Smart Home Platform

Overview

This was a two-year project to build a coherent connected home product range — three hardware SKUs sharing a common software platform. I joined when the hardware was already in tooling and the app was a blank Figma file. My job was to make sense of everything that sits between those two things.

The Problem

What needed solving

The hardware team had made good hardware. But each device had been specced independently — different engineers, different timelines, different assumptions about how users would set them up. By the time I arrived, we had three products that were physically incompatible in ways that weren't obvious until you tried to build the companion app.

The deeper problem: nobody owned the seam between the hardware and the software. Engineers cared about the device. The app team cared about the app. Users would only see the experience — and the experience was going to be incoherent.

My Role

What I did

I became the person who owned the seam. Concretely, that meant running weekly syncs between the hardware and software teams that hadn't previously talked to each other, writing a shared "platform spec" that both sides could design against, and making calls about which incompatibilities we'd live with versus fix before launch.

Some of those calls were genuinely hard. Fixing the Bluetooth inconsistency between Device 2 and Device 3 would have cost six weeks and a partial re-spin of the PCB. We decided to paper over it in software and be honest with users about the limitation. That was the right call, but it required convincing a hardware team that software could make it not matter — and then holding the software team to actually delivering that.

Process

How I approached it

The first thing I did was map the setup journey for all three devices simultaneously — every state, every failure mode, every moment where the app needed something from the firmware and the firmware might not give it. This produced a 12-page document that immediately became the source of truth for both teams.

Halfway through, we ran a round of unmoderated user testing on the setup flow and discovered that the two-step pairing process we'd assumed would be obvious was being abandoned by 40% of participants at step one. That was the pivot — we rebuilt the onboarding from scratch with a single-step flow, which required a firmware change we'd said was out of scope. I made the case and got it approved.

Outcome

What happened

The platform launched on schedule. Setup completion rate in the first month of live data was 78% — significantly above the 60% benchmark from comparable products. The Bluetooth workaround held; we had 0 critical bugs related to the cross-device incompatibility in the first 90 days.

The more durable outcome was the working relationship between the hardware and software teams. The joint reviews became a standing meeting that outlasted the project, which is the kind of thing that doesn't show up in a product metric but matters for the next thing.