Addressing the question that was stalling sales — before anyone asked.

Searching for connections between companies wasn’t on the roadmap yet, but graphing company networks was. I made graphing the Phase 1 and connection searching became Phase 2. Each Phase 1 decision would build towards Phase 2: the graphing platform, rendering model, and feature sequencing. Phase 1 would build the momentum for Phase 2, and when Phase 2 was green-lit, we would have the foundation.

The team was evaluating lightweight graphing libraries. I pushed for ReGraph — a more powerful, enterprise-grade library — because of what it could do at scale. It came with a rich component model, flexible APIs, and rendering capabilities that would support the complex networks Pathfinder would eventually need to reveal.

Entity networks are dense. A single company can have dozens of shareholders, investments, subsidiaries, and officers. Rendering everything at once creates noise, not insight. I designed a rendering model that flattened bi-directional relationships — cutting down on noise while still enabling filtering. Users didn’t have to think — it just worked. It helped the platform scale and visualize complex networks without overwhelming.

Simplifying how relationships are rendered on the graph helped display complex networks without adding extra noise.

I prioritized what was necessary to ship value. My read was that an unfiltered graph would open to visual noise: hundreds of nodes, no way to orient, no clear entry point. Analysts would open it once, decide it didn’t work, and never come back. A bad first impression on a feature this visible would be harder to recover from than a delayed launch, and a failed Phase 1 meant Pathfinder never gets approved. Filters shipped with v1. The graph was usable on day one.

Once analysts were exploring graphs the next problem became obvious. They needed to show the network graphs to someone; their coworkers, a procurement committee, a decision-maker. Export as SVG was the next feature — high resolution, small file size, designed for transport to the boardroom presentation. Download unlocked the external use case.

As analysts began exploring and creating more graphs — filtering, expanding, repositioning company nodes — they were investing real effort. Losing that work on logout wasn’t a minor inconvenience. It was a reason not to use the feature at all. The Save feature came next.