eBay Listing Guidance and seller insights

The mandate

Listing Guidance began as a hackathon project and a fast external proof of concept: machine-learned models that told a seller, while they were still writing the listing, what would actually make the item sell — classifying the item into the right category, guiding price against live demand, and improving the title and attributes. It grew into a full seller insights program and became the mechanism the marketplace used to drive feature adoption and enforce listing standards at platform scale.

The organization

Started as a hackathon team, then a funded product team spanning the listing experience and seller insights.

Outcomes

• The first AI-driven guidance program built for the supply side of a B2C marketplace — classification, pricing guidance and title optimization, a decade before LLMs made the pattern obvious.
• That pattern is now industry standard: Amazon's Seller Coach, Airbnb's Opportunities, and an equivalent in almost every marketplace since.
• Drove roughly $2B in incremental GMV through seller-facing tools.
• Became the platform mechanism for driving feature adoption and enforcing listing standards, rather than a reporting dashboard sellers had to go looking for.
• Went from hackathon prototype to funded program by proving the concept externally before asking for the platform investment.