How we run discovery before writing production code
Workshops, spikes, and decision logs that keep builds aligned with business goals.
Technical deep dive
Jumping straight into tickets feels fast until the wrong product ships. Discovery does not have to be a six-week theatre production.
What we always capture
- Primary user jobs and success metrics.
- Non-goals (explicitly out of scope for v1).
- Integrations and data sources with real owners.
- Constraints: compliance, performance, devices, offline needs.
Spikes over speculation
If an integration is uncertain, schedule a time-boxed technical spike. A two-day spike that kills a bad approach is cheaper than two months of code.
Decision log
Write down the choices and why. Future you — and the next engineer — will not remember the meeting. The log also stops reopening settled debates every sprint.
Discovery is finished when the team can explain the product, the risks, and the first vertical slice — not when every wireframe is perfect.
