Book Shopping Cards
Create once render everywhere

Company
Amazon

Role
Product Designer

Year
2022

About the project

The Amazon Books Shopping team owned dozens of pages across seven surfaces in fourteen market places, adding up to hundreds of customer facing end points.

The primary objectives were to minimize the inconsistency across card patterns used on these pages by designing reusable components and establishing a framework called “Create once, render everywhere” that will effortlessly scale card patterns for all surfaces with a single click.

Thank you page

Books landing page

Discovery

The discovery phase started with an audit of the existing strategies for Amazon’s books shopping cards such as “Recommended for you”, “This week’s best-seller”, “Next in series”. Following that, an ideation session was facilitated to define a set of card patterns that could seamlessly cover for 80% of the current strategies. The outcome of this collaborative effort resulted in an initial set of card patterns.

Ideation session

Initial set of card patterns

Reason, Entity and Action

The cards were intentionally designed to align with common business goals, metrics, use-cases and customer expectations. Considering that the cards were going to be used by multiple teams, we created a framework called "Reason, Entity, and Action" to assist in governing the card pattern.

“Reason” is the reason why a strategy was created, “Entity” could be a book, author, an offer or an article and “Action” was the end goal of the card and how the success was going to be measured.

When an interaction solves a user pain point, identifying design solutions that work universally becomes crucial.

Challenges

The most compelling interaction patterns only worked on some surfaces where native features are available, and accessibility was always a challenge when images and text were not displayed on load.

If the right solution was a progressive reveal of meta-data to achieve the right density, we chose the right interaction mechanism that is best for each surface.

Reason, Entity, Action

Card interaction x Volume of metadata

Surfaces

  • Desktop

  • Tablet

  • Mobile (iOS, Android)

  • eReader

  • Email

Outcome

Five distinct card patterns capable of scaling to 6 surfaces showcasing 3 content types (book, author, article) were delivered.

Card patterns:

  • Shoveler

  • Hero

  • Standard

  • 4-Pack

  • List

Full set of card patterns

Design annotations

A shoveler card scaling to various surfaces

Impact

  • The investment in time, effort, and number of team members dedicated to a single card pattern decreased significantly

  • Design system documentation played a pivotal role in reducing the number of meetings and Slack discussions

  • The new book shopping card patterns coincided with Kindle Vella's successful launch

Thank you page (before)

Thank you page (after)

Kindle Vella