Common Story Splitting Patterns - Which Work Best?
We keep getting stories that are too large (13+ points) and struggle to split them effectively. What are the most reliable story splitting patterns you use? I've heard of workflow steps, CRUD operations, and business rule variations but need practical examples.
SPIDR method works brilliantly: Spike (research), Paths (happy/error), Interfaces (UI/API), Data (CRUD), Rules (business logic).
Example: "User registration" (originally 21 points) becomes 5 stories: Research OAuth providers (spike, 3pts), Basic email signup (happy path, 5pts), Error handling (3pts), Admin user management UI (5pts), Password strength rules (2pts). Each is independently deployable and testable.
Workflow steps are my go-to. For "payment processing" I split into: capture card details, validate payment, process transaction, send confirmation email, update order status.
Each step delivers value and can be demoed independently. Works great for backend-heavy features where user doesn't see intermediate steps but each builds on the last.
CRUD splitting saved our team. "Product catalog" split into: Create product (POST), List products (GET), Update product (PUT), Delete product (DELETE), Search/filter.
Simple, predictable, and each story has clear acceptance criteria. No more 21-point monsters. We can also prioritize - often Read/List comes first, Create second, Update/Delete later.
Business rule variations work when CRUD feels too granular. "Discount system" becomes: Percentage discounts, Fixed amount discounts, Buy-one-get-one, Volume discounts.
Each rule is independently valuable and testable. Product owners love this because they can prioritize business value - maybe we ship percentage discounts first to support Black Friday, save BOGO for later.