Repositioning PairAnything as a product recommender
I transformed a legacy wine pairing tool into a guided quiz experience that personalizes wine buying. The PairAnything Wine Recommender Quiz invites shoppers to imagine how they’ll enjoy the wine, whether that’s during a weeknight dinner, or celebrating your friends’ birthday, and recommends wines that match the vibe. I wanted shoppers to feel confident buying wine they’d never tasted before, online.
Role
Product Lead & Designer
DURATION
2 months
team
CEO, developer
Shoppers answer simple, lifestyle-based questions like: “What kind of meal are you enjoying?” Based on their answers, they receive three wine recommendations, each paired with a meal suggestion to help visualize the wine in context. Shoppers can save results via email to access later.
Redefining our product
With the original pairing tool, shoppers could search for any food or wine and get a pairing. In theory, it worked well for shoppers who already knew what they wanted a pairing for. However, after analyzing the data, I noticed that only 5% of visitors searched for a pairing; a majority of users were abandoning the tool before engaging or seeing any value from it.
Through conversations with wineries, we realized that the true value of our product wasn’t the pairings themselves but rather our ability to simplify the decision-making process for shoppers. This insight shifted our perspective from framing our product as a pairing database to a conversion tool using personalized recommendations.
Before the redesign: The original PairAnything product only had a mobile view and required shoppers to type and search to get a pairing.
Before the redesign: The original PairAnything product only had a mobile view and required shoppers to type and search to get a pairing.
the problem
Nobody buys new wine online.
98%
of our survey respondents only buy wines they’ve tasted before when shopping online.
90%+
of wine is purchased in-person, where shoppers can get a recommendation from an employee. Only 6% of wine sales happen online.
This highlights a major trust gap: most shoppers don’t feel confident buying unfamiliar wine online without guidance. Through Reddit threads and conversations with wine drinkers, I kept hearing the same thing: people trust people. Interviewing wineries told us they struggle to sell wine in situations where they can't converse with shoppers in-person.
“We do really well when we’re there at the store, talking to people about our wines... but that just isn’t sustainable for us in the long term, especially when we want to expand our online DTC (direct to consumer) channel.”
- Winery co-founders describing their sales strategy
“The easiest way to pick a good bottle of wine is to go to a respected wine store and talk to the staff.”
- Redditor replying to: “I am a Newbie to wine. What are the most important things for me to know?”
“AI can’t taste wine. I really don’t want a world where AI is picking my wine.”
- Redditor replying to: “Are Soms, critics and store recommendations in danger from generative AI?”
RATIONALE
Nobody buys new wine online.
Recreate the trust of talking to a sommelier or winemaker, but online?
Help online shoppers to feel confident buying wine they’ve never tasted before?
I hypothesized that:
- If we help shoppers imagine how a wine fits into their life, they’ll be more likely to purchase and enjoy it.
- A quiz lowers the barrier to entry by guiding users through simple choices, unlike a search bar, which requires them to know what to look for before they start.
how we did it
How do sommeliers currently recommend wines?
We interviewed sommeliers and winemakers to understand how they guide customers. They start by asking for contextual information like:
“Well… what are you having? What’s the vibe?”
- The first thing a sommelier asks when making a recommendation
We also tested a competitor product, Tastry, which asks technical questions like, “Do you like bell pepper?” While this reflects how experts evaluate wine characteristics, user interviews showed that these questions didn’t resonate with casual wine drinkers. One interviewee shared:
“These quizzes don’t account for the fact that sometimes I want chocolate and sometimes I want vanilla…they should also account for mood and pairings.”
- User, after testing Tastry
This feedback reinforced our decision to focus on how users want to enjoy wine and to design the quiz for repeat use. For example, someone might take it once for a birthday party and again for a casual weeknight meal.
Tastry’s quiz qualifies traditional wine characteristics, while ours focuses on the context of how users want to enjoy their wine.
Finally, after interviewing our client Domaine Dardagny, we learned their customers typically spend over $100 per order and aren’t price sensitive. To streamline the experience and reflect the brand’s premium positioning, we omitted price-based questions and focused on lifestyle and flavor preferences instead.
Tastry’s quiz qualifies traditional wine characteristics, while ours focuses on the context of how users want to enjoy their wine.
USABILITY TESTING
Making results feel rewarding
During user testing, we noticed the results felt anticlimactic. Early designs used too much negative space, and key information required excessive scrolling, leading to drop-offs. I redesigned the quiz results page to:
- Display wine recommendations first, side-by-side for easy comparison
- Collapse quiz responses into a drawer to reduce clutter
- Vertically condense pairing section for easier skimming
Previously, quiz responses appeared first, taking attention away from the wine recommendations. Users had to scroll excessively to view all the information, making the results feel anticlimactic.
The final design shows all 3 wine recommendations side-by-side above the fold, and collapses quiz responses in a drawer. Pairings are vertically condensed into one scroll for easier skimming.
Adopted by Wineries, Expanding to Retail
Prior to launching the quiz, we found strong signals by testing the prototype with customers:
5+ wineries signed
New and returning customers of PairAnything have chosen to adopt the quiz instead of the old pairing search tool for their ecommerce site.
Expanding into retail
A wine retailer plans to use it on an in-store kiosk, prompting me to design an accessible tablet version
We launched the PairAnything Wine Recommender Quiz on Domaine Dardagny’s website in late August 2025. We’ll consider the wine quiz successful if at least 10% of quiz-takers purchase a wine, validating that the product converts shoppers, and at least 20% save their results via phone or email, indicating that end-users found our recommendations helpful.
Next, I’m designing the enterprise setup experience to make the quiz work for any beverage seller, allowing PairAnything to expand into more industries outside of wine.