Reachable
Don’t describe the value — make it visible.
Reachable pools a team’s collective rolodex to find warm referral paths into sales targets. It worked well as a Salesforce plugin — but its standalone web app was a cumbersome afterthought. My job was to make the standalone product stand on its own.
Great inside Salesforce. Lost on its own.
The plugin lived where salespeople already worked, so it felt useful. The standalone web app had inherited the plugin’s mechanics without its context — a scroll-log of relationships that made you hunt for the one thing the product was supposed to hand you: the warmest path to the person you’re trying to reach.
From a vertical scroll-log to a comparison-first canvas.
The signature move was to stop stacking relationships in a list and lay them out horizontally — you on one side, the target on the other, connectors between — so multiple referral paths are visible at once and comparable at a glance. A network you can seebeats a network you have to reconstruct in your head.
Twenty thousand connections, narrowed to the useful one.
A canvas only helps if you can get to the right slice of it. I layered on filtering (narrow tens of thousands of connections by channel and relationship type), a relationship-strength profile, a contact-card system, and real error states — the unglamorous cases that decide whether a tool feels trustworthy.
Channel and relationship, as first-class filters.
The filter view turned an overwhelming graph into a workable shortlist — connection paths sliced by channel and by relationship type, so a rep starts from “the paths that matter” instead of “all the paths.”
Strength is the whole point of a warm intro.
A relationship-strength profile made the intangible legible — weak to strong, in both directions, grounded in shared history — so a rep could tell a genuine connection from a business-card acquaintance before spending social capital on the ask.
The unglamorous cases that build trust.
A contact-card system carried the person’s essentials at a glance, and I designed the error messaging as a real state — not an afterthought — because a referral tool that fails silently is worse than no tool at all.
I named the risk, then designed against a clear vision anyway.
The whole premise — that a standalone market existed at all — was an unvalidated assumption owned by marketing, outside my scope to test. The honest thing was to say so, and then do the job well against the product owner’s clear vision. On a short engagement, you pick your battles and design the best version of the bet you were handed.
Make the value visible rather than describe it.
The redesign turned a cumbersome list into a comparison-first canvas where a rep could see, filter, and judge warm paths at a glance — the product’s actual value, finally on the surface instead of buried in a scroll.
The engagement had target KPIs (more standalone adoption, new standalone customers) but no measured results were captured, so I don’t quote numbers.