Building any multitiered solution is not just creating a User Interface to render the data, there is most likely a service tier that fetches data from a database, and serves that data up to the UI to then be rendered. How do you derive work items in your product backlog? One User Story, and multiple child tasks, one task per tech stack tier, UI, service tier, and database? Or three user stories, one per tech stack tier?
User Stories Defined, Per Tech Stack Tier
There are clear advantages of representing most work items with User Stories such as deriving story points, determining team average velocity, and a more accurate burndown chart depicting a downward trending scope and implementation of user stories.
Using child Tasks of user stories may obfuscate the total work required to implementation of the solution unless baked into the parent story points. Tasks are typically tracked in terms of hours, and separately user story points are calculated/derived from a collective, relative effort estimation, e.g. Fibonacci sequence; 1,3,5,8,13,20…, and many teams may overlay this scale to fit their sprint duration.
Feature and Story Planning – At a Glance
In order to organize each feature, and correlated user stories, teams may use a prefix in the title of the user stories, such as [UI] or [DB]. At a glance, a product owner, or the implementation team can see if all the required stories for a given feature have all the elements required to implement the feature. For example, if a new report needs to be created, multiple stories must contain [UI], [API], and [DB] stories.
Drawbacks – Accepting a User Story as Complete
If you segment your product backlog user stories based on tech stack, you may need to wait until all related stories, UI, API, and DB have been implemented. For example, If your API and DB stories are developed, and not the User Interface (UI), you’re QA/Testing may not start until the UI story has been deployed. Of course, your tester could test the API using testing tools like SoapUI.