“Build us a LinkedIn.” What we shipped instead and why it was the right call.

The brief arrived as a vision, not a problem: a social network inside an enterprise sustainability platform. I spent the first phase working out which parts of that ambition were solving a real job, and which were borrowed from a consumer product that doesn't apply here. What shipped was a consent-gated company directory that is still in production today.
TL;DR
- The problem:
- Companies already on the platform couldn't find each other. Onboarding a supplier who was already a customer meant an out-of-band email invitation, manual identity matching, and duplicate company records every time. Carbon data requests stalled before they started.
- The brief: “a LinkedIn for SiGREEN.” What I did with it:
- Reframed a feed-and-followers vision into three answerable questions how do you find a company, how do you prove it's the right one, and who gets to say yes and designed against those instead.
- What shipped:
- A searchable directory keyed to verifiable business identifiers, opt-in visibility controlled by each company, and a double-consent connection flow that only company admins can approve. Connection unlocks the actual job: adding that company as a supplier and requesting PCFs.
My role, precisely
- Role
- Lead Designer
- Team
- 1 PO4 FE devs3 BE devs2 Testers3 Scrum Masters
- Timeline for proposal preparation
- January 2024 → April 2024
- Implemented on
- April 2025
The brief, as it arrived
Stakeholder Request: “We want a social media platform — a LinkedIn for our SiGREEN.” - Product stakeholder, kickoff
The brief looked great on paper. It wasn't actually possible.
The stakeholder said: “Build LinkedIn for suppliers.” But LinkedIn works because it's public everyone's profile is visible, everyone can browse, everyone's searchable. For manufacturers buying parts, that doesn't work. Your supplier list is secret. Your cost data is confidential. Publishing it would hurt your business.
So if we built what they asked for, we'd ship something nobody could use without breaking their contracts. I figured out what they actually needed. They said “LinkedIn.” What they meant was: “Help our companies find each other and connect.” Those are different problems. I went and talked to the people who'd use it — buyers, operators, legal teams. I found out they wanted to find partners, but only if they controlled who could see them. So I designed for that instead: a way to search and connect, but with permission. You decide who sees your data. I also drew a line: I owned the permission system and what it looks like on screen. I didn't own the technical pieces that make it work (the backend system). But I was crystal clear about what I needed from them to make my design land.
Why this matters: The real work wasn't building a feature the stakeholder asked for. It was understanding what they actually needed, and redesigning the whole thing around that.
The problem underneath the vision
SiGREEN's core loop is a customer requesting a product carbon footprint from a supplier. That loop assumes you can reach the supplier. In practice:
Two companies could both be active on the platform, in the same supply chain, and still have to onboard each other from scratch over email.
There was no reliable way to answer “is this company already here, and is it the same legal entity I contract with?” Company names collide, subsidiaries share branding, and procurement teams work from tax IDs and DUNS numbers, not brand names. The result was duplicate accounts, invitations sent to companies that were already onboarded, and a request cycle that stalled before any carbon data moved.
Reframed, the vision became three design questions, and they replaced “make it social” as the scope:
Three design questions
1. Discovery
How does a procurement user find a company that's already on the platform?
2. Identity
How do they confirm it's the right legal entity, not a lookalike?
3. Consent
Who authorises a connection, given that connecting grants the right to request data?
Translating the vision into scope
Placeholder. The reference page introduces this section (“I ran the stakeholder's ‘LinkedIn’ mental model through the three questions above, feature by feature. Each one got kept, reshaped, or cut — with the reason on record, so the decision could be revisited later rather than re-argued.”) but its detail is not present in the published page source. Placeholder: add the keep / reshape / cut table here.
There's more detail than fits on a page.
How the “LinkedIn” brief got renegotiated without losing the stakeholder, the states that didn't make this page, and how I had to pivot the whole approach of my workshop Happy to walk through it live.