| name | north-star-selection |
| description | Use this skill when the user asks "which metric should we focus on", "how do I choose between these metrics", "what's the best metric to track", "help me select our primary metric", "our metrics are confusing", "we have too many metrics", or wants to select a primary North Star from a set of competing metrics. This is the selection and evaluation skill; for defining and setting a North Star from scratch, use strategy/north-star. |
| version | 2.0.0 |
North Star Selection
You are helping the user choose the right North Star Metric from a set of candidates. This is different from defining a North Star from scratch — here, the user has options and needs a framework for deciding.
Framework: Lenny Rachitsky (North Star guide, survey of 40+ top companies), Sean Ellis (North Star selection criteria), Dan Olsen (prioritizing metrics).
Step 1 — Load Context
Read memory/user-profile.md for product stage, business model, and current metrics. Read context/company/analytics-baseline.md for available metrics.
Step 2 — Gather the Candidates
Ask: what metrics are being considered? For each candidate metric, capture:
- Name and definition
- How it's currently measured
- Who has access to this data
Step 3 — Evaluate Against NSM Criteria
Score each candidate on the 5 NSM criteria (1–5 each):
1. Represents value: Does it go up when users genuinely get value, not just when they're active?
2. Predicts revenue: Is it a leading indicator of long-term revenue or retention? (Revenue itself is a lagging indicator — avoid as a primary NSM)
3. Measurable: Can you track it accurately with current infrastructure? Any data quality concerns?
4. Actionable: Can the product team run experiments that meaningfully move this metric?
5. Understandable: Can any team member immediately understand what this metric means? (Avoid complex composite metrics as NSM)
Score each candidate across all 5. The highest total score is the strongest candidate.
Step 4 — The Vanity vs. Value Test
Apply the most important filter: does this metric go up even when users aren't getting value?
- Total signups: Goes up with any acquisition — users could sign up and never use the product
- Total sessions: Goes up if users open the app and immediately give up
- Total pageviews: Goes up with any traffic, regardless of engagement
Eliminate vanity metrics from consideration.
Step 5 — North Star for Your Business Model
Cross-reference with the known best-practice NSM by business model (from memory context):
SaaS: "Weekly active users who complete [core action]" — captures genuine engagement, not just logins
Marketplace: "Completed transactions with positive outcome for both sides" — prevents optimizing for volume at the expense of quality
Consumer: "DAU/MAU ratio" — measures habit formation, not just reach
Usage-based: "Value-generating actions per period" — not all usage, specifically value-generating
Freemium: "Users who hit the activation threshold" — the specific action that predicts conversion
Step 6 — Recommendation
Present:
- Ranked candidates with scores and rationale
- Recommendation with explanation
- Why the top alternative was rejected
- Any supporting metrics that should be tracked alongside the NSM
- One metric to explicitly exclude (and why — prevents gaming)
Step 7 — Output
Produce:
- Evaluation table (all candidates × 5 criteria × scores)
- Final recommendation with rationale
- Supporting metric suite (3–4 metrics that indicate whether the NSM will move)
- Implementation note: if the recommended NSM isn't currently being tracked, what would it take to start?
