Product Discovery

Several frameworks structure effective discovery. Continuous Discovery (Teresa Torres) — runs discovery in weekly cycles alongside delivery, with regular automated customer touchpoints (interviews, surveys) providing a constant data stream. Design Thinking — sequences discovery through five stages: Empathize, Define, Ideate, Prototype, Test. Jobs-to-Be-Done (JTBD) — frames customer needs as "jobs" they are trying to accomplish, and identifies the functional, emotional, and social dimensions of each job. Opportunity Solution Trees — visually map the path from a desired business outcome through customer opportunities to potential product solutions, making prioritization and dependency logic explicit. Product Ops facilitates discovery ceremonies and tracks which opportunities are being explored vs. validated vs. ready for delivery.

Discovery and delivery should run in parallel, not in sequence. A dual-track agile approach maintains two concurrent workstreams: the Delivery Track (Sprint work building validated features) and the Discovery Track (research and experimentation to validate near-future roadmap items). Product Ops schedules both tracks with appropriate capacity — typically 70–80% delivery, 20–30% discovery — and ensures discovery work feeds the backlog 2–4 sprints ahead. This prevents the dangerous pattern of "building first, learning later," where engineering effort is committed before the solution is properly validated. It also prevents research from running ahead of development capacity, creating untested insights that go stale before they can be acted upon.

Solution validation uses the minimum-fidelity artifact that can generate valid learning. In order of increasing cost: Narrative prototype (describe the concept in words and ask "would this solve your problem?"). Clickable wireframe (demonstrate the interaction flow without visual design). Design mockup (high-fidelity visual presented as if real, tested for comprehension and usability). Smoke test (fake door) landing page(test demand before building). Concierge MVP (manually perform the service before automating it). Technical spike prototype (minimal code to test technical feasibility of a critical component). Product Ops tracks validation signals from each stage and documents the evidence that supports progressing an opportunity to the engineering backlog.