Launch a Working MVP with Real Users
20 weeks · 3 milestones
Run user research with at least five affected people, write an evidence-based MVP specification, build a deployed V1.0 product, and document real usage by at least three people outside the family. Proof requires: user research notes, a feature-to-evidence specification, a change log tracing V1.0 decisions to user feedback, and a faculty attestation confirming the launch is real.
Milestone map
Milestone map
3 milestones
Before writing a single line of code or making any product artefact, interview at least five people who experience the problem you identified in Discovery. Run structured conversations using the problem scoping you produced — ask what they actually need from a solution, not what you assume. Synthesise the findings into a product specification documenting: the target user description, three specific user needs derived from the interviews, the core functionality a V1.0 would deliver to meet those needs, the tools and platform you will use to build it, and what success looks like for a single user after their first use. Features in the specification must trace to interview evidence — if a feature cannot be traced to a specific user need from the research, it is cut from V1.0.
Proof required
Submit: (1) notes or transcripts from at least five user research conversations — real names or anonymised descriptions, specific responses quoted, and dates; and (2) your MVP specification document (minimum 400 words) covering: target user description, three specific user needs with the interview evidence that produced each, core V1.0 functionality list with each feature traced to a user need, tools and platform selected, and a one-paragraph success definition specific enough to be testable.
What gets checked
- User research is documented with specifics — five conversations with real responses recorded, not paraphrased summaries or hypothetical needs invented after the fact
- The specification traces each core feature to a specific user need from the research — for every feature in the V1.0 list, there is a corresponding interview finding; features without evidence are absent
- Success is defined specifically enough to be testable — 'the user completes X in under Y steps without asking for help' not 'the user finds it helpful'