# Case: a north star for Halves

> Pick and defend a north star metric for Halves, with six months of numbers behind it.

- Course: [Product Analytics: From Data to Decisions](https://productovi.com/courses/product-analytics.md)
- Module: Which numbers should we steer by?
- Lesson 8/32, ~41 min
- Source: https://productovi.com/courses/product-analytics/choosing-a-north-star

## A goal for the second half

**Alex Tutor:** Goal-setting time at Halves. Maya wants a recommendation from you, the product team's numbers person.

**Maya:** Next week the board sets goals for July to December. The CEO wants MAU as the company goal: “it grew 58% since January”. Theo wants sign-ups. I need a recommendation, with numbers.

![Three people at a meeting table each hold up a different card: a crowd, a blank sign-up form and a small group of friends.](https://productovi.com/courses/product-analytics/inline/choosing-a-north-star-illustration-2.webp)

*Three candidates for the company goal and one board meeting to choose.*

**Alex Tutor:** Then let's put the candidates side by side. Here are six months of monthly numbers. `weekly_active_groups` is the average per week in that month: groups where at least two members added or settled an expense.

**Dataset (CSV):** [Halves monthly metrics, January–June 2026](https://productovi.com/datasets/product-analytics/choosing-a-north-star/halves-candidate-metrics-h1.csv)

Staff accounts excluded. Paid social campaigns started in April.

- `month`: Calendar month
- `new_signups`: New accounts created
- `mau`: Distinct users who opened the app in the month
- `app_opens`: Total `app_opened` events
- `expenses_by_people`: Expenses added by people (recurring bills excluded)
- `weekly_active_groups`: Average per week: groups with at least two members who added or settled an expense
- `plus_subscribers`: Paying Halves Plus subscribers at the end of the month

**Sofia:** I'll start with the CEO's number. Growth from January to June is June divided by January, minus one.

**Alex Tutor:** Check it rather than trusting the slide.

**Maya:** Please do. If the board catches one wrong number, nobody listens to the rest of the deck.

> **Exercise:** By how much did **MAU** grow from January to June 2026, in %? One decimal.

## Reach grew, sharing didn't keep up

**Alex Tutor:** The CEO is right about MAU. Now the whole row: sign-ups +119%, MAU +58%, app opens +48%. Those are reach. On the value side, expenses added by people grew 14.8%.

**Theo:** And weekly active groups? That's the one Alex keeps pushing.

**Alex Tutor:** Compute it the same way. Then we'll put the two growth rates next to each other.

> **Exercise:** By how much did **weekly active groups** grow from January to June, in %? One decimal.

**Sofia:** So MAU grew about five times faster than the groups that actually share costs.

**Alex Tutor:** Put them in one ratio: weekly active groups per 1,000 MAU. January: 1,480 ÷ 9,800 × 1,000 = 151. June: 107. For every thousand monthly users, fewer groups are sharing costs.

![From January to June reach metrics grew 48–119%, while weekly active groups grew only 11.8%: fewer groups share costs per 1,000 monthly users.](https://productovi.com/courses/product-analytics/diagrams/m2-l4-growth-by-candidate.png)

**Theo:** My campaigns added 5,700 monthly users since January. That's not nothing.

**Alex Tutor:** It isn't, and it's real reach. Over the same months, weekly active groups grew by 175. The question is why so few of those users ended up in working groups.

**Alex Tutor:** Try the same normalizing trick on another value number: expenses added by people, per monthly user.

> **Exercise:** Compute **expenses added by people per MAU** for June 2026 (`expenses_by_people` ÷ `mau`). Two decimals.

![A broad stream of people pours into a square, but only a few reach the tables where friends split receipts together.](https://productovi.com/courses/product-analytics/inline/choosing-a-north-star-illustration-3.webp)

*Reach grew fast; groups that actually share costs grew far slower.*

**Maya:** Let's be careful before blaming anyone. Theo's campaigns have brought us great users before.

> **Exercise:** Weekly active groups per 1,000 MAU fell from 151 in January to 107 in June, while paid campaigns ran from April. What is the best **first hypothesis** to check?

