Perfect Pair Quiz for MeUndies


What We Did

Strategy, Product Design, Prototyping, A/B Testing

My Responsibilities

  • Design strategy

  • User experience

  • Task flow & wireframing 

  • Definition of logic for development

The Problem

New customer conversion was below our target goal for the year.

Possible Issues

  • People coming to for the first time are having trouble finding the right product

  • People unfamiliar with MeUndies don’t know enough about the brand

  • We’re overwhelming people with choice

Our Hypothesis

People will be able to find the right product more quickly and with greater ease if we explain the value of our underwear and narrow their selections to just our core product line.

We Started with Testing

The potential impact of a product is difficult to measure precisely, so we wanted to move quickly and test our hypothesis with the least amount of effort from all members of our team: product, design, and engineering. Testing our hypothesis would also make sure that we were focusing on a project that would move the needle toward our goal.

In the initial test, we focused on introducing key value props earlier in the customer journey and reducing the amount of choice presented to new users. We created a homepage banner that was only shown to new users and introduced MeUndies as The World’s Most Comfortable Undies. Additionally, we backed up this claim with details that would normal be found on a Product Detail Page. The call to action (CTA) on this homepage banner–Find Your First Pair–would lead users to a product listing page only displaying the core underwear products for men and women (a total of eight products).

Our control in the experiment, the original homepage banner, focused on a new print release and led users to either reserve the print (if they are a member) or shop the collection for it. 

The test was completed and launched in a day with minimal effort from design and no engineering. We saw an increase in new customer conversion after running this test for 10 days.

Control banner, test banner, and test product list page.

Rethinking the Shopping Experience

When we started this project we suspected that it could potentially have an impact on improving new customer conversion, but we didn’t have any data to back it up. The results of our initial test validated our hypothesis and gave us confidence that we were pursuing an exciting and relatively well-defined idea. We also gained approval from key stakeholders to continue ideating on this initiative, allowing more time for design and development to work on this project. 

We wanted to continue moving quickly, so design and development began white-boarding potential solutions with a focus on mobile, since approximately 70% of MeUndies users are on mobile devices. During this process we identified two distinct approaches–curation based and self-service selection–which could designed with shared elements and patterns to reduce the amount of time spent in development.

Prototype 1: Perfect Pair Quiz

Curation Based Recommendations

The Perfect Pair Quiz is an interest based flow that generates a set of curated results based on the user’s responses to questions about preferred cut(s), color, and activities. A user can end up with a list of up to six pairs of underwear that are recommended based on inventory available in their size and their indicated preferences.

Perfect Pair Quiz wireframes used to begin development

Perfect Pair Quiz Development Prototype

Prototype 2: Perfect Pair Picker

Self Service Shopping

The Perfect Pair Picker is a streamlined flow that generates a single result based on preferred cut and the user’s size. The colors shown are only those available in the size selected by the user.

Perfect Pair  Picker  Development Prototype

Perfect Pair Picker Development Prototype

Focus on the New User

With the new user in mind, we wanted to figure out a way to include information that would be helpful in making the most relevant selections throughout the task flow. For example, in the cut step we included a brief description each product that would help someone tell the difference between cuts (e.g., boxer brief has a longer inseam than the trunk). We also pulled information out of the fit guide to include the natural waist sizes within each size selector and a recommendation for what to do if a user fell between two sizes.

It’s Working

Numbers and Results

Data and findings pulled from Google Optimize and Analytics

Data and findings pulled from Google Optimize and Analytics

Our goal was simply to create an experience where new users could provide some details about the type of underwear that best fit their needs and interests, and not have to worry about finding the styles, colors, and prints most relevant to them. 

We launched the two prototypes against our original control (the product listing page experience) and after 10 days of data we learned that both the Quiz and Picker Experience improved conversion and AOV compared to our control. Ultimately, the Perfect Pair Quiz outperformed the Picker.

During the test we found that:

  • Perfect Pair Quiz earned 18.8% more per session than the control

  • Perfect Pair Quiz had a subscription rate improvement of 25.32% over the control (despite not have a direct callout to membership until the cart screen)

  • Perfect Pair Picker had a subscription rate improvement of 29.11% over the control

What’s Next

The team decided to move forward with the Perfect Pair Quiz. The look and feel of the Quiz has been updated since completion of the test and is now ready for production. The feature has been prioritized for development and should be released by the end of the fall–just in time for the 2018 holiday season.

We also have plans to continue iterating on the Perfect Pair Quiz once the feature has been developed and deployed to production–testing the addition new questions, order of existing questions, introducing The Membership value props, and more.



Desktop, Tablet, Mobile Web


  • User Flows

  • Sketching

  • Design + Dev Pairing

  • Medium Fidelity Wireframes

  • A/B Testing


  • Whiteboard

  • Sketch App

  • InVision