Designing for Experimentation

The Thinking

The old platform had a ceiling — and we'd hit it. Testing was shallow, layouts were fixed, and marketing autonomy was limited by system constraints. The migration was a chance to change that, so I designed the foundation around flexibility and growth before focusing on visuals.

The Solutions

Redesigning for Flexibility

The redesign centered on one primary goal: give the business room to move. I embedded content slots at every touchpoint in the path to purchase — homepage, PLP, PDP, cart, and checkout. Where the old platform offered one fixed homepage layout, we now had fifteen configurable content areas. Teams now had the opportunity to add content, test promotions, and iterate with full autonomy — something that had never been possible before. Once the foundation was in place, the visual redesign followed.

To support the new architecture, I also built the organization's first shared design system — giving the design team a consistent foundation to work from across all four brands

Questioning a Legacy Assumption

The PZA flow had been in place for nearly eight years — a multi-step personalization and upsell sequence designed to increase order value. Through recurring Hotjar session analysis, I noticed a consistent pattern: high-intent users, particularly last-minute holiday shoppers, were hesitating and dropping off within the flow. I formed a hypothesis, aligned with my manager, and ran a test. Hiding the flow during peak season drove a 30%+ lift in conversion. Net revenue impact was positive despite a drop in average order value. The assumption that the flow was always an asset turned out to be conditional.

The System in Action

The new platform gave us something we'd never had: the ability to test meaningful changes without development cycles. That flexibility opened the door to questioning assumptions we'd never been able to test before — including some long-held beliefs about what was actually driving revenue.

Two years after launch, Hotjar session reviews revealed that a long-standing upsell flow — long viewed internally as a major revenue driver — was actually causing high-intent users to drop off during peak season. After reviewing Hotjar videos and forming my hypothesis, I coordinated with my manager in choosing a peak traffic window and ran a controlled test. Testing validated my hypothesis: removing the flow drove a 30%+ lift in conversion — net revenue positive despite a drop in average order value.

Outcomes

The migration produced lasting structural changes to the sites — new content flexibility, new testing capabilities, and a fundamentally different way of making decisions.

• Structured experimentation surfaced a single behavioral insight that drove a 30%+ peak-season conversion lift
• Evolved from zero structure to a fully mature, repeatable experimentation program
• Content slots across all touchpoints gave the entire organization autonomy to test, iterate, and move without dependency
• Hotjar session reviews became a recurring practice, surfacing behavioral insights shared across teams
• UX and accessibility standards formalized for the first time
• Built the organization's first shared design system, giving the design team a consistent foundation across all four brands

Reflection

This project changed how I approach design. Watching high-intent users struggle through a flow that everyone assumed was working taught me that clarity matters most when stakes are highest. It also taught me to question my own assumptions and sit with the results rather than jump to conclusions. But the biggest shift was this: designing for growth, building the foundation that allows a product to learn and improve over time, can be more valuable than designing for the moment.