02 / ZERO-TO-ONE AT SCALE · ~3 min read
Four zero-to-ones: founding teams, and shipping the launch that lost clicks on purpose
What this case should tell you about me: Strategic, deep product sense: trusted with ambiguous zero-to-one, self-directed.
- The product
- Founded the Japan Travel UX team: I led user-centered UX vision and product strategy across functions, from Engineering to Partnerships, as the only designer on a team of roughly fourteen.
- The problem
- Google's global, rooms-based travel product wasn't meeting localized Japanese needs in an OTA-dominated market, and every localized fix had to survive a western-first design system and global metrics.
- The outcome
- Our first big launch caused a ~9% booking-click loss; we shipped it anyway. Within five months the loss mostly closed, ads conversions swung from ~-2% to ~+12%, partners who adopted the format saw ~+15% gains, and the pattern scaled to Google Maps (~+46%) and the Search booking module (~+24%).
- My role
- The same 0→1 muscle ran before and after Japan: sole designer and UX Lead of YouTube Gaming (50B+ yearly watch hours), QR UX lead for India at hundreds of millions of monthly offline transactions, and strategist of the concept voted hero feature of the Brazil team's formation workshop.
I'd rather be hired for my product sense. What I do well: charting a creative vision to inspire the team while keeping it grounded in a principled system.
Context & constraints
Japan is highly valuable to Google and to Travel, with clear gaps to solve. Foundational research in 2019 identified unique needs; from 2019 to 2022, Google wasn't meeting them. The engineering team was staffed in May 2022; I arrived that November as the first and only designer among roughly fourteen people: no dedicated researcher, no established roadmap, no precedent for what "travel UX for Japan" should even mean. The work before the work was deciding what the team's first real bet would be, and building the evidence to defend it.
One designer. During a UXR shortage, research still had to happen: I collaborated with data science and initiated foundational research to assemble user-centered, data-backed rationale.
A global product surface. International solutions get automatically labelled non-compliant by a western-first design system, or called "cluttered"; a Japan win that dented a global number would die by default unless argued for. And an entrenched market: Japanese OTAs provide hundreds of filters to give users confidence, and the average Japanese user is part of 20+ loyalty programs. We weren't going to win by copying them.
Three decisions I pushed for
- 01
Ship the user's mental model, even when the launch metric dips
OTAs have millions of packages and plans. Literally. Plans = Room + Meal + Activity, so users book plans rather than rooms: 80% of people staying in popular onsen destinations choose with-meal plans (Ministry of Tourism, public source). Google's global product modeled rooms. We added crucial booking-plan details to hotel listings and let users filter by them. Showing more info upfront and fewer partners after filtering caused a ~9% booking-click loss. We believed this was still a critical launch for JP users: focus on the user and all else will follow.
What we didn't do: A metrics-safe cosmetic localization (translated labels, local photography) that protected the headline number but changed nothing about the mismatch. After 5 months: more major partners onboarded, ads conversions swung from ~-2% to ~+12%, and partners who adopted the format saw ~+15% conversion gains. The feature became an onboarding incentive rather than a tax.
- 02
Bet on the open standard, kill your own proprietary rail
On payments in India, I took over the scanner system and led several pillars to improve and grow QR scans. The data said users had already voted on our proprietary merchant-code format: under 1% of scans. Despite an obscure UX, people kept trying to find their national-standard QR ("I didn't know I could scan PhonePe QR with Google Pay"). Countries with a national QR standard want interoperable codes so users can pay to and from any app. Our POV: make the national-standard code primary.
What we didn't do: Keep spending design and eng effort propping up our own proprietary rail for lock-in. Killing your own team's artifact is politically expensive; the counterfactual, a scanner optimized for codes nobody scans, was more expensive.
- 03
Ship the playbook, not just the launch
A 0→1 win that stays local is a feature; a 0→1 win that others can replicate is leverage. I established the global QR scanner system (a scalable, reusable bottom sheet across Brazil, India, Singapore) and aligned the scanner working model across 4 designers, 7 teams, 4 regions. In Japan, the localized booking pattern's scaling to Maps was approved and built with Maps eng, deliberately validated via holdbacks. And I pulled together the efforts to improve the design system for international needs, so the next international team wouldn't re-fight my fights.
What we didn't do: Hoard the wins as personal scope. Singleton designers are structurally tempted to stay indispensable; I optimized for the opposite.
How I did it
Act 1: found the practice (Japan travel, 2022→2023)
Founded the Japan Travel UX team. I led user-centered UX vision and product strategy across functions, from Engineering to Partnerships. First-quarter foundations: user archetypes, the journeys we’d win or lose on, a visualized roadmap, and research priorities I could execute without a staffed researcher. Partnered with an external design firm on the founding concept sprint, then carried the first bet single-handedly from ideation through launch review to public ship.
Act 2: scale across verticals (2023→2024)
Led the expansion from JP lodging to JP travel to JP journeys. Onsen is the #1 reason Japanese people travel domestically, and onsen-town queries were underserved: Google showed Wikipedia results while JP OTAs showed ryokan recommendations. I built the case the way I always do: collaborated with data science and initiated foundational research. We did a team trip to Kusatsu Onsen to form our user journeys (we entered an outdoor day onsen, had to bring our own towels, and there was no shower, but the temperature was just right). We recruited Googlers to share case studies of their favorite onsen trips (one had visited 50 onsens in the past year), and, of course, we aren’t the user: 25% of JP travelers use magazines for travel inspiration, and magazines go into a LOT of detail, down to water chemistry. Led the onsen-town workshop with PM, Eng, BD, and partner-team guests. Leadership greenlight from VPs; started a new Food vertical with a strong UXR foundation. The launched hot-spring page drove a ~+22% lift in time-on-page.
