Father's Day — A/B Test Optimization

Overview

After the Father's Day campaign page launched, I noticed that the click-through rate was significantly lower than expected. Rather than accepting the initial performance, I proactively identified the problem and proposed an A/B test to validate a hypothesis about the design's impact on user engagement.

This project became a textbook example of how rapid experimentation and data-driven iteration can rescue an underperforming campaign — and ultimately led to a lasting design principle adopted across seasonal campaigns.

Process

• Performance monitoring — Tracked real-time metrics after launch, identifying an unexpectedly low CTR within the first few days. Compared against benchmarks from similar seasonal campaigns (Mother's Day, Valentine's Day).
• Root cause hypothesis — Analyzed the page layout and user behavior data. Hypothesized that the hero banner's visual design was not clearly communicating the page's purpose, leading users to bounce rather than scroll.
• A/B test design — Designed a variant that addressed the hypothesis — making the call-to-action more prominent and restructuring the above-the-fold content to immediately show gift categories.
• Test execution & analysis — Ran the A/B test with a statistically significant sample size. Monitored CTR, ordered product sales (OPS), and abandon rate across both variants.
• Postmortem & knowledge sharing — Documented the full process, findings, and design principles in a postmortem report. Presented insights to the broader UX and marketing teams.

Results

The A/B test variant delivered clear improvements across all tracked metrics:

• Significantly improved CTR compared to the original design
• Higher ordered product sales (OPS) driven by better category visibility above the fold
• Reduced abandon rate as users found relevant gift options more quickly
• The insight — "show products, not just branding, above the fold" — was adopted as a design principle for all future seasonal campaign pages at Amazon Japan