# Lesson 4: Decide what's worth building

> Prioritization with priority, effort, and RICE: how to pick the next thing deliberately when there's one of you.

Source: https://productos.ca/documentation/guides/academy/lesson-4-decide-whats-worth-building (last reviewed 2026-09-22)

A **feature** is a specific capability you build to address a theme. **Prioritization** is deciding which one comes next when you can't build everything, and solo, you can build almost nothing at once. Choosing wrong for a month costs a month you don't get back.

## The two cheap tools

**Effort / Impact.** Size the cost (`xs`–`xl`) against the value. The 2×2 has a quadrant with your name on it: *high impact, low effort*. Live there as long as the market lets you.

**RICE.** When gut-ranking stalls, score it:

```
RICE = (Reach × Impact × Confidence) ÷ Effort
```

- **Reach**: how many users does this touch per quarter?
- **Impact**: how much does it move each of them? (3 = massive, 1 = medium, 0.25 = minimal)
- **Confidence**: how sure are you about the above? (100% / 80% / 50%)
- **Effort**: person-weeks. For you, calendar-weeks.

The score isn't the point. The point is that scoring forces you to say *reach* and *confidence* out loud, which is exactly where shiny ideas die.

## Do it now

On [Features](https://productos.ca/documentation/pages/golden-thread/features):

1. Create a feature from your Lesson 3 theme and link the theme.
2. Set **priority** and **effort** honestly (hover the ⓘ on either column if the scale is fuzzy).
3. Open the AI overlay (`⌘.` / `Ctrl+.`) and ask **How do I prioritize?** (it's a suggested prompt when Learning Mode is on): the Copilot ranks your real list by priority, effort and RICE and explains itself. Disagree where you disagree, the model ranks; you decide.

> **Saying no is a feature**
>
>   "Won't do" is a real status, on purpose. A recorded no beats a backlog of polite maybes, and the user who asked deserves the honesty.