The Magic of Small User Stories

by Dwight Kingdon

I often run into teams that like big User Stories. Why spend time writing, planning, and estimating a bunch of small stories when you can write, plan, and estimate just one big story? Bigger stories mean better efficiency, right? Not necessarily.

Why smaller stories?

  • Transparency: Small stories provide much higher transparency/visibility into sprint progress during the Daily Stand-up and for stakeholders. Progress is easy to track when we see a series of completed, small stories; it’s more difficult to estimate and visualize true progress of a large story each day.
  • Predictability: Small stories tend to result in a more accurate and reliable velocity trend, which reduces variability and improves predictability. Relative story point estimates using the Fibonacci sequence are, by design, increasingly less accurate for larger estimates — like the “cone of uncertainty”. So, a 13 or 20-point story is likely much less predictable than several 2, 3, or 5-point stories.
  • Flexibility: Small stories provide more frequent opportunities for feedback and course correction. They provide more flexibility to adapt as we learn. When circumstances change or stories become irrelevant, there’s less waste in revamping or purging a small story.
  • Throughput: Small stories allow testers to begin testing sooner in the sprint, and work on smaller chunks of code.
  • Reduced Risk: Large stories increase the risk that the team will not deliver anything at the end of the sprint that is 100% done. Team capacity is like a funnel. When we try to force a large object (large stories) through it, the funnel gets clogged. For example, a Developer working on a large story may also be the only one with the skill set to complete another critical story.

How do I create smaller stories?

  • If there are several outcomes in a single user story, are they all necessary right now? Is there an 80/20 split that provides most of the value? Can you develop the 20% in a different story?
  • Are there multiple business rules or personas in the user story? Can rules be built separately, or different persona’s handled in separate stories? Can simpler rules suffice for now to get it working?
  • Can the “happy path” be coded first without all the exception conditions? Exception conditions can often be put into additional stories.
  • Are there multiple platforms or interfaces (inputs or outputs) associated with the user story? Can we develop them one at a time?
  • Are there multiple operations involved in the user story i.e., CRUD? Can we develop one operation at a time?
  • What test scenarios apply? Are some scenarios complex and not very relevant? Can a simpler first iteration be developed to prove the design?
  • Does most of the story complexity come from non-functional requirements like security or performance? If so, can the story be separated to first make the functionality work, then modified later to satisfy non-functional requirements?

How small is “small”? What’s the right size?

This is an Agile Alliance community blog post. Opinions represented are personal and belong solely to the author. They do not represent opinion or policy of Agile Alliance.

Originally published at https://www.agilealliance.org on July 22, 2019.

Agile Alliance is a nonprofit global member organization, supporting and serving the Agile community since 2001. AgileAlliance.org

Agile Alliance is a nonprofit global member organization, supporting and serving the Agile community since 2001. AgileAlliance.org