Skip to content

Slicing & Prioritization

Which stories ship together, in what order. Value-driven, not effort-driven. Required-for-prediction stories must all be in the slice — or the slice can't test what it claims.

Owners: PO, Trio Phase it lives in: What We Build (Volume III) The corpus principle this enacts: Predictions over plans.

Where it lives in the chain

How to do this

The PO slices the story map into releases by asking, in order:

  1. What is the prediction this slice tests? "Slice 1 tests whether the shortcut moves the grading time from 47 min to under 25 min — the partial step before the full target."
  2. Which stories are required for that prediction to be testable? "Stories A, C, E, G are required. The shortcut needs to work, the next-exam navigation needs to work, the rollback needs to work. Without all four, the test is invalid."
  3. What can be deferred to the next slice without invalidating the test? "Stories B, D, F can wait — they add convenience but the prediction still tests without them."
  4. What's the smallest possible thing that still proves something? "This is slice 1. It's smaller than we'd like to ship. That's the point."

What good practice looks like

A team slicing the grading-shortcut Epic:

SlicePredictsRequired storiesDeferred
1 — shortcut MVPShortcut reduces session to under 25 minOpen, shortcut-key, next-exam, basic state preservationUndo, multi-section, statistics view
2 — full predictionSession under 15 minAdd undo, performance optimisationMulti-section, statistics
3 — broader rolloutSustained at scaleMulti-section, batch operationsStatistics
4 — extrasPower-user adoptionStatistics view, customisation

Each slice is a thing the team can ship, a thing the named person can use, and a thing the prediction can test against. Three properties that, together, define what a slice is.

The MoSCoW alternative — Must / Should / Could / Won't — produces a feature list with vague priorities. Walking-skeleton slicing produces a list of things you can ship in order, each one testing a sharper version of the prediction.

200apps · How We Work · NWIRE