xoai/opportunity-map

<!-- sage-metadata cost-tier: sonnet activation: auto tags: [product-management, opportunity, prioritization, planning, sequencing] inputs: [discovery-output] outputs: [opportunity-map] requires: [] -->

Opportunity Map

Produce a structured assessment of which discovered opportunities to pursue, based on both customer evidence (outside-in) and organizational capability (inside-out). The output is a living decision document — expected to be reviewed and updated regularly as new data arrives.

Deliverable type: document

Mode Behaviors

FIX (update): Update an existing opportunity map with new information — revised scores after research, changed decisions after a competitor move, resolved dependencies after shipping a feature. Don't re-run the full assessment. Update the affected scores/decisions, adjust sequencing if needed, and note what changed and why. Minutes.

BUILD (light): Quick assessment. Score the top 3-5 opportunities from discovery, do a lightweight capability check (existing capability? unique advantage?), make pursue/defer decisions without detailed dependency mapping or sequencing rationale. Produces a focused decision sufficient to move into a PRD. 10-15 minutes.

ARCHITECT (full): Complete assessment. All opportunities scored with capability fit, full dependency map, detailed sequencing with rationale, monitor/defer lists with signals, confidence tracking per opportunity, review plan with triggers. The comprehensive decision document. 30-45 minutes.

When to Use

• After discovery work (JTBD, research, lean canvas) has identified multiple opportunities and the team needs to decide where to focus
• At planning cycles when reassessing priorities against new data
• When the competitive landscape shifts or new capabilities emerge
• When a PM says "we have all these needs — what should we actually do?"

Prerequisites

This skill requires structured discovery output. Accepted inputs:

• JTBD analysis: Desired outcomes with opportunity scores (ideal)
• Research findings: Identified needs with importance/satisfaction signals
• Lean Canvas: Hypotheses with risk levels
• Any structured list of customer needs with some measure of priority

If the input has opportunity scores (Importance + Satisfaction → Score), use them directly. If not, Step 1 applies lightweight scoring.

If NO discovery work exists: recommend running a discovery skill first (e.g., jtbd). Opportunity mapping without customer evidence produces opinion-based prioritization — flag this prominently if the user insists.

Process

Step 0: Gather Context

Read first: references/opportunity-methodology.md

Confirm with the user in ONE message:

1. Discovery input: Which analysis is this based on? Where are the scored opportunities?
2. Capability context: What are the team's core technical capabilities? What data, infrastructure, or distribution advantages exist? What's the approximate team capacity and timeline horizon?
3. Strategic context: What direction has leadership set? Are there constraints or mandates that override pure opportunity scoring?
4. Scope: Is this for a specific feature area, an entire product, or a new product? (Affects the granularity of assessment)

If the user provides discovery output but no capability context, ask for it explicitly: "The opportunity scores tell me what customers need. To assess which opportunities YOU should pursue, I need to understand your capabilities and constraints. What does your team do well? What data or technology advantages do you have?"

Step 1: Score Opportunities

If discovery output includes opportunity scores, import them directly.

If not, walk through each identified need and apply the lightweight scoring from references/opportunity-methodology.md:

• Importance (1-10): How critical is this need to the job performer?
• Satisfaction (1-10): How well do current solutions serve this need?
• Opportunity = Importance + max(Importance - Satisfaction, 0)

Tag each opportunity's confidence level based on the evidence behind it:

• High: Validated by quantitative data (survey n≥100) or multiple qualitative sources (8+ interviews)
• Medium: Supported by limited qualitative data (3-7 interviews) or consistent indirect signals (metrics, behavioral data)
• Low: Hypothesized from context or 1-2 data points — needs validation

Present the scored landscape to the user for review before proceeding. Scores are judgment calls — the user may adjust based on context the discovery didn't capture.

Step 2: Assess Capability Fit

For each opportunity scoring ≥10, assess organizational capability:

Three questions per opportunity:

1. Existing capability: What do we already have that serves this need, even partially? (Feature, data, infrastructure, expertise)
2. Unique advantage: What do we have that competitors can't easily replicate? (Data moat, distribution, trust, technology)
3. Effort level: How much work to address this opportunity? (Quick win / Moderate / Major / Foundational)

Capture as a capability assessment table. Keep it structured but qualitative — false precision is worse than honest judgment here.

If the user hasn't provided enough capability context, this step will surface the gaps: "I can't assess capability fit for opportunity O3 because I don't know what transaction data you have access to. Can you clarify?"

Step 3: Map Dependencies

Examine the relationships between opportunities:

• Requires: O1 can't work without O3 being solved first
• Enables: Solving O2 makes O4 significantly easier or more valuable
• Shares foundation: O1 and O2 both need the same underlying capability

The job process map from JTBD (if available) provides the natural dependency structure — needs at earlier job stages often need to be addressed before needs at later stages.

