The Engineering Firm's Approach to Product Discovery
How we blend technical feasibility with user desirability early in the product lifecycle — before a single line of code is written.
Discovery before the backlog
Most failed products did not fail at sprint three — they failed because the problem was never validated with the people who would pay for it. In an engineering firm, the temptation is to jump into architecture because that is where we are strongest. Discovery exists to protect that strength from being spent on the wrong problem.
We treat discovery as a time-boxed workstream with the same discipline as a build phase: owners, artefacts, and exit criteria. The goal is not endless workshops. It is a shared decision that a problem is worth solving, for a defined buyer, under constraints we can actually ship against.
Triangulating desirability, feasibility, and viability
Every discovery engagement maps three lenses. Desirability comes from interviews, shadowing, and job-to-be-done framing. Feasibility comes from spike work, constraint mapping, and an honest look at data readiness. Viability asks who buys, how budget is approved, and what switching costs look like.
When one lens is weak, we do not paper over it with optimism. A delightful workflow that requires unavailable data is not a roadmap item — it is a research debt ticket. Product and engineering co-own that call so nobody feels overruled later when scope changes.
Artefacts that survive the handoff
The output of discovery is not a slide deck that dies in email. We produce problem statements, opportunity scores, technical risk notes, and a thin prototype or clickable flow when it clarifies debate. Those artefacts become the north star for the first delivery increment.
We also document what we chose not to build. Explicit non-goals prevent scope creep when a stakeholder rediscovers an idea three months later. Discovery that leaves decisions undocumented forces the build team to rediscover politics under deadline pressure.
From insight to first slice
A successful discovery ends with a first vertical slice that can learn in production: one persona, one critical workflow, measurable success criteria. We resist platform-first builds that delay contact with reality.
Engineering estimates during this phase are ranges with confidence labels, not false precision. That honesty keeps commercial conversations grounded and makes it easier to re-plan when discovery uncovers a harder integration than expected.
Marcus Chen
Head of Product
Product leader focused on discovery, packaging, and shipping B2B software that engineers and buyers both trust.
