Examples of personalization concepts on a mobile phone

Personalized Shopping

Creating a seamless, relevant and useful personalized omni-channel shopping experience for the AEO customer.
The Project

This project was founded in the fact that American Eagle Outfitter customers share the same digital shopping experience despite having unique needs and intentions. Our nimble cross-functional team was hyper-focused on solving customer and business problems via data-driven tactics with the ultimate goal of creating a highly personalized shopping experience for every AEO customer.

My Role: Lead UX site designer, user research, wireframing, prototyping, user testing, project prioritization, workshop facilitation, vision design and presentation

The Team: User Experience, Data Analytics, Site Optimization, UI Design, Engineering, Site Merchandising

Timeline: 2+ years

The Problem
Customer Goals
KPIs

Process

This project began with a dual approach of testing easy to implement, tactical data-driven recommendations while simultaneously creating a future-thinking personalization vision in order to gain business buy-in and create a backlog of work.
Feature Prioritization

To start, the team created a backlog of ideas that we felt addressed our project goals. These ideas were based on competitive examples as well as known functionality and data gathered on our site. Each team member performed competitor research and submitted ideas that I maintained in a shared wiki page.

I facilitated and participated in workshops with our working team and key business partners to genereated further ideas and prioritize them based on implementation effort and customer impact.

The project began with small, tactical ideas that had high customer impact and lower implementation effort in order to prove the value of data driven features and to justify and obtain an annual project budget.

Marketing Workshop
Concepts genereated in an ideation workshop I facilitated
mobileExample
Impact/effort prioritization matrix session I hosted
Design + Iteration

Once the team selected an idea from a prioritized backlog, I created a draft of the mobile concept. I then reviewed the wireframe at our weekly team meeting. We would discuss what the concept was trying to accomplish, any possible assumptions we were making and how we would build it. Updates to designs were made based on these discussions.

Below is an example of a high fidelity wireframe iteration for a product recommendation module in the bag.

Product recommendation in bag version 1
V1: Multiple recommendations
Product recommendation in bag version 2
V2: Single recommendation
Product recommendation in bag version 1
V3: Single recommendation with Add to Bag
Note: We A/B tested the final "single recommendation" concept that showed a +.17% lift in RPVr.

This project allowed me design more quickly than most AEO projects because we had dedicated developement resources that didn't have to follow the standard engineering workstream. Basic ideas were generated and tested quickly, iterated on and, if needed, tested again before a concept was finalized.

Another unique aspect to this project was that our small cross-fucntional team had much more influence over the designed concepts because we worked so often and closely together. I believe this created better final products.

desktop design version 1 desktop design version 2
Dual A/B and usertesting of two design iterations of a "Featured Offers" tab that would display deals most relevant to the customer first.
Ultimately discoverability issues and technical limitations prevented this feature from going live.
User Testing

I performed interactive prototype and live site listening labs via usertesting.com to better understand the usability of our concepts.

In addition to generating interactive prototypes, I also created test scripts, ran and evaluated tests and presented findings to the working team and key business partners.

A/B Testing

Our project had great support from our data analytics team. We were able to test risky design assumptions quickly and were therefore able to learn faster and implement successful concepts swiftly.

Standard KPIs such as CVR, UPT, AOV, AUR, Revenue, RPVr, NTF and when possible, return rates, were evaluated to determine feature success.

Most of the features tested and implemented in the first year consisted mainly of product recommendation modules fueled by aggregate customer data. During this timeframe we were able to learn a lot about what helped our customer shop and what got in the way of their shopping experience.

a future vision

While learning from our basic aggregate data recommendation tactics, we simultaneously developed a strategy to create a more aggressive one-to-one site experience based on individual customer’s needs and preferences.

I designed a future-thinking customer story that highlighted concepts for individualized customer content grounded in our project goals. The story presented an omni-channel shopping experience that touched communications, digital shopping as well as a new in-store app experience.

I presented this vision to partners from all over the buisness, including our brand presidents, in order to illustrate how a personalized shopping experience could help our customers and our business.

This vision, paired with promising data from tested recommendation tactics, had the entire business excited about the project and asking how they could help make this vision come to life. This included securing additional budget and resources for the project.

A personalized landing page from the future vision story

Results + Next Steps

In 2 years this project tested and/or implemented over 18 tactics on the AEO site.

Examples of implemented tactics include:

frequently bought together module on pdp
"Frequently Bought Together" recommendations
More Looks to love recommendations on pdp
"More Looks to Love" recommendations
Underwear recommendations on pdp
"Get the deal" undie recommendations
Quick filter tabs on category page
Category popular filters
Added to Bag modal
"You May Also Like" recommendations

In 2018, we reported the following KPIs for the five combined personalization site features illustrated above:

6%
interaction rate
4x
more likely to convert
+26%
in UPTs
+21%
in ADS

In addition to driving standard KPIs, we also saw an increase in customer acquisition:

+22%
Acquisition Value
+80%
New Dual-Brand Customers
+17%
Annual Omni Spend
Evolving Personalization

After a site replatform at the end of 2019, I was tasked with designing true one-to-one personalized features. Some of these tactics included:

mobile screen showing size 4 recommendation above the size picker on pdp
Size recommendation
Added to bag modal with jean recommendations
"Get the Deal" recommendations
picked for you flag on a product image on a category page
"Picked For You" flags

Some of my favorite, not-yet-implemented one-to-one personalization ideas include:

  • Replacing AEO model images with user generated photos that look more like a known customer's size/shape. This would make it easier for the customer to assess how clothing could look on them.
  • Reordering navigation, product guides (ex: Jeans) and category page content for known customers to allow them find the products they are likely to want faster.
  • Sending a custom promotional offer to a customer at a relevant moment in time (hour/day/month/year)to incentivize them to shop.