Case Study ❂ Amazon

Trusted Book Recommendations

Project overview

The product is

  • Amazon Books connects readers to content that aligns with their interests while streamlining their discovery and buying process.

Why this product exists

  • To support book customers with personalized recommendations, seamless purchasing, and flexible reading formats.

What does success look like

  • Improve the personalization and clarity of book recommendations using contextual metadata, editorial content, and user behavior.

  • Enhance trust and engagement with recommendation widgets using micro UX patterns and streamlined content integration.

The challenge

  • Despite robust data science behind Amazon's book recommendations, users often perceived them as impersonal or irrelevant, often leading to lower engagement and missed conversion opportunities.

  • The challenge was to earn trust by humanizing the book recommendations so they did not feel like the algorithm.

  • Design lightweight but meaningful UX interventions that make the UX/UI smarter, more transparent, and trustworthy.

My approach to the project

❂ Learn the ecosystem to see how might we use existing patterns to create new solutions

  • Conducted UX audits of the Amazon Book Review, book detail page, and book thank you page to see where the customer experience overlapped and needed connection points.

  • Worked cross-functionally with UX research, Amazon Editors, and the Goodreads Team to gain insights into all the moving parts of the team.

  • I worked with UX research to review customer feedback to identify where the lack of trust around recommendations comes from.

    • A key insight: users don’t trust recommendations when there was no visible photo of a person or the context sounded like a machine.

  • Partnered with the Goodreads UX team to compare how they present similar content and digest how their users perceive trust.

❂ Design scalable mirco book recommendation for anywhere in the shopping experience

  • Introduced an editorial blurb component to make book recommendations feel personal and trusted by including a photo and quotations.

  • Accessed content from the Amazon Book Review and Goodreads to populate the book recommendation blurbs.

  • Designed and tested several micro interactions with the editorial blurb component to show users an actual human was recommending the book.

  • Created a north star design pattern to include other human centered voices like authors, Kindle highlights, and more.

Outcome

  • Leveraged Goodreads book recommendation reputation by adding Goodreads star ratings to book detail page to build Amazon book recommendation trust.

  • Created design road map for integrating other human voices into the book recommendation engine.

  • Mirco UX interactions, when backed by user insight and executed thoughtfuly, can have an large impact on how customers interact with your product.

If I had more time

I would have liked to see where the blurb card fit into the search, detail page, and thank you page, resulting in the growth of trust of human centered book recommendation.

Additional case study details

mirco UI interactions

Book customers often value accolades such as awards and bestseller status as indicators of a book's quality and relevance, helping them make informed choices in their reading selections. These accolades provide customers with a sense of assurance that they are investing their time and money in a worthwhile literary experience.

Currently, customers are presented with several types of accolades throughout the book's detail page, both above (ATF) and below the fold (BTF).

In the glance evaluation section (ATF) customers see a single ranked badge. We estimate ~30% of our top titles will have more than one badge at a given time.

In the editorial reviews section (BTF) customers see an array of trusted voices and media sources. We hypothesize these accolades are being missed by the customer due to their location on the detail page.

In Q4 2022 we ran an email campaign with multiple badges on a single title. This experiment resulted in a statistically significant increase of click through rate (CTR) by 6.25%.

How might we create a combined multi-badge and editorial customer experience ATF to help customers simplify their evaluation of titles and increase their purchase confidence?

COMPONENT SYSTEM

Many book customers struggle to find curated book recommendations from trusted sources, leading to a frustrating experience of sifting through vast amounts of irrelevant or low-quality titles. As a result, customers do not discover books that align with their interests, preferences, and reading habits.

Currently, a majority of customers come to the bookstore with a book title already in mind due to low prices, large selection, and ease of consumption. The bookstore lacks discoverability mechanisms throughout the customer's shopping journey.

The most simplified customer's shopping journey is through browsing the book store navigation or from search/results. Customers select a title, and after purchase, see a thank you page.

The Trusted Voices team has an opportunity to present customers with curated book recommendations throughout their entire shopping journey.

Our hypothesis is by creating a UX solution that provides customers with personalized and reliable book recommendations throughout their shopping journey, it will help customers simplify their evaluation of titles and increase their purchase confidence.

View Full Visual Story