Pmg Pick North Star Metric

---
name: pmg-pick-north-star-metric
description: Builds a north star metric tree with the north star, three to five named input metrics with definitions, data sources and owning teams, and a counter-metric for each input. Use for "run pmg-pick-north-star-metric", "pick our north star metric", "north star framework", "which number are we moving", "input metrics", "metric tree", "we ship a lot but cannot show impact", part of The Complete Claude Guide for Product Managers Pack by Polar Bear.
---

# Pick the North Star Metric

## When To Use
The team ships a lot and cannot say which number it is moving, so every dashboard review becomes a debate about which chart matters. Use this when you have a strategy and need one product-level number that captures customer value, plus the few inputs teams can actually move.

## When Not To Use
If the question is whether one release worked, run Set the Success Metrics; the north star moves too slowly to judge a single feature. If there is no strategy yet, a north star just freezes today's habits into a number, so run Write the Product Strategy first.

## Inputs
- The strategy kernel or vision, and the moment a customer gets value from the product
- The metrics you track today, with where each comes from
- Candidate north stars already floated by the team or leadership
If you have none of this, I start from a description of the value moment and mark the output as a first draft. With the Amplitude, Mixpanel, PostHog or Pendo connector on, I read event names and definitions from your analytics instead of guessing them.

## Approach
I use the North Star Framework as described in Amplitude's North Star Playbook: one output metric that reflects customer value, moved through a handful of named inputs the team can act on. A good north star passes three tests: it expresses the value customers get, it represents the product strategy, and it leads business results rather than trailing them. The failure it prevents is picking registered users or revenue, which climb while customers quietly stop getting value.

## Workflow
1. Ask up to three questions: what the core value moment is for the customer, which data sources are trustworthy today, and who approves the north star.
2. List two or three candidates. Rate each against the three tests (customer value, represents the strategy, leading indicator) as pass, partial or fail, with a reason. Reject revenue and vanity counts, such as registered users or daily actives with no value action, and say why.
3. Write the chosen candidate's exact definition: what counts, the time window, what is excluded.
4. Break it into three to five inputs the team can act on. Breadth (how many customers), depth (how much they engage), frequency (how often) and efficiency (how fast they get value) is a starting heuristic, not a rule.
5. For each input, write the definition, the data source and the owning team (a team, never a person). Flag inputs with no reliable data as "not yet measurable".
6. Add one counter-metric per input so gaming it does not hurt customers (for example, [activation speed] guarded by [support contacts per new account]). The user sets each threshold.

## Output Format
```markdown
# North Star Metric Tree
## Candidates
| Candidate | Customer value | Represents strategy | Leading indicator | Verdict |
|---|---|---|---|---|
| [metric] | [pass / partial / fail: reason] | [...] | [...] | [chosen / rejected] |
## North star
[Name]: [exact definition, time window, exclusions]
## Inputs
| Input | Definition | Data source | Owning team | Counter-metric and threshold |
|---|---|---|---|---|
| [input] | [definition] | [source or "not yet measurable"] | [team] | [guardrail, threshold set by user] |
## Decision
[Named person] approves the north star and the inputs by [date]; [team] confirms each data source by [date].
```

## Done When
- The north star passes all three tests, or each partial is explained
- There are three to five inputs, each with a definition, a source and a team
- Every input has a counter-metric with a threshold the user set
- Revenue and vanity counts are named as rejected, with the reason

## Quality Bar
- Metrics are aggregate; no per-user or per-employee metric appears in the tree
- No invented baselines; unknown values stay as [placeholders]
- Inputs are things a team can move this quarter, not another team's outcome
- A named person approves the north star; Claude never picks it alone.

## Next
Run pmg-build-product-roadmap (Build the Product Roadmap) to place the work that moves the inputs.

## About the makers

This pack is made by Polar Bear, a consultancy built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).