Etsy

Adding an AI-generated summary feature to help improve clarity and insight on the reviews of a product.

Background

Etsy, an online marketplace for handmade items, attracts buyers who value authenticity and creativity. Reviews play a crucial role in helping buyers assess custom products, but sifting through them can be tedious and time-consuming.

To improve this experience, this project focuses on an AI-generated summary feature - this is introduced at the top of the review section, highlighting key points and overall pros and cons.

Role
UX Researcher
UX / UI Designer

Project TypeAdding a feature

Timeline
75 hours

Tools
Figma
FigJam
Optimal Workshop
Whimsical

1 / Research

What did people look for when looking at product reviews? What were their priorities and frustrations? These were the questions I started off with as I began research.

What did users think about online product reviews?

While most online marketplaces had ways to categorize reviews, it was still a time-consuming process to decide on purchasing a product. Everyone has their own way of prioritizing how they want to purchase a product, and I wanted to understand what users focused on the most when reading product reviews.

What is the problem we're solving?

Buyers find it time-consuming to comb through hundreds of individual reviews to find specific details, and not only can this present a time commitment for the buyer, but in some cases provide little to no results.

User Goals

Higher trust of product authenticity with the new feature.
Reviews are now personalized for the user.

Business Goals

Increased buyer / user engagement.
Reduced return rates.
Improved buyer experience.

What is our research goal?

To understand how buyers currently assess reviews, their method and process on doing so
To identify buyers' top pain points when reading reviews for online products, and what can be done to make this experience more seamless

User Interviews

I conducted interviews with 5 people, and focused on their experiences with:

how they went through product reviews
what they prioritized
how time-consuming was the process of reading reviews

Affinity Mapping

Organizing the findings and data from user interviews produced some interesting results on what users prioritized, when reading reviews and what they desired for.

Key Insights

Users would like to see a breakdown of all reviews, or a summarization.
Users wanted to see reviews personalized to their preferences.
Users tend to look at the most recent reviews first.
Users liked to see the product attributes from reviews (e.g. color, material, quality, etc.)

Competitive Analysis

To note, the AI-generated summary feature was already present in several other online marketplaces - Amazon was the most notable one as it had already implemented this with its own review section. However, many of these were not personalized for the user, and still warranted users to read through reviews for further info.

Key Insights

Very few online marketplaces have the feature to generate a summary of all reviews at this time.
The ones that do, still lack personalization for the user's product preferences.
The review summaries also lack additional info on the product details.

2 / Define

User Personas

Based off our user interviews and research, meet Melanie and Janet!
‍Both embody the traits, goals, and challenges they encounter as identified from my research goals and statements.

Feature Roadmap + Prioritization

The primary features were identified, and would be added in this order based on what was identified during the research process.

AI generated review summary of product
Product Preferences Quiz
Product customization filters

User Flows

To see how this could be put into practice, I created a user flow to visualize how a user would interact with the new AI-generated summary review feature.

Using the AI-generated summary review feature

Low-fidelity wireframes

With a user flow in place, I sketched out low-fidelity wireframes, highlighting the new features as well focusing on personalization & efficiency of reading reviews.

3 / Design

Etsy's current design system

Etsy already had a strong design system in place, and they focused on utilizing re-usable components to ensure consistency and stability within their platform.
‍
To ensure that the new feature(s) would integrate seamlessly with their platform, I had to ensure that it was:

a non-unique component (can be used anywhere on their platform)
was clearly visible for the consumer (on the screen)

Initial concepts

Adding the toggle

For the toggle feature, this would allow the user to search for products based on their personalized preferences, which many users expressed a desire for during interviews. In turn, this would also improve the results that users received, as it was tailored exactly to what they were looking for.

Before adding the toggle, users would need to input their preferences for products by taking a short survey upon visiting the Etsy platform. It wouldn't take more than 2 minutes of their time and could easily be skipped if the user preferred to do so.

Product Preferences Quiz

Now, the challenging part here was incorporating it into the main search screen for products, and where to fit it accordingly. I decided to place the toggle under the search bar, in the middle of the screen so the user would have clear visibility of the toggle.
Selecting the toggle would change the current products on the screen based on the user's product preferences prior.

Adding the AI review feature

Adding in the final touch, the AI-generated review summary, was a step that required looking back at the current competitors' features.
‍
First, to acknowledge the users' personal product preferences, I added in a section that identified key features about the product, so the user knew it adhered to their personal preferences.
‍
Then, the AI-generated review summary was added, which included a detailed summarization of the product reviews, key features about the product, and quotes from buyers of the product to enhance trust and authenticity of the product.

4 / Testing

Usability Tests & Results

Tasks

Interact with the product preferences toggle.
Successfully interact with the AI-generated review summary.

Success metrics

There was a 95% or higher completion rate of the tasks.
Less than 5% of errors encountered (this included navigating to a different page, mis-clicking, unable to complete the task, etc.)

Testing method

Via video call, all participants conducted in a moderated test by running through the final prototype.‍

Results

All users successfully completed both task flows with no issues, and did not encounter any errors!

Iterations

90% of the participants did find this feature helpful and intuitive, but there were some slight modifications to be made:

Final Product

View the prototype here

5 / Conclusion

Challenges

Working with an existing product and learning their current design system, as well as where to fit new features within the product, was quite tricky.
Etsy, in particular, is a product known to have made many iterations over the years due to consumer feedback and observation. So, figuring out where to insert this new feature meant understanding how their platform operated, how their current features worked, and what options were visibly available for the user.

Final thoughts & next steps

This project allowed for me to gain insight and experience on designing a feature for a pre-existing product, as well as understanding the current use cases of how users utilized the product, and how to work with their needs.
With the completion of adding in this feature and receiving positive feedback from the results, this project truly demonstrated how users want to feel informed and empowered to make their own decisions.
Providing a way to give them this feedback, as well as cutting out the work for them, plays a huge role in retaining user loyalty and satisfaction for a platform.
Next steps would be to observe and improve upon the frequency on how often the AI-generated review summary is interacted with by users.