App, Case Study, Product, Reverb, Web

Related Searches

The Team

Product Designer: Alyssa Welch
Product Manager: Colleen McClowry
Web Developer: Natalee Anfuso
App Developer: John McCartney (iOS), Sean Larson (Android)

Client

Reverb, Engage Team in collaboration with Search Science

Overview

Reverb’s Search Science team uncovered an interesting user behavior: more than half of the top 500 search queries were broad strings. About one-third of those searches were followed by another within one minute. This pattern of behaviors raised some flags on the team. Could we be leading our users to better utilizing our search functionality? How might we begin to lead them down a path where they are narrowing down on what they really want?

The team hypothesized that surfacing query suggestions for broad terms will make it easier for users to get to their intended items faster.

The Goal

To an increase in conversions from search, with clicks on search suggestions and clicks on listings in the search grid laddering up to that.

Exploration and Discovery

A brief round of competitive research and analysis was performed. To achieve timing that worked best for the teams involved, this served the purpose of a simple gut check to validate visual patterns and potential placements.

Process

The Variants

The team wanted to test a variety of module strategies on the Search Results page.

Pills versus Image Cards
Text-based pills seemed to be the standard in related search modules, however we saw value in testing imagery in this space as our instrument-loving audience had previously responded favorably to this style of wayfinding.

Placements
One experience would test the Related Search pills near the pagination, which visually matched patterns seen elsewhere, with the Related Search image cards above the grid. The contrasting experience would flip those placements, putting pills above the grid and image cards below.

We planned to test both related search placements on web and iOS platforms. For Android, due to a need to balance resources, we chose one direction and decided to draw conclusions based on comparisons between other platforms.

Engineering Handoff

I prepared the file for a smooth engineering process by documenting differences between responsive sizing on the web platform, detailing from where images were to be sourced, outlining the minimum and maximum requirements to surface this module, and accounting for best and worse case search string scenarios (truncation rules).

Impact

The experiment was successful and ultimately rolled forward.

On web, we saw an 8.91% increase in CTR and a 3.5% increase in search conversions. The new design was particularly impactful for logged out buyers, presumably less familiar with either music gear or the Reverb brand, who saw a significant increase of 8.85% in search conversion.

On iOS, we saw a 2.23% increase in search conversions. Additionally, increased search functionality resulted in a 1.66% increase in the app’s daily active users.

8.91%
increase in CTR on web
3.5%
increase in conversions from web search
8.85%
increase in conversion among logged out users
1.66%
increase in app daily active users

Final Designs

The variant with Pills above the grid and Image Cards below the grid won on web and was rolled forward.

I hypothesize this variant performed better as the text-based pills were less distracting at the top of the page. This allowed a user to complete more of their intended task in a linear fashion–scrolling down through the search results and then interacting with the Image Cards beneath when the results did not meet their expectations.