Act 3: scale across geographies, the physical→digital thread (India)
Before Japan travel, I spent 18 months on payments in India and Japan. I founded and led UX for the merchant-payments team, India’s big rock. As India’s economy opened back up, QR scans passed hundreds of millions of monthly offline transactions; I led QR UX for India and landed 10+ impactful user launches, several proposed by me and rooted in foundational scanner research, with ~+18% QR transactions received and ~-17% scanning latency. On the Japan side: UX Lead on the payments app for Japan, coordinating across 15+ UX designers and researchers. Then the physical→digital idea ran the other direction: printed QR codes that turn a real-world storefront into a live digital surface. I developed strategy and UX for this zero-to-one program: prioritized use cases, developed QR design principles, designed the stickers and merchant guides. Our first QR? The PM ran to a copy shop to print test stickers to paste at a relative’s salon. Piloted with thousands of merchants across India and Japan, it showed a measurable (~2%) uplift in Search usage after scanning. This is the “it scaled to India” proof: the same 0→1 muscle, exported.
Act 4: the muscle is repeatable (YouTube Gaming before; Brazil after)
YouTube Gaming (2018–2021). UX Lead of YouTube Gaming, the world’s largest gaming content platform with 50B+ yearly watch hours; more than 200 million people watch YouTube gaming content every day. I was the sole designer for the whole vertical, a 100+ person cross-functional org, in a role previously held by a designer two levels my senior. We partnered with the world’s biggest publishers and launched loot drops to millions of users, with tens of millions of dollars in rewards and a ~+10% lift in gaming watch time; our partner pipeline was larger than our eng capacity. For live raiding I researched, prioritized, roadmapped, drove alignment for, designed, and tested end-to-end feature flows: a project requiring alignment with 18+ stakeholders across 6 major product areas. Proactively identified, validated, scoped, and launched personalized gaming surfaces with a ~+22% watch-time lift. And when off-platform signals raised privacy questions, I didn’t route around them: I authored the transparency-and-control framework that let any team at YouTube use such signals safely. A designer-originated policy artifact, adopted org-wide.
Brazil (2024). Most recently the playbook ran in a market I’d never worked: a co-creation program with local youth in São Paulo that generated 100+ ideas and converged on three concepts. I conceived the UX strategy for the selected concept, solving for complex entrypoints spanning Android gestures and 3P messaging apps like WhatsApp, and formulating a full loop for saving places, personalized recommendations, and collaborating on weekend plans with friends. After my initial proposal received the most excitement from senior stakeholders, I provided UX leadership to partner designers and engineers. It was voted the hero feature of the Brazil team’s formation workshop and influenced resource asks for the following year. Sized by data science at roughly 3x the estimated impact of features then being greenlit; a VP-level award for human-centricity; from first workshop to working prototype, about six weeks.
Breadth, briefly
Android premium segment (2022, ~3 months). I started the Android premium team in the Singapore office, but was only in it for three months before an undeniable offer to start Travel Japan. By then I had set up a phone-testing program so Googlers could test Chinese OEM phones (Xiaomi, OPPO, Samsung) and propose ideas collaboratively; I set it up physically in the office to look like a mini phone store. If it helps the team find a vision together, I’ll set up the foundations for 0→1, even if that means building a mini phone store.
YouTube verticals playbook (2020). Owned the YouTube-wide verticals UX initiative across 8 vertical teams spanning the entire org. Established the org playbook and identified 5 themes common to verticals.
Impact
Founded the Japan Travel UX team, then scaled its scope to new verticals (the launched hot-spring page drove ~+22% time-on-page) and impacted international design systems beyond Japan. On YouTube Gaming: loot drops to millions of users with tens of millions of dollars in rewards and ~+10% gaming watch time; personalized game pages at ~+22%; plus a privacy framework adopted beyond my own team. The Brazil concept was sized at roughly 3x the features then being greenlit and prototyped in about six weeks.
“It is a major win for our users and our partners, and would not have happened without your hard work, insights, and leadership.”
Julie Farago, VP of Engineering, on the Japan launch · permission pending
What broke / what I'd do differently
The launch that went backwards
The ~9% booking-click loss meant the team's annual goals were missed on both topline measures. What saved it wasn't spin: it was having pre-committed to the user-model argument before launch, instrumenting the recovery, and building the partner incentive loop that turned the format into something OTAs wanted to join. What I'd do differently: size and socialize the expected dip quantitatively before launch, not just directionally. The months of metric anxiety were spent re-winning an argument I could have pre-won with a forecast.
We built the wrong rail first
The proprietary merchant-code format was already live when I took over the scanner, and users had already rejected it: under 1% of scans. The failure wasn't mine to prevent, but the pivot was mine to argue. What it taught me: in infrastructure UX, interoperability beats ownership; the sunk-cost defense of a proprietary format is the most expensive design decision a team can quietly keep making.
Pilots that fade
The printed-QR pilots showed real scan-to-usage lift, and honest telemetry also showed the other thing: merchant retention decreasing over time. Novelty effects are a 0→1 trap; the first metric a founder-designer should design is the one that detects their own product's decay. We logged the lowlight instead of burying it, which is why I still trust the uplift number.
The discipline that made the wins believable: holdbacks
The reason I can say the Japan pattern scaled with double-digit gains is that we held back cohorts and measured, rather than narrating. When a result was noisy we said so, in the review, on the slide. Every 0→1 claim in this case survived a counterfactual: exactly the standard I'd bring to evaluating agentic-UX features, where the temptation to demo-and-declare is even stronger.