Search Results

User Research | UX Design | Visual Design
Project Overview
StyleSeat is a two-sided beauty marketplace for beauty professionals and clients to find each other. I was hired to develop the demand-generating side of the marketplace by building an amazing discovery experience for everyday users looking for beauty services.
Search is the backbone of the client experience. Matching discovery users with the perfect stylists either coming in from SEO or downloading our native app, and letting returning users book with their favorite pro.
First time user booking rate (+4.5)
Bounce rate
Positive qualitative feedback from clients and stylists

Existing Design

In-person interviews with independent stylists and salon owners to get a sense of how they want to showcase their business. sessions with non-StyleSeat users to understand their first impressions of the current product as well as subsequent sessions with prototypes throughout the process.

In-product client-side surveys asking things like "why did you not book with this provider?"if a user bounced after a profile viewing.

Google surveys for non-StyleSeat users to get a sense of how consumers make decisions about who to book with for their beauty needs.

AB testing small things like badging and information hierarchy on list cards.
The Research Plan
Booking and engagement for first time users was low. Most bookings were pro-driven, i.e. stylists handing out their URL, rather than users coming to the site to discover new stylists on our platform.

Users were typing very general terms like "hair" and bouncing because they weren't seeing the results they wanted.

Browsing was arduous. Users were pogo sticking between search and profile pages looking for key information like booking times and services provided, producing a high bounce rate for first time users.

Cover photos were un-dynamic and default selected by the stylists while setting up their profiles. Because of this, the photos users were seeing while scrolling through search results were often irrelevant to their gender, hair type, or style preference which eroded their trust in the salon's ability to cater to their specific hair needs.

Users couldn't easily understand we were a booking site. Most users on landing assumed we were a yelp-like directory rather than a booking site.

More information wasn't exactly better. Relevant information was helpful but anything else was distracting to users and might deter them from clicking through to a profile page.

The Solution on Mobile

Switching cover photos to photos of salon interiors was crucial to help reduce image bias.


New Features

Availability pills
On-demand booking
Mobile experiences are increasingly about last minute, on-demand needs for people who don't plan ahead. Available time slots, especially for lower consideration bookings like nails and barbering, are the main considerations before booking on mobile. Putting time slots up front prevents frustrating back and forth from search to profile as users know at a glance when stylists are available. This UI also has the added benefit of letting new users know right away StyleSeat is a booking site.
A cheap AB test showed a 4.5% increase in booking
We had a hunch being able to book directly from search would be a quick and valuable for users, but couldn't invest in building out an entire flow before testing that theory. We surfaced the availability pills and watched tap targets and bookings, and there was a 4.5% increase in bookings in the AB test group with availability pills, which gave us signal that building out a booking flow from search was worth while.

Book from search

Improved Search UX 
The new nav bar acts as a value prop that lets users know right away StyleSeat is a booking site. Surfacing inputs for all of your needs right up front de-emphasizes the need for filters, an element that was barely engaged with in the previous UI. Interstitials with suggestions for inputs do a lot of the heavy lifting for a user, making them more likely to find what they need.

AB Test Results
Our data scientists created a service map with categories and sub categories based on raw user inputs. This inspired us to create the service menu picker, helping guide users with their inputs to a second tier of specificity, resulting in more accurate search results.
Control Search Interstitial
One single level of specificity with location only surfaced after the search bar has been clicked. Overly reliant on users knowing how to spell things like balayage or know the names of things like nail gels.
Two levels of specificity to choose from in the service menu picker. The location and time picker was fired separately. This led to more usage of the search bar, more bookings and cleaner data.