derisk-measurement-advisor

## Purpose

Guide product managers through identifying the most important things to measure, test, or track across 10 risk dimensions — 4 internal and 6 external — then triage each into **act on immediately** or **start tracking**. The output is a prioritized risk register that tells you what to do first, not just what could go wrong.

This is not a generic risk list. It's a guided scan tied to *your* specific product, customer, and market context. The question it answers: **is the orange worth the squeeze?**

## Input

**Works best with:** The product or AI idea you want to stress-test.
**Also useful:** Its stage (concept, prototype, in-market), what you've already validated, and any dimensions you already know are risky.

Anything supplied with the invocation itself — text after the skill name, a pasted context dump, or an appended `ARGUMENTS:` line — counts as answers already given. Use it and skip whatever it covers; don't re-ask.

**Arriving empty-handed? That works too.** The advisor opens by asking for the idea in a sentence or two, then scans the 10 risk dimensions against it.

**Example invocation:** `De-risk this: an AI agent that auto-drafts insurance claim responses for adjusters — prototype stage, no compliance review yet.`

## Key Concepts

### Two Lenses, One Question

Every product idea carries two kinds of risk. Most PMs check one and ignore the other.

**Internal risks** ask whether the product itself will work:
- **Product Outcome Metrics** (Desirability + Usability) = "What's in it for the customer/user?"
- **Business Outcome Metrics** (Feasibility + Viability) = "What's in it for our business/organization?"

**External risks** ask whether the world will let it work:
- **PESTEL forces** (Political, Economic, Social, Technological, Environmental, Legal) = "What could blindside us?"

The single triage question that applies to both sides: **Do I act on this now, or start tracking it?**

---

### The Internal Risk Framework (DUFV)

Adapted from Marty Cagan's four product risks, split into two measurement categories:

| Category | Dimension | Core Question |
|----------|-----------|---------------|
| **Product Outcome Metrics** | Desirability | Will customers value it enough to buy it? |
| **Product Outcome Metrics** | Usability | Will customers figure out how to use it? |
| **Business Outcome Metrics** | Feasibility | Can we build it? Can we sustain it at scale? |
| **Business Outcome Metrics** | Viability | Will it work as a viable business? |

**Product Outcome Metrics** answer: *What's in it for the customer/user?* They measure whether the product is solving meaningful problems and delivering value to its intended users.

**Business Outcome Metrics** answer: *What's in it for our business/organization?* They measure how effectively and reliably the product supports a seamless customer experience while remaining feasible to deliver.

The connection to positioning: Product Outcome Metrics validate your **Problem Positioning Statement** (the customer promise). Business Outcome Metrics validate your **Problem Framing Statement** (the business case).

---

### The External Risk Framework (PESTEL)

Six categories of forces outside your organization that can reshape, delay, or kill your product plans:

| Category | What It Covers | Example Risk |
|----------|---------------|--------------|
| **Political** | Government policies, trade regulations, political priorities | New AI safety rules for state contractors |
| **Economic** | Market conditions, funding, inequality, supply chains | VC AI funding dries up — focus shifts to profitability |
| **Social** | Demographics, sentiment, digital literacy, cultural shifts | AI accused of bias — can you prove fairness? |
| **Technological** | Platform dependencies, standards, competing tech | New AI model outdates yours — time for rapid innovation? |
| **Environmental** | Energy consumption, sustainability, carbon footprint | AI criticized for carbon footprint — how to go green? |
| **Legal** | Data privacy, compliance, IP, evolving regulations | AI deepfakes lawsuit crossfire — can you protect your product? |

**Key insight from the classroom:** Market conditions don't just change — they potentially change us, our business, and our products. PESTEL helps PMs categorize and respond to that change, not just observe it.

---

### The Act/Watch Triage

Every surfaced risk gets one of two labels:

- **Act on immediately:** This risk is present, material, and requires a response now — run a test, change a plan, mitigate a gap.
- **Start tracking:** This risk is real but not yet urgent — monitor it, set a trigger threshold, revisit on a cadence.

The triage applies equally to internal and external risks. A desirability gap you can't answer is an "act now" just as much as a regulatory deadline. A demographic trend you see coming is a "watch" just like a feasibility concern you've already partially mitigated.

**Anti-pattern:** Labeling everything "act now." If every risk is urgent, none of them are prioritized. Force yourself to split roughly 40/60 act/watch.

---

### When to Use This Skill

**Use this when:**
- You have a product or AI idea and need to decide whether to commit
- You're about to put something on the roadmap and want a structured risk check
- You've been focused on internal risks and haven't scanned external forces (or vice versa)
- Leadership asks "what could go wrong?" and you want a better answer than "nothing, we're good"

**Don't use this when:**
- You don't have a specific product idea yet (use `problem-framing-canvas` first)
- You need a deep PESTEL analysis with data and citations (use `pestel-analysis` component skill)
- You've already validated the idea and need to pick a validation method (use `pol-probe-advisor`)
- The idea is already in production and you're optimizing, not de-risking

---

### Facilitation Source of Truth

Use [`workshop-facilitation`](../workshop-facilitation/SKILL.md) as the default interactio