MidCase studyBacklog management
Split a large checkout epic into stories
Try answering first
Model answer
The aim is stories that are vertical slices: each delivers a small piece of working value end to end (UI, logic, data), can be finished within a Sprint, and can be tested on its own. I avoid splitting by technical layer, such as "build the checkout database".
For a checkout epic I would begin with the thinnest complete path, a walking skeleton:
- Guest pays for a cart with one card type and receives an order confirmation.
- Validate card errors and show clear failure messages.
- Add a second payment method, such as a digital wallet.
- Add shipping options and cost calculation.
- Apply discount codes.
- Saved addresses and payment details for returning users.
- Order confirmation email and order history.
Useful splitting patterns: by workflow step, by business rule variation, by data type, by happy path versus edge cases, and by simple versus complex.
Then I order the slices by value and risk, so the team learns early and stakeholders can see progress each Sprint.
Key points
- Split into vertical slices that deliver working value, not by technical layer
- Start with the thinnest end-to-end path
- Use patterns: workflow step, rule variation, happy path first
- Order slices by value and risk
Likely follow-ups
- How small is small enough for a story?
- How do you handle a story that depends on another team's API?
Common mistakes
- Splitting by layer so nothing is usable until all parts are done
- Making stories so small they have no user value