epic-breakdown

Apply the nine splitting patterns to the spec's functional requirements: workflow steps, business rules, data variations, platform differences, CRUD operations, spikes, performance, persona variations, and happy path versus edge cases. Each resulting sub-epic gets a Fibonacci estimate and dependency mapping.

Common patterns include splitting by workflow steps, business rules, data variations, platform differences, CRUD operations, and happy path versus edge cases. Choose the pattern matching the feature's structure, such as workflow steps for multi-step flows or persona variations for role-specific behavior.

Split any sub-epic estimated above 8 story points using the nine patterns. Sub-epics at 13 points are considered very large and must be decomposed, while items under 1 point should be grouped together rather than tracked separately.

List each sub-epic's blocking dependencies, render them as a Mermaid graph, then allocate sub-epics to sprints so no item is scheduled before its dependencies. The dependency map prevents sprint planning surprises during engineering handoff.

Avoid platform-based splitting when differences are only cosmetic, such as CSS variations. Use it only when behavior genuinely differs across web, mobile, tablet, or offline contexts, otherwise the split creates artificial overhead without delivery value.