## Testing each candidate

**Alex Tutor:** Now run every candidate through the five north-star checks: can't rise while users lose, tracks value, leads revenue, splits into inputs, has a written definition.

> **Exercise:** Match each rejected candidate to its main weakness as the Halves north star.

**Maya:** Plus subscribers grew 21.5%. That's between MAU and groups. Doesn't that make it a fair compromise?

**Alex Tutor:** Before I answer Maya, what do you think?

> **Exercise:** **True or false:** Plus subscribers grew 21.5%, between MAU (+58.2%) and weekly active groups (+11.8%), so they make a fair compromise north star.

**Alex Tutor:** Plus follows active groups with a delay: people pay for receipt scanning when they share lots of costs. It's the right number to check monthly, next to the north star. Steering by it invites paywalls in the way of sharing.

**Sofia:** So every candidate is useful for something. The question is which one the company steers by.

## Defending it in the room

**Maya:** The CEO will say: “Groups grew 12%, MAU grew 58%. Why pick the goal that looks worse?”

**Alex Tutor:** Because it's the one that can't be bought. A MAU goal can be hit by doubling the ad budget while groups stall. Weekly active groups only rise when people actually split costs together.

![A pump fed with banknotes inflates a big balloon of people, while beside it a small plant of friend groups grows only when watered.](https://productovi.com/courses/product-analytics/inline/choosing-a-north-star-illustration-4.webp)

*A good goal rises only when people get value, not when the budget does.*

**Sofia:** And MAU doesn't disappear. It stays on the dashboard as a volume metric, next to the north star.

**Alex Tutor:** Two versions of the goal line made it into Maya's draft. Which one goes to the board?

> **Exercise:** Which goal for July–December would you put in front of the board?

**Alex Tutor:** Every recommendation needs one honest line about its limits. Which sentence says what this monthly table can and can't show?

> **Exercise:** Which sentence belongs in the recommendation as the limit of this analysis?

**Maya:** Give me the short version I can say out loud when the CEO pushes back.

> **Exercise:** The CEO says: “MAU grew 58% since January. Let's make it the company goal.” Reply in 3–4 sentences with numbers from the table.

## The recommendation

**Maya:** Now the one-pager I'll bring to the board: which north star, how it's defined, what feeds it, what we watch next to it and the numbers that make the case.

**Alex Tutor:** And one honest line about what the table can't tell us. These are monthly totals, so they don't show which users came from the campaigns.

**Sofia:** Got it: a pick, a definition, a small tree, a guardrail, the numbers and one limit.

> **Exercise:** Write Maya's one-page north star recommendation for the board.

> **Key ideas**
>
> - Put candidates side by side over the same months. January–June: sign-ups +119.4%, MAU +58.2%, opens +47.9%, expenses +14.8%, weekly active groups +11.8%, Plus +21.5%.
> - Normalize to see what reach turned into: weekly active groups per 1,000 MAU fell from 151 to 107.
> - Test each candidate: reflects value, leads revenue, movable through inputs, hard to push up without helping anyone.
> - A good north star can look worse than a vanity goal. That's why it can't be bought.
> - Keep volume metrics and revenue on the dashboard as context and guardrails.
> - Monthly totals can't say which users came from the campaign: check by cohort or channel before blaming it.

**Alex Tutor:** The board still gets MAU. It just stops being the goal the whole company steers by.

- Previous lesson: [DAU, WAU, MAU and stickiness](https://productovi.com/courses/product-analytics/active-users.md)
- Next lesson: [Funnels: conversion window, step order and who counts](https://productovi.com/courses/product-analytics/funnels.md)

---
Productovi Learning. Free, no sign-up needed to read.