e.l.f. Cosmetics tailors the beauty regimen to every shopper

Top beauty retailer delivers highly-targeted experiences across categories, product detail pages, and mobile menu, achieving 3.2X ROI from deep learning-based PLP personalization

Download the PDF * Email not necessary to download
increase in mobile menu clicks after personalizing the menu based on users’ past shopping behavior
3.2x ROI
from testing the real-time reordering of site visitors’ product listing pages (PLP) using AdaptML
uplift in revenue per user by scaling up the delivery of personalized product listing pages during Black Friday and Cyber Sale week


Founded in 2004, e.l.f. Cosmetics launched with a simple mission to create high-quality skincare, makeup, and beauty products at an accessible price point. From day one, e.l.f. has aspired to place its customers at the core of all product and marketing decisions. So when e.l.f. sought to provide an optimal shopping experience for all, its team partnered with Dynamic Yield to help them deliver.

After first utilizing the platform’s first-party integration, recommendations, and menu personalization capabilities, e.l.f. witnessed a 4.2% increase in revenue per user and an 18% uplift in customer engagement. The team went on to leverage the technology’s Personalized Product Listing Page solution to further showcase the most relevant items and streamline the discovery process.

Once able to automatically reorder category pages using the experience optimization platform’s deep learning algorithm, AdaptML™, e.l.f. saw a 29.7% uplift in revenue per user and 3.2x ROI from the tool. Further, based on projected annual incremental revenue gain, a 13.5x return on investment is expected upon rolling out personalized sorting across all category pages.

“Listing pages can be cumbersome to navigate. With Dynamic Yield’s AdaptML™ PLP solution, we have made them fully contextual and adaptive based on who the shopper is during a critical phase of the buying journey. Now, products displayed are more interesting and relevant to the individual, allowing for not only simpler but deeper exploration.”
Ekta Chopra, Chief Digital Officer,
e.l.f. Cosmetics
“Dynamic Yield has been instrumental in helping us uncover the different types of audiences coming to and interacting with the e.l.f. site, enabling us to truly cater to each beauty lover’s specific needs. The platform has allowed us to easily test new strategies and optimize on the fly for quick, meaningful results.”
Shana Rungsarangnont, Digital Products Manager,
e.l.f. Cosmetics

The Challenge

While e.l.f.’s products are sold at big-box retailers, including Target, Walmart, and Ulta, its eCommerce team was focused on growing the company’s digital presence. Realizing PLPs, PDPs, and the mobile menu are all critical to the shopping experience, the beauty retailer wanted to move beyond a one-size-fits-all approach for these areas of the site. Set out to optimize them towards each visitor’s needs, this meant they would need a solution to:

  • Ingest first-party data and identify high-value segments
  • Optimize the mobile experience for product discovery
  • Surface the most relevant product recommendations
  • Personalize category page results according to real-time customer behavior


Onboarded and activated historical first-party customer data from day one
Before integrating with Dynamic Yield, e.l.f. used Custora, a cloud-based customer analytics software to help identify insightful audience segments. However, the team wanted to further leverage the Custora segments to take immediate action on the site and customize experiences. The ability to ingest Custora’s first-party data with Dynamic Yield ensured the beauty retailer would be able to create highly-targeted campaigns while not losing the rich first-party historical data.

First-party audience examples:
Customized mobile menu based on user affinity and browsing behavior
With limited real estate, e.l.f. wanted its mobile menu to resonate with each shopper. To effectively do this, the team decided to personalize the menu based on browsing history, allowing customers to quickly find products best suited to their needs. New visitors were shown high-level categories while returning visitors were exposed to subcategories of the main category they had previously shown interest in. For example, if a shopper had browsed the “skincare” category, the menu bar would include “cleanser,” “face masks,” and “treatments.” With these menu optimizations, e.l.f. saw an 18% increase in click-through rate, driving visitors deeper into its product catalog.
Optimized recommendations on PDPs after a bakeoff between DY and alternative recommendation vendor
Knowing PDPs are a valuable piece of the product discovery phase, e.l.f. chose this space to optimize its recommendations. And to ensure the most relevant products were shown to each visitor, the team decided to run a bakeoff between its two recommendation engines. Each vendor served a “Viewed Together” strategy in e.l.f.’s “You May Love” section and after reaching statistical significance at 95% in two weeks, Dynamic Yield proved the strongest engine, generating a significant uplift in CTR (+ 23.2%) and revenue per user (+4.2%) compared to the other vendor.
Generated the optimal sorting order of items by exploiting the co-occurrence of products in a user’s browsing history
With 48% of e.l.f. shoppers visiting its product listing pages, the beauty retailer’s main goal was to help visitors find the items they were looking for more easily. Recognizing the massive optimization opportunity at hand, the team hypothesized that it could increase key business metrics across the site by tailoring the sorting order of its category pages.

To test this theory, e.l.f. ran an experiment targeting nine of its most popular category pages for eyeshadow, lipstick, lip gloss, lip balms & treatments, face primer, and more with a 50/50 traffic split against its native default sorting order. After running the test from the end of October to late November, the product listing page personalization led to a 3.3% increase in revenue per user across all categories, yielding a 3.2x ROI on its initial investment in the solution, which leverages a deep learning strategy to surface items the user is most likely to interact with next.

Primed to not miss out on any incremental revenue gains during the busy Black Friday / Cyber sales events, e.l.f. then scaled its tailored PLP experience to the majority of the brand’s traffic. This change led to 29.7% more revenue per user being generated over the short-lived Cyber week. And today, the team continues to leverage Dynamic Yield’s deep learning algorithm to automatically optimize the sorting order of products on its listing pages, which based on projected incremental annual revenue, translates to a 13.5x in ROI.

The Key Takeaway

e.l.f. combined its commitment to providing the best possible products with that of a superior customer experience by matching its visitors to the most relevant items. After deploying Dynamic Yield’s technology, the beauty brand was not only able to tap into the important audiences engaging with the brand, but also reflect their needs and preferences across the shopping journey. By strategically implementing personalization across its PLPs, PDPs, and mobile menu, the eCommerce team was able to increase customer engagement by 18%, boost revenue per user by 29.7%, and will ultimately produce a 13.5x return from investing in personalized sorting on category pages.