HEINEKEN · 2024-Present · Design Manager
One platform, 35 markets
HEINEKEN ran 35+ separate ordering sites, one per market, each with its own branding, its own features, and its own queue of one-off requests. I led design on replacing them with a single configurable product, and on deciding what a market was still allowed to change.
Decisions
- One frontend, not 35. Localization only where it changes the outcome, argued market by market.
- Every initiative names the pain, the target behavior, and the metric before design starts.
- Prototypes over decks for anything a static frame can't answer.
- Ship at good enough and measure, rather than hold for launch-perfect in every market.
Context
Over 35 markets ran separate sites with different branding and features, creating duplicate work, inconsistent experiences, and a standing queue of one-off requests. The vision was one global configurable frontend that localizes only where it matters. I co-led the customer journey vision, set the north star through design sprints, and led internal and agency designers against the roadmap.
How I led
Strategic design reviews and focused critiques that forced decisions rather than collecting opinions. With our Director of Product I built an eazle ecosystem demo covering customer, sales rep, and customer service in one story. It made the vision concrete enough for executives to fund.
Execution
I introduced AI prototyping so teams could explore flows, copy, and edge cases earlier than a Figma file allows. I built a GPT aligned to our tone and UX guidelines to draft and lint microcopy, because we had no dedicated copywriter. I built a white-label replica of the customer app so we could run unmoderated tests with external pools, and worked with CX Research on personas calibrated from analytics, order data, prior A/B tests, and surveys. That was enough to surface likely failure modes in hours rather than weeks.
One-way-door and high-risk changes still go to live research. The point was faster direction, not less evidence.
Where it landed
Multiple markets now run on one product with a reusable playbook. Teams track CSAT, active customers, reorder rate, search success, and task time with a clear read per release. It behaves like one product instead of many projects.
Learnings
Lead with consistency and bend for context when it helps users. Ship when it is good enough, then measure and improve. Users reward speed, reliability, and clarity.