Northwind · B2B SaaS · 2024

Zero to one, with users who hate apps

A field operations tool for technicians who work with gloves on

Role: Design lead
Team: 2 designers, 5 engineers, 1 PM
Duration: 7 months
Year: 2024

Technicians were logging jobs on paper and re-entering them at night. We shipped a mobile tool they'd actually use on a roof, in the rain, one-handed.

The problem nobody had measured

Ops leadership believed technicians disliked the old tool. Nobody knew why, and nobody had watched one use it.

I spent four days riding along with technicians in two cities. The tool wasn't the problem — the assumption behind it was. It expected both hands, good light and a stable connection. Field work offers none of those.

The average technician touched their phone 40 times a day and typed on it twice.

What we changed

We stopped designing a data-entry form and started designing a recording device. Everything became optional, resumable and offline-first.

Voice-first job notes, transcribed later, editable at the desk
Single-thumb layout: every primary action inside a 44px zone at the bottom
Offline by default — sync became invisible instead of a button
Photos replaced four text fields, because technicians already took them

How I ran it

Two designers, both mid-level, both wanting more ownership. I split the work by user journey rather than by screen so each of them could own outcomes end to end.

We ran a fortnightly session with four technicians on retainer. That single ritual killed more bad ideas than any critique I've run.

What changed

average time to log a job: 11 → 3 min
weekly active technicians, up from 31%: 78%
support tickets about the old tool: −62%

What I’d do differently

I front-loaded too much research. Three of those four ride-along days told us the same thing. Now I timebox discovery and buy the rest back in weekly contact with users.