My Role
Title
Associate Director, UX Design at eSalon, a D2C eCommerce company
Timeline
Five weeks
Tools
Figma, FullStory, Baymard Institute
Key Outcomes
Four attempts had already tried to fix this upsell and failed. The problem wasn't the design, it was that nobody had aligned what users wanted with what the business needed.
I started from scratch, diving into research and data to understand why people were saying no. Then I built a strategy that connected customer value with our revenue goals.
A/B testing confirmed that results exceeded expectations: -2% conversion drag vs our -7% threshold, while customers started taking the upsells, boosting both average order values.
Context & Challenge
I transformed eSalon's confusing upsell experience into a clear value proposition.
In late 2023, eSalon faced a common subscription challenge: how to grow Average Order Value (AOV) and Customer Lifetime Value (LTV) without increasing customer acquisition spend.
haircare products to a customer’s recurring shipment. If you’re already getting personalized color, why not complete the routine?
Upsells are a critical revenue lever for eSalon's subscription model—helping offset subscriber acquisition cost (SAC) by increasing lifetime value early in the customer journey.
One clear opportunity was in the subscription upsell step of the custom hair color ordering flow, a natural moment to suggest adding
But the existing upsell wasn’t working. The page was cluttered and text heavy, with abrupt price jumps that left customers feeling surprised and skeptical, especially those arriving from deep-discount ads. Most skipped the offer entirely, never seeing the real value behind it.
UX expert Felix Haas explains why the cluttered, overwhelming approach was fundamentally flawed. “Great UX is about reducing decisions... When we’re overloaded with choices - even small ones - our brains get tired. And tired brains don’t convert.”
I led the redesign of the upsell experience, aligning product, marketing, and engineering around a simple principle: make the value clear, fast. Using behavioral analytics, competitive research, and UX best practices, we restructured the flow to surface benefits upfront, simplify pricing, and give customers a compelling reason to say “yes.”
Discovery & Strategy
I used research to uncover why users weren't taking upsells and then aligned stakeholders on the opportunity.
I started by diving into the behavioral data to understand what was actually happening. FullStory sessions revealed a clear pattern: users were toggling between price tabs but barely scrolling to see what they'd get for their money.
The timing made things worse. Users hit the upsell right after completing a lengthy quiz to customize their hair color, just when they thought they were done. Instead of feeling like a natural next step, it felt like an interruption that added decision fatigue at the worst possible moment.
The problem became obvious: people were fixated on price comparisons while completely missing the value proposition.
I looked at how other subscription brands like Madison Reed, Glossier, and Billie handled similar moments. Most were introducing upsells earlier in the flow or making the value much clearer upfront. I proposed changing the placement to leadership, but they did not feel it was worth the development effort to restructure the entire flow.
The research insights led to a hypothesis: users arriving from deep-discount ads were highly price-sensitive, and our bundled pricing without clear benefits was killing adoption.
I pushed the team to align on a baseline metric for success: we could accept up to a 7% conversion drop if enough customers took the upsell to hit our LTV goals and drive profitability despite the initial loss.
Design Process
The solution: show what mattered, not everything.
But getting there meant solving several challenges across different teams.
I stripped each option down to essentials: one benefit-focused sentence, large product imagery, and lots of white space. This went against stakeholder instincts to "show more value" by listing everything.
I advocated for simplicity using our FullStory data to show how users were scrolling past the price, and examples of successful competitors using simple upsells helped make the case.
The biggest hurdle was convincing engineering to decouple the pricing structure. The current system bundled everything together, creating sticker shock. I needed each upsell to feel like an individual choice with clear value, not just "pay more for more stuff."
The engineering team initially flagged the complexity. They wanted leadership buy-in for the extra effort to decouple the pricing. When I presented my research findings and recommended solution, leadership agreed the development work was justified.
One design decision came directly from user behavior I'd observed in Fullstory: users were tapping on static images expecting more information. So I made each upsell tappable, opening detailed overlays for users who wanted to dig deeper. This let me keep the main page clean while still supporting transparency.
Before
A wall of text and jarring price jumps left users feeling overwhelmed.
Most just ignored the upsell and moved on.
After
Clean, personal recommendations that actually felt helpful,
making it easy for customers to see the value and savings.
Outcomes & Impact
I proved that simplicity could drive both user satisfaction and business results.
Results Beat Expectations
Conversion drop was limited to -2%, beating the stakeholder-approved threshold of -7%. We had prepared for some user drop-off but ended up with results that exceeded our projections.
Upsell attach rate increased, resulting in a meaningful lift in LTV for new subscribers. The clearer value proposition meant customers weren't just seeing the upsells, they were actually choosing them.
Broader Organizational Impact
I proved that focused upsells could outperform "kitchen sink" offers, changing leadership's default approach to bundle design.
The research-driven approach shifted how teams evaluated upsells, prioritizing behavioral evidence over assumptions about customer preferences.
Leadership Reflections
I rebuilt stakeholder confidence by grounding design decisions in evidence.
The hardest part of this project wasn't designing the upsell, it was overcoming skepticism from four previous failed attempts. Everyone was wary of another redesign, and honestly, they had good reason to be.
The difference this time was grounding the work in research and behavioral evidence instead of gut instincts about what customers wanted.
I knew I had to get everyone aligned on what we were actually trying to achieve before we designed anything. So I pushed the team to agree on a baseline metric for success and called out the real risk of hurting conversions if we didn't handle the placement carefully. Those guardrails upfront gave me room to design something the team could actually get behind.
Trust came from being transparent about what went wrong before: showing the cluttered layouts, the abrupt price jumps, the engagement data that told the real story. Then demonstrating how each design decision directly solved those specific problems.
The leadership lesson hit me clearly: credibility comes from evidence, and getting alignment on success upfront is what lets a design solution actually stick and make an impact.
What I’d Do Differently
Think Beyond the Immediate Fix
If I were tackling this today, I'd start by thinking big picture. Instead of just fixing the broken upsell, I'd push for a conversation about where and how upsells should live throughout the entire customer journey.
Hitting people right before checkout feels pushy and can make them abandon their cart entirely. There could be better moments, like after they've received their order or when they're browsing related products, where an upsell actually feels helpful instead of like we're just trying to grab more money.
Personalize the Experience
What if we could look at someone's purchase history, how they browse the site, or even signals from the ads that brought them to us, and use that to predict which products they'd actually want?
AI could make these adaptive experiences possible, and customers expect that level of relevance. It could turn static offers into something that actually gets better over time as we learn more about what each customer values.
