Measuring Emissions

The first gate was internal. Before running a single external session, I spent weeks with CPG and supply chain experts already inside Indigo - building vocabulary, stress-testing assumptions, mapping where the industry had tried this and failed. That domain fluency was what made external recruitment possible at all; the subject was specific enough that cold outreach to participants wouldn't have worked.

Once I had a frame, I made the case to the VP of Product, recruited a junior researcher, and stood up the formal discovery project. Third-party recruiters sourced participants we couldn't have reached through our own network.

Each session split its time: half structured research guide, half visual reaction to rough sketch concepts. The concepts were conversation starters drawn at a fidelity meant to be thrown away - concrete enough to push back against, not finished enough to defend.

PM and I divided signal ownership: PM owned GTM (what could actually be sold, to whom, in what window); I owned user signal (what jobs the tool needed to do and for whom). We synthesized at the end to scope something both real for users and sellable. The PM carried the project forward from there.

The map was the right educational surface and the right sales narrative. It was also harder to design than expected, because users came in with Google Maps muscle memory but the actual workflow was GIS authoring - creating and managing spatial objects, not navigating.

Consumer conventions help on familiarity and hurt on authoring tasks. On Google Maps, a tap navigates or surfaces information. In this product, a tap means “select this object to edit.” Long-press is the standard context menu gesture, but it's low-discoverability in a professional workflow where people don't know to try it. Slide-up info panels work for reading; they're awkward for editing or associating data to a spatial object you just created. Object creation is the clearest example: Google Maps technically supports drawing areas, but it's buried in a navigation-first flow. We needed it to be primary.

Engineers on the team had deep mapping backgrounds - not just rendering knowledge but interaction pattern knowledge. They flagged where consumer conventions would break before design had to discover it the hard way. Most of the decisions about mode switching and gesture differentiation came out of that collaboration.