Engagement Prompt



Engagement prompt are social proof messaging that are becoming more and more prominent in many e-commerce websites. It is information on a product's popularity that could help users decide on purchase decision making especially when deciding between similar products.

Prototyping Tools: Adobe XD


As the lead product designer for the project, I worked closely with the Product Manager, UX Researcher and Developers for this experimentation. I provided all documentation, competitor analysis, gathered questions, designed prototypes and specifications.


At Macy's, we offer plenty of similar products and aside from customer reviews, shoppers do not see any other content that validates the popularity of a particular product on product detail pages. In this experiment, we wanted to test if social validation can help make decision-making less overwhelming for customers. 

How might we provide social validation that would help users in purchase decision making?


What We Knew

• In 2018,  Macy's a/b tested "XX customers viewed" vs. "XX customers purchased" which showed a clear incline towards "XX customers purchased

• Technical limitations to getting real time views

• Technical limitations to show scarcity (2 items left)

• Our numbers for purchased items are much higher than competitors (see competitive analysis below)

What We Didn't Know

•  What is the goldilocks range for the "XX customers purchased" that would make the social proof compelling for the customer?

•  Can/ should "XX" be dynamic and updated in real time?

•  For saliency, what is the UI for engagement prompt—temporary or permanent?​


If  we add social proof on our product page to show the number of customers who purchased the product 


then the customer more likely gain confidence in ownership of the product


Resulting in an increased ATB rate and conversion.

Competitive Analysis

Below are screenshots from competitors with engagement prompts.

Comp analysis@2x.jpg


UI Analysis

UX Research:

Optimal Range

Our goal was to find the optimal  range of what would be the compelling number that would make the social proof compelling. 

Through survey monkey, we tested five different numerical categories; each with a low and high number. We did this to identify any inflection points between numbers that were too low, just right and too high.


We also tested across a basic t-shirt  vs. unique t-shirt but only within the 4-digit numbers. 

We asked on the survey if the number of customers who purchased this is too low, just right or too high.

engagement prompt.jpg

Key Takeaways

Engagement prompts are helpful.

63% of all shoppers agree/strongly agree that “x customers purchased” is helpful.  Of the 20% who stated this information isn’t helpful (decliners) the top reasons why include:  1) Not influenced by others; 2) Prefer reviews; 3) Want to be unique

However, “Best Seller” was deemed a better alternative by this decliner cohort

Engagement prompt graph@2x.jpg

The overall “just right” range falls broadly within the high triple digits to lower five digits

MVP Test


Treatment 1

Treatment 2 — Winner

For items sold > 500, show the units_sold metric

to drive the “XX customers purchased” 

 For items sold > 500 to 20,000 use the units_sold metric

to drive the  “XX customers purchased” 

For items sold above 20K, show “Best Seller”

UI Rapid Test


Red Text Near CTA

Pill text over image

Notification badge over image

Additional Credits


Year: 2020

UX Designer: Leopoldo Udan Jr.

Product Manager: Allison Burns

UX Researcher: David Hughes

UX Content: Kimberly La Spada

Developer: Rama Krishna Madamanchi & Gilbert Medina