Reviews

Overview

We know that buyers look to reviews to assess other buyers’ feedback and social proof to form an opinion about seller credibility.

Our task was to make reviews more helpful by tailoring them to quick-decision and deep-dive buyers, and by exploring how review feature tags could support other moments in the shopping journey

Problems to address

• No ability to filter or easily find negative reviews

• Unable to view more than 3 - 4 reviews at a time

• Confusion around counts/numbers especially for new buyers

• Confusion around item vs shop reviews

• The reviews module is not super attention-grabbing and blends into the rest of the page

Outcomes and learnings

• 14 tests run

• We found no evidence that deep divers convert at a higher rate, but conversion rate is higher among visitors that engage with reviews.

• Even though reviews are assumed to matter more for expensive items, only about a third of high-stakes shoppers scroll to them

• Visitors usually check reviews at least once per visit, but not on every listing, which suggests we’ve reduced mental effort and made it easier to buy.

Next steps

Our reviews update initially drove a GMS lift, but we’re not confident it builds long-term seller trust. As we add more seller credibility signals to the listing page, we need reviews to be non-disruptive during early triage, easy and useful for buyers who want to dig deeper, and transparent and simple to interpret.