How Kendra Scott scaled customer research 30x

Opportunity: Bringing customer insight into a fast-moving design process

Kendra Scott's success rests on distinctive design, entrepreneurial intuition, and a deep understanding of how customers use jewelry to express themselves and celebrate meaningful moments. Guided by a commitment to family, fashion, and philanthropy, the brand now operates more than 185 stores in North America, alongside wholesale channels and digital.

That kind of growth brings a new opportunity: identifying and embracing the unique personas within a customer base that is now larger and more varied than ever.

Kendra Scott's e-commerce and digital team leads that work by personalizing the online experience for each customer profile. The team already used digital analytics, competitive research, A/B testing, and feedback from customers and store teams, with external research partners supporting larger strategic initiatives. Moving at the pace of a growing omnichannel business, the team relied mostly on A/B tests to validate designs after the build.

The team wanted to bring qualitative insight in earlier, during the design phase, where changes were faster and less expensive to make.

"A lot of the decisions that we would make were based on internal experience and opinion, or anecdotal information we might get from our customers when we interacted with them in the store," said Brad Power, Senior Director.

Direct customer evidence would help the team commit to a design decision with more confidence, said Edward Rendon, Senior UX designer.

The work to add Buy Now, Pay Later and other payment options to the site showed where that evidence could help. The team explored 20-30 design versions, and hearing from customers who used those services would help them narrow the options sooner.

The team wanted to hear from customers directly, and to do it properly. Traditionally, even a simple survey can translate into weeks’ worth of spreadsheets, emails, calls, collecting information and synthesizing. For a lean design team of one or two people covering a wide remit, that time was best saved for the biggest decisions. So they focused in-depth customer research on about two of the largest projects each year, run with an external agency. Those studies produced rich, high-quality insights from real customers. The team wanted to bring that same customer voice to the many smaller decisions made throughout the year.

The team also had internal research tools, which ran on credits, so each question was weighed against what it would cost to ask. Precision mattered even more. Rendon said those tools could only target participants on broad questions, like "Are they a Kendra Scott customer?" or "Are they a jewelry customer?" Matching studies to the team’s customer segmentation strategy would bring every answer closer to the decision in front of them.

Solution: Fast, targeted research built into the design workflow

The team searched for an end-to-end research platform that would help them access customer opinion quickly at scale, with clarity and accuracy. They ultimately chose Askable because the panel let them recruit participants from their existing customer segments, the AI took on the moderating and synthesis that used to take weeks, and the pricing compared favorably with the other platforms they considered.

Rendon said setting up a study took very little effort. Study results came back already summarized. That spared him the week he used to spend synthesizing them and writing them up.

"When we ran our first study with Askable, the initial reaction from the team was 'This is too easy. We're missing something, right?'" Power said.

So they went looking for the catch.

"When we looked into it, we realized that we were getting the amount of data we had in the past, but we were able to get to those same conclusions much faster," Power said.

They could also target participants by demographics, or by the specific features they used. They could trust that the answers came from customers like their own.

Impact: From a few projects a year to continuous customer learning

Research has gone from about two large projects a year to about five studies a month, a 30x increase in research frequency.

Askable's AI-moderated studies reduced turnaround times from weeks to minutes or days, giving the team a lightweight way to test ideas while designs were still in progress.

Faster cycles also mean faster course correction, Power said.

Rendon called the method lightweight but rigorous. The team can leave studies running in the background while they design, bringing customer input to more questions than ever.

For Buy Now Pay Later, the team tested designs with customers who use those payment options. That took them from 20-30 versions down to two or three.

"We were really able to narrow down those design options with confidence," Rendon said.

The team has also been building out their persona libraries. They run studies on how different customers navigate the site, then use Ask AI to ask questions of the results.

Power welcomed how well that mapped to their existing segmentation. It meant they could research how to best personalize experiences for each group.

"For new customers, we were finding that they wanted to see our classic and most popular pieces. But for returning customers, they were looking for what just released, what is new," Rendon said.

Promotional pop-ups were a recurring design question. The team wanted to know not whether customers disliked offers, but when those offers felt usefu and when they felt intrusive.

The answer came down to control. Customers did not object to the offers themselves; they objected to being interrupted and made to dismiss something before continuing to browse. Instead of a modal, the team created a collapsible element that surfaced the offer in context without blocking the page.

"We wouldn't have been able to have the confidence to design in that direction without hearing the voice of the customer," Rendon said.

Daniel Kim, Senior UX/UI Designer, said the study is what changed the outcome.

"There was a real possibility we would have shipped another standard pop-up because it's a familiar pattern and performs well in isolation," he said.

The team used the same approch to test whether customers wanted a chatbot. The answer was yes, giving them the confidence to design one around what customers said they needed.

"Now we're able to include customer research on a lot more of our projects, which gives us higher confidence that we're going to have product fit when it gets to market," Power said.

Looking Ahead: Expanding research beyond the e-commerce and digital team

Power describes where his team has landed as a tipping point for the company. They’ve built enough confidence with the tool to start a community of practice around it. He already asks where else it could go. Could physical product design teams use it? Could it inform market-level decisions?

"We're realizing that because of the speed and simplicity of the tooling, we can expose it to a lot more internal teams than we have in the past, which helps us make better decisions across more areas of the business," Power said.