Amazon Ads

Behavioral Segmentation for Advertiser Education

Context

Educational programs for Sellers and Vendors relied on broad email and banner distribution to maximize webinar participation. The dominant logic was volume: reach as many advertisers as possible and hope attendance followed.

Structural Problem

This approach burned high-value advertisers with repeated generic invites, wasted resources on low-propensity segments, and ignored differences in HVA completion, engagement depth, and format preference. The system optimized for gross reach rather than behavioral relevance.

My Role

I designed the segmentation approach, partnered with a senior analyst on experimentation, collaborated with channel owners on rollout, and wrote SQL to operationalize the audiences. The objective was to redesign the education distribution model around advertiser behavior, not just improve campaign metrics.

Outcome

The new model increased open rate from 35% to 40–45% and CTR from 1.3% to 2%. It was reviewed, approved, and rolled out across all regions offering webinars at Amazon. The work later expanded into a broader multimodality strategy based on format fit and structured learning progression.