Why New Products?
Better meet need
Meet unmet need
Offer new features and functions
Hard but necessary to verify market for software product
Features, Ilities, Design Constraints
Feature can be captured by "story card": for example "As an author I need a way to organize the text of my book into chapters and sections."
"-ility" is something that needs to be true of the software but isn't really a feature: for example "every software transaction must complete within 60 seconds"
Design constraint is a requirement that exists for reasons outside the software product: for example "The software must be written in UCSD Pascal."
Software is always for people
Making "representative" people up helps in understanding what software needs to be / do
Be a fiction writer — give names and humanizing details, for example: "'Lando' (gamer tag) is a 14-year-old male. He lives with his Mom in a suburb of Portland. He mostly plays multiplayer FPS and TPS games, especially Halo and Counterstrike. He uses Discord for in-game voice and text chat."
Maybe five-ish persona depending on product scope? Be sure to cover all major roles that might interact
Validate your personas with domain experts
Detailed stories about using the software, for example "Lando sits down at his gaming desktop and logs into the game using his Steam account. He starts to create a new character, selecting female gender and combat infantry class, and starts to fill out cosmetics. Lando's mother calls dinner — Lando leaves the screen open and leaves the room to eat. While he is out, there is a momentary power failure. When Lando returns he reboots his computer, logs back in, and proceeds to complete character creation by selecting a starting zone."
Detail helps catch missing stuff, and gives basis for organization
Again, be a fiction writer, but validate your secenarios with domain experts
Should be scenarios for all personas; should be scenarios that cover most normal use
Story cards contain actual requirements, for example "As a player I want to create a customized character"
Developed from scenarios, personas, product vision
"Epic" is story that is too big to be directly implementable: include, but then break up into multiple story cards. For example, above might be "As a player, I want to select my character's gender", "As a player, I want to select my character's class", etc…
User stories should be directly actionable as code
Reqs Development Tips
Ruthlessly eliminate anything not absolutely necessary (can always add back after release)
Update/add scenarios to reflect changing vision and requirements
Don't forget the -ilities!
Avoid design constraints as much as possible, as they can really limit solution space