Document as a simple dependency map (text or diagram). Identify:

• Any "foundation" capabilities that multiple opportunities depend on
• Any circular dependencies (these indicate the opportunities may need to be tackled as one initiative)
• Any opportunities that are independent (can be pursued in parallel)

Step 4: Decide

For each opportunity, assign one of three decisions:

Pursue (limit to 2-4 opportunities):

• High opportunity score (≥12)
• Strong or buildable capability fit
• Dependencies resolved or resolvable within the planning horizon
• For each: document WHY pursue, WHAT's our advantage, WHAT happens if we don't (the do-nothing alternative)

Monitor (track, don't commit):

• High score but weak capability fit, or
• Strong fit but moderate score, or
• Key dependency not yet resolved, or
• Confidence too low — needs validation first
• For each: document WHAT signal would change this to "pursue"

Defer (explicitly not now):

• Low score relative to alternatives, or
• Requires fundamentally new capabilities we can't build in the horizon, or
• Already well-served by competitors, or
• Strategic misalignment
• For each: document WHY deferred and WHEN to reconsider

Hard rule: If more than 4 opportunities are "pursue," push back. "You've selected [N] opportunities to pursue. With [team size/capacity], this means each gets roughly [fraction] of available resources. Is that enough to make real progress on any of them? Consider: which 2-3 would you pursue if you could only pick that many?"

Step 5: Sequence

For the "pursue" opportunities, determine order using the criteria from references/opportunity-methodology.md:

1. Dependencies first (B before A if A requires B)
2. Quick wins before bets (build momentum and learning)
3. Foundation before features (shared capabilities first)
4. Higher confidence before lower (lower risk first)
5. Higher score as tiebreaker

For each position in the sequence, explain why it's in that order. The rationale matters more than the ranking — it helps the team understand and commit to the plan.

Step 6: Define Review Plan

The opportunity map is a living document. Define:

• Next scheduled review date (typically quarterly, or end of current planning cycle)
• Event-based triggers for earlier review:
  • Research results that change confidence levels
  • Competitor launch that shifts satisfaction scores
  • Shipped feature that resolves dependencies
  • Strategic direction change from leadership
  • Team capacity change

Step 7: Quality Check

Before presenting, validate:

• [ ] Every opportunity has a score, confidence level, and decision
• [ ] No more than 4 opportunities are "pursue"
• [ ] Every "pursue" has a rationale including do-nothing alternative
• [ ] Every "monitor" has a specific trigger signal
• [ ] Every "defer" has a rationale and reconsider condition
• [ ] Dependencies are mapped and reflected in sequencing
• [ ] Capability context from the user informed the assessment
• [ ] Confidence levels are tagged per opportunity with validation needs
• [ ] Review plan is defined with date and event triggers

Output

Save to .sage/docs/opportunity-map.md using the template from templates/opportunity-map-template.md.

Append decisions to .sage/decisions.md recording pursue/ monitor/defer decisions, key rationale, and sequencing. Update the "Current Artifacts" section.

Present to user: "Here's the opportunity map. I recommend pursuing [N] opportunities: [X, Y, Z]. [X] comes first because [rationale]. [M] opportunities are on the monitor list — the key signal to watch is [signal]. Want to discuss any of the decisions?"

Rules

MUST:

• MUST read references/opportunity-methodology.md before starting
• MUST have structured discovery output as input — mapping without customer evidence produces opinion-based prioritization
• MUST gather capability context from the user — opportunity scores alone don't tell you what to pursue (outside-in without inside-out produces wish lists)
• MUST limit "pursue" to 2-4 opportunities — if everything is a priority, nothing is
• MUST include the do-nothing alternative for every "pursue" decision — this prevents pursuing opportunities that don't justify the cost
• MUST define a review plan — the opportunity map expires
• MUST tag confidence per opportunity and recommend validation for low-confidence items

SHOULD:

• SHOULD present the scored landscape to the user for review before making decisions (scores are judgment calls)
• SHOULD use the job process map (from JTBD) to inform dependency mapping when available
• SHOULD identify foundation capabilities that unlock multiple opportunities
• SHOULD note when an opportunity moves from "monitor" to "pursue" would require additional discovery work

MAY:

• MAY produce a lightweight map in BUILD mode (fewer opportunities, simpler assessment) when scope is a single feature area
• MAY include brief solution direction notes (1-2 sentences per opportunity) to help evaluate feasibility — but not detailed solutions
• MAY recommend specific discovery or research skills to run for low-confidence opportunities

Failure Modes

• No capability context provided: Don't produce a map with only outside-in scores. Ask for capability context first. If the user can't provide it, the map will note "capability fit unassessed" for all opportunities and recommend the user complete this assessment with their team.

• All opportunities score high: This usually means the satisfaction scores are too uniformly low (common when the product is new or the category is immature). Differentiate by asking: "Among these, which need do users complain about MOST vocally?" or "Which of these, if solved, would be the strongest reason to choose your product?"

• User wants to pursue everything: This is the most common failure mode in real PM work. Push back with: "Pursuing [N] opportunities with [capacity] means each gets [fraction] of focus. History suggests this produces mediocre progress on everything rather than breakthrough progress on anything. Which 2-3 would move the needle most?"

• Discovery input is a feature list, not needs: If the user provides "we want to build: notification system, dashboard, AI categorization" instead of customer needs, redirect: "These sound like solutions. Let's back up — what customer needs do these address? Running a JTBD analysis would give us the needs to map these solutions against."

• Opportunity map becomes a one-time artifact: Remind the user that the map is designed to be reviewed regularly. "This map reflects what we know today. I've set a review trigger for [date/event]. When new data comes in, we should revisit the scores and decisions."