / ↓ Degrades and note the time lag (immediate vs. delayed).
Step 4: Identify Second-Order and Feedback Effects
Look one loop further out — the effects that surface after both sides react:
- Cross-side flywheel effects: Does the change strengthen or weaken the loop where more supply attracts demand and vice versa?
- Adaptation/gaming: How will participants change behavior to exploit or evade the change once they understand it?
- Selection effects: Does it change the mix of who joins or leaves (adverse selection of low-quality supply, cherry-picking by high-value demand)?
- Disintermediation/multi-homing risk: Does it push either side to transact off-platform or lean harder on a competitor? Rank these second-order risks High / Medium / Low by likelihood × severity.
Step 5: Score the Tradeoff
Summarize into a compact scorecard:
- Net first-order effect per side (from Step 2)
- Health-vector impact table (from Step 3): Liquidity / Quality / Trust / Unit Economics
- Highest-ranked second-order risk (from Step 4)
- Directional verdict on the primary objective: likely moves it, likely neutral, or likely moves it at unacceptable health cost
Step 6: Recommendation, Guardrails, and Rollback
Deliver an opinionated call:
- Recommendation: Proceed / Proceed with modification / Do not proceed / Run as time-boxed test
- If proceed: State the single most important modification to reduce the worst second-order risk
- Guardrail metrics: 3–5 metrics to watch, each with a direction and a threshold that means "the other side is being harmed"
- Rollback trigger: The specific condition — a named metric crossing a named threshold within a named window — that should halt or reverse the change
- Sequencing note: If both sides are affected, which side to protect first and why
Output Format
- Structure: Six labeled sections following the steps; Step 3 rendered as a 4-row health-vector table; Step 5 as a compact scorecard
- Length: 700–1,000 words
- Format: Structured document — short prose per section plus tables and rated bullets, not long essays
- Sections: Tradeoff Frame → First-Order Effects (per side) → Health Vector Impact Table → Second-Order Risks (ranked) → Scorecard → Recommendation & Guardrails
- Tone: Marketplace strategy director — decisive, mechanism-driven, willing to say no; every impact claim names the causal mechanism, not just a direction
Quality Checklist
Before delivering output, verify:
- [ ] The tradeoff is stated as an explicit give-and-take, naming what each side gains and loses
- [ ] Every one of the four health vectors (Liquidity, Quality, Trust, Unit Economics) has a direction and a stated mechanism — none skipped
- [ ] Second-order risks are ranked High/Medium/Low, not just listed
- [ ] At least one gaming or adaptation effect is identified
- [ ] Guardrail metrics include a numeric or directional threshold, not just a metric name
- [ ] A specific rollback trigger names a metric, a threshold, and a time window
- [ ] The final recommendation is a clear verdict, not a hedge
Notes
- The most common failure this tool catches is a demand-side win that quietly suppresses supply participation — always test the constrained side's willingness to keep showing up.
- Treat "improves the primary objective" and "improves marketplace health" as separate verdicts; a change can win the metric and still be the wrong call.
- When two options are being compared, run each through Steps 1–5 and reserve Step 6 for a head-to-head recommendation.
- If the change is reversible and cheap to instrument, bias toward "run as time-boxed test" with the guardrails from Step 6 as the test's kill criteria.
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills