What Shipping a Smartwatch Taught Me About Cross-Functional Execution
Software teams can slip a week. A factory booked for a production run cannot. Wearables compress every hard problem in cross-functional execution into a device the size of a cookie.
When I led program management for Wear OS and Pixel Watch software at Google — three OS releases and a dozen watch launches across Pixel Watch 1–3, Samsung Galaxy Watch 4–7, Xiaomi, and OnePlus — I inherited a truth that reshaped how I think about execution: in hardware, the ship date is physics. Components are ordered months ahead. Manufacturing lines are reserved. Retail launches are choreographed across carriers and countries. The date doesn't negotiate.
That constraint sounds brutal, and sometimes it was. But it taught me more about cross-functional leadership than a decade of software-only launches, because it forces disciplines that fluid deadlines let you avoid.
One date, many clocks
A smartwatch launch runs on incompatible clocks. Silicon and sensors move on multi-year cycles. Firmware and OS releases move on quarters. Health features move at the pace of clinical validation and regulatory review. Apps and partner ecosystems move in weeks — for Pixel Watch, we coordinated 22 Google product teams, from Fitbit to Maps to YouTube, to land optimized apps on day one. The TPM's job is to make these clocks strike midnight together — once, in public, with no patch window for the hardware.
What made that possible wasn't a master schedule, though we had one. It was a shared understanding of which decisions were reversible and which weren't. We spent disproportionate energy on the irreversible ones — sensor selection, battery budgets, the feature commitments printed on the box — and gave teams real autonomy on everything else. Most execution failures I've seen since come from inverting that: agonizing over reversible decisions while irreversible ones slide through unexamined.
Health raised the stakes
Android Health added a dimension consumer software rarely faces: when your product touches heart rhythms and health data, "move fast" meets "first, do no harm." We worked with health scientists and regulatory experts whose definition of done was fundamentally different from an engineer's — and correctly so.
Cross-functional leadership isn't getting everyone to work the same way. It's building a system where different definitions of rigor can coexist and still ship.
The failure mode is treating the most cautious function as a blocker to route around. The winning move is the opposite: pull them in earliest, because their constraints are load-bearing. A validation requirement discovered in month two is a design input; discovered in month ten, it's a crisis.
Three practices that traveled with me
Name the ship-date owner of every risk. Not the workstream owner — the person who will personally explain to leadership if this risk eats the date. Ambiguous ownership is where schedules go to die.
Run the launch backwards. We rehearsed from launch day in reverse: what must be true the week before, the month before, the quarter before. Backwards planning surfaces the quiet dependencies — certifications, translations, store approvals — that forward planning treats as afterthoughts.
Make status emotionally safe and factually brutal. The moment a team learns that reporting red invites punishment, your dashboard becomes fiction. We celebrated early bad news like the gift it is, because the alternative is late bad news, which is a catastrophe.
Every organization has its version of the factory that can't wait — a moment when reality stops negotiating. The teams that thrive are the ones that built their execution muscles before that moment arrives.