north-star-metric

What this skill does

# North Star Metric (NSM) Expert

## Overview

A North Star Metric (NSM) is the single number that best represents the value your product delivers to its customers. Sean Ellis popularized the framing; Amplitude codified the playbook; Lean Analytics calls a related concept the "One Metric That Matters" (OMTM). The NSM is **one** number, not a dashboard. Its job is to align the entire team -- engineering, marketing, sales, support -- on a shared definition of "we won this quarter."

This skill produces a complete NSM specification: the NSM itself, 3-5 **input metrics** that the team can directly influence, the **leading indicators** that move days or weeks before the inputs do, the **anti-metrics** (things that must NOT move in the wrong direction), and **counter-metrics** that guard against gaming the NSM. The Python tool (`metric_tree_builder.py`) emits the spec as JSON, Markdown, or a Mermaid tree diagram suitable for a README or a Confluence page.

This is the first artifact a team should produce after defining its strategy and before writing OKRs. Once the NSM is set, OKRs typically map directly to moving the input metrics, and roadmaps justify themselves by which input metric they target.

### When to Use

- **New product or new strategic direction** -- before OKRs, before roadmap, define the NSM and inputs.
- **Strategic re-alignment** -- the existing dashboard has 47 metrics and no one knows which to optimize.
- **Cross-functional friction** -- marketing optimizes acquisition, product optimizes activation, neither agrees on success.
- **Investor or board reporting** -- the NSM becomes the headline metric in monthly/quarterly reviews.
- **A/B experimentation guardrails** -- every test reports the NSM impact plus the counter-metric impact.

### When NOT to Use

- For very early-stage discovery (use `discovery/brainstorm-ideas/` and `discovery/identify-assumptions/` first; you do not yet know what value you deliver).
- For pure technical infrastructure work where the user-value chain is too indirect.
- When the organization has not aligned on strategy -- the NSM exposes strategic disagreement but does not resolve it.

## What makes a good North Star Metric

A defensible NSM passes five tests:

| Test | Question | If it fails... |
|------|----------|---------------|
| **Customer value** | Does this number going up mean a customer got more value? | Pick a different metric (e.g., move from "MAUs" to "weekly active accounts that completed a task") |
| **Strategic alignment** | Does it match how the company makes money? | Pick a metric closer to monetization (e.g., "paid weekly active" not "weekly active") |
| **Leading, not lagging** | Does it predict long-term success, not just measure last quarter? | Move upstream: revenue lags; activation leads |
| **Single number** | Can it be a single number on a dashboard? | Compress: "DAU/MAU ratio" is one number; "DAU and MAU separately" is two |
| **Movable** | Can the team affect it within a quarter? | Decompose into input metrics the team can move |

### NSM archetypes (Amplitude framework)

| Archetype | What value is | NSM example | Used by (kind of company) |
|-----------|---------------|-------------|---------------------------|
| **Attention** | Time spent or engagement depth | Time spent reading, video minutes watched | Media, social networks |
| **Transaction** | Purchases or trades | Gross merchandise volume, rides booked | Marketplaces, e-commerce |
| **Productivity** | Tasks completed | Weekly active teams creating >=1 deliverable | SaaS productivity tools |
| **Communication** | Messages or interactions | Messages sent per WAU | Messaging, social |
| **Subscriber** | Recurring revenue retention | Paid weekly active subscribers | Subscription products |

Pick the archetype that matches your business model. Mixing archetypes (e.g., chasing both "messages sent" and "subscribers") produces incoherent priorities.

### NSM examples (real companies)

| Company | NSM (publicly discussed) |
|---------|--------------------------|
| Airbnb | Nights booked |
| Spotify | Time spent listening |
| Slack | Active accounts sending >=2000 messages |
| Facebook (historically) | Daily active users with 7 friends in 10 days |
| Quora | Number of questions answered |

Each NSM directly represents the value the customer is getting -- not what the company is doing.

## The metric tree

An NSM alone is not actionable. It must decompose into **input metrics** the team can directly move. The metric tree formalizes this decomposition.

### Structure

```
                     North Star Metric
                            |
        ----------------------------------------
        |             |             |          |
    Input #1     Input #2     Input #3    Input #4
        |             |
   --------     --------
   |      |     |      |
 Lead   Lead  Lead   Lead
```

- **NSM at the top** -- the single number
- **Input metrics (3-5)** -- the levers the team can pull. These multiply or sum to the NSM
- **Leading indicators (2-3 per input)** -- the earliest measurable signals that an input is moving

### Input metrics: the math matters

Input metrics should have an explicit mathematical relationship to the NSM. The most common patterns:

| Pattern | Example |
|---------|---------|
| **Multiplicative** | Active users x Sessions per user x Actions per session = Total actions (NSM) |
| **Additive** | Web signups + Mobile signups + API signups = Total signups (NSM) |
| **Funnel** | Visitors -> Signups -> Activated -> Retained (each step is an input that gates the next) |
| **Ratio** | Engaged users / Total users = Engagement rate (NSM); inputs are engagement events and user counts |

Write the formula explicitly: `NSM = Input_1 * Input_2 * Input_3`. If you cannot write the formula, the inputs are not really inputs -- they are loosely related metrics.

### Leading indicators

A leading indicator is a metric that moves **before** an input metric does. It gives the team an early-warning signal.

| Input metric | Lagging signal | Leading indicator |
|--------------|----------------|-------------------|
| Weekly retention | Measured at day 7 | Day-1 retention; first-day session count |
| New paid users | Conversion at day 30 of trial | Day-3 activation; trial usage depth |
| Support ticket resolution | Average days to close | First-response time; auto-deflect rate |
| Revenue per account | Quarterly billing | Expansion features adopted; seats added |

Leading indicators are the team's daily/weekly dashboard. Inputs are the monthly review. The NSM is the quarterly headline.

## Anti-metrics and counter-metrics

The NSM is a powerful lever -- which means it can be gamed. Two guardrails prevent this.

### Anti-metrics

An **anti-metric** is a number that must NOT move (or must move in a specific bounded direction) while the NSM moves up. They protect customer outcomes that the NSM might inadvertently degrade.

Examples:

- **NSM: video minutes watched** -- Anti-metrics: complaint rate, time-to-skip, churn within 30 days of first watch
- **NSM: messages sent per WAU** -- Anti-metrics: messages reported as spam, abuse reports, blocked users
- **NSM: nights booked** -- Anti-metrics: cancellation rate, host complaints, "would not book again" rate

Set explicit thresholds. "Churn must not exceed 4%" is enforceable; "churn must stay low" is not.

### Counter-metrics

A **counter-metric** is the metric the team owns ensuring does NOT degrade because the NSM is chasing the wrong thing. Where anti-metrics protect the customer, counter-metrics protect the business.

Examples:

- **NSM: signups** -- Counter-metric: paid conversion rate (signups are useless if they do not convert)
- **NSM: time on site** -- Counter-metric: revenue per visit (time spent without purchase is loss)
- **NSM: messages sent** -- Counter-metric: server costs per message (engagement at infinite cost is bankruptcy)

Every NSM should ship with at least 2 anti-metrics and