Must
Without this, the product fails. The user can’t do the thing the product exists to do.
Why prioritize at all
You wrote 8–12 stories last lesson. If you ship all of them, you’ll ship none of them — not in any reasonable amount of time, and not at the quality that would convince a real user to come back. Prioritization is the act of admitting, in writing, that the first version of the product is not the whole product.
An unprioritized backlog is a wish list. A prioritized backlog is a plan.
MoSCoW, in one paragraph
MoSCoW comes from DSDM in the mid-90s and has stuck around because it’s blunt. Every story gets exactly one of four labels:
Should
Important but not critical for v1. The product works without it, but feels half-finished. Plan to add it next.
Could
Nice if it’s cheap. The kind of polish you’d add on a quiet afternoon, not a feature you’d staff a sprint for.
Won’t (this time)
Explicitly out of scope. Not “we forgot,” but “we decided.” This category is the most important one.
The trap: everything is Must
The first time you do this, every story will look like a Must. That’s the warning sign. If everything is Must, nothing is. Force yourself to a ratio that hurts a little:
A workable distribution for v1
- Must: 3–5 stories. The exact set you’d ship if the deadline moved up two weeks.
- Should: 2–4 stories. Strong candidates for v1.1.
- Could: 1–3 stories. Drop them if anything else slips.
- Won’t (this time): at least 2 stories. Naming what you’re explicitly skipping is what makes prioritization real.
If your Must list has eight items, you’re not prioritizing — you’re negotiating with yourself. Cut.
The discipline: name the Won’t
Most teams skip the Won’t column. They write Must, Should, and Could, and leave the leftover stories floating in some unwritten “maybe later.” This is a mistake. Naming the Won’t is the move that prevents scope creep three weeks from now, when someone — possibly you — asks “wait, weren’t we going to do X?”
You weren’t. You decided not to. Write it down.
Use Claude as a sparring partner, not an oracle
This is one of those decisions where the value of an AI assist isn’t the answer — it’s the argument. Paste your stories in, ask Claude to assign MoSCoW with reasoning, and then disagree where you disagree. The reasoning is what makes it useful.
Operator story · Lesson 2
From Alon’s notebook
The PM ritual Alon learned at Intuit for cutting a Must list in half. Suggested hook: the question that breaks the tie when the team can’t agree (it isn’t “is it important?”), and the meeting that turned a 9-Must list into a shippable 4.
[OPERATOR STORY — Lesson 8]Tonight’s assignment
Run the prompt. Argue with the output. Drop your final MoSCoW labels into the workbook below, and then write the Must list as one block of text — that’s the input for Lesson 03.
Assignment · Section 08
Prioritize with MoSCoW
Sign in to save your answers.Label each story M / S / C / W. Then distill the Must list.
Done with this section? Review all your answers in one place.
Review my workbook →