Picking a Good North Star

A north-star metric is a single number that captures how much value the product is delivering.

What a good north star looks like

1. Tied to value delivered to the user, not just the company.
2. Leading, not lagging. It should move when you do good work.
3. Controllable. Product changes should be able to move it.

Testing North-Star correlation in Python

import numpy as np
import pandas as pd

# 200 customer accounts tracking candidate leading metric vs lagging outcome
np.random.seed(42)
df = pd.DataFrame({
    "account_id": range(1, 201),
    "signup_page_views": np.random.poisson(lam=15, size=200),  # Vanity metric
    "weekly_active_collaborators": np.random.poisson(lam=4, size=200), # Leading candidate
})
# True revenue at Month 6 depends heavily on active collaborators, not page views
df["m6_arr_usd"] = (
    df["weekly_active_collaborators"] * 1200 +
    df["signup_page_views"] * 10 +
    np.random.normal(0, 500, size=200)
)

corr_matrix = df[["signup_page_views", "weekly_active_collaborators", "m6_arr_usd"]].corr()
print("--- Correlation Matrix: Candidates vs 6-Month ARR ---")
print(corr_matrix["m6_arr_usd"].round(3))

Takeaway

Validate that your candidate north star actually predicts downstream retention and revenue before committing your team's OKRs to it.