We spent months designing and building Spaces, a home for a team's ongoing work in Enterpret. We launched it to a few beta customers to learn. It looked good, and almost none of them created one. This is about noticing that, saying it out loud, and rebuilding the idea with the people it was for.

Role

Lead designer. Wrote the spec, coded the base UI (8 stacked PRs), ran the design-partner research

Team

A PM, two engineers (one on frontend) and a CSM

Timeline

May to Nov 2026

Outcome

A paused rollout, a product rebuilt around what customers told us, and a habit of putting users in the loop before building

TL;DR

• We built for months before a customer touched it. Spaces gave a team's ongoing work one home in Enterpret. Once it reached beta customers, only about one in eight Spaces was created by a customer. Our own team set up the rest.
• I called it a build trap. In our weekly PM meeting I said we'd taken on too much without a single user in the loop, and owned my part of that.
• The next morning I went to users. I picked design partners from our product data, then built each of them a Space from their own work before we ever spoke.
• Three bets, five calls. Two bets came back half right. The third, that people would set up their own Space, was simply wrong.
• The product flipped. The thing that watches became the product, Spaces became packaging around it, and the setup now asks for confirmation instead of configuration.

Coming soon

The full case study covers the design decisions I'd still defend, the moment I said it out loud, what five customer calls changed, and what I'd do differently. Check back soon.