Savage X Fenty keeps shoppers engaged with rapid, reliable experiments
Before
Manual data exports slowed decision making
Unreliable targeting undermined confidence in test results
Poor multi-platform support for experiments and feature testing
Feature flag setup and cleanup took hours and diverted developer resources
Lack of environment flexibility limited QA coverage for releases
After
Real-time data exports into Segment and Snowflake for more accurate, in-depth analysis
15% improvement in site performance and stability compared to legacy tools
3 hours saved on average per feature flag during setup and maintenance
1 hour saved on average to clean up permanent flags
Multi-environment testing improved QA for different scenarios

About Savage X Fenty
Savage X Fenty is a bold and inclusive lingerie and fashion brand founded in 2018. The brand stands out in the retail market with its commitment to inclusivity, offering products that embrace all shapes, sizes, and genders.
As part of the Techstyle Fashion Group, Savage X Fenty leverages advanced technology and data analytics to optimize customer experiences and stay at the leading edge of fashion and e-commerce. The brand’s mission is simple: to make every customer feel empowered, confident, and ready to embrace their unique beauty.
The Challenge
As Savage X Fenty and its customer base grew, the company’s product and central analytics team had increasing difficulty maintaining a consistent and reliable experimentation process. The existing platform was not always reliable or easy to use, which caused delays in testing new ideas and prevented teams from acting quickly.
“What keeps me up at night is that my product managers are blocked from being able to activate the data and make critical business decisions or run experiments,” said Alan Chang, Product Management Director, Techstyle Operations Services.
The previous A/B testing platform was brittle, hard to maintain, and stakeholders lacked trust in both its targeting and results. Manual data exports, delayed reporting, and inconsistent tools for different platforms prevented teams from running experiments efficiently.
One of the reasons Savage X Fenty chose LaunchDarkly was because the existing A/B testing platform we had was unreliable in choosing the audiences. LaunchDarkly gave the business teams the confidence that experiments could be run reliably and the data could be trusted.
Alan Chang
Product Management Director, Techstyle Operations Services
Solution
With LaunchDarkly, Savage X Fenty unified its experimentation processes across iOS, Android, web, and backend platforms. Teams now work in one tool and benefit from reliable targeting and streamlined workflows.
“The ability to target experiments and analyze results efficiently has been a game-changer for us,” said Riley Duncan, Data Analyst. “It’s enabled us to run more tests and gain actionable insights much faster.”
Their integration with Segment and Snowflake now allows for real-time data exports, reducing their reliance on manual reporting and helping them achieve deeper and faster analysis.
Benefits and Results
With LaunchDarkly, Savage X Fenty has implemented more efficient experimentation, so product and data teams can improve customer experiences and continue pushing the boundaries of innovation in the fashion industry.
Savage X Fenty has also seen measurable improvements in speed and stability compared to their previous feature flag setup. The team now saves an average of three hours per new feature flag in setup and maintenance time across environments, and an additional hour during cleanup. Testing across multiple environments has also improved their quality assurance process, leading to more stable, performant releases. Overall, the team reports a 15% boost in site performance and stability compared to their legacy systems.
“It’s really great to collect and share A/B testing results, because everyone on the team—stakeholders, product managers, the business team—can all trust the data. The LaunchDarkly reports match what we see in our site analytics reports, so that consistency in the instrumentation has been very helpful in building trust in the data,” said Alan Chang.
A recent experiment showed how impactful this shift has been. The team introduced a progress bar on the shopping bag page for VIP members, showing how much more a customer needed to spend to unlock member credits. These were the results after running the experiment for one week:
- The conversion rate increased by 2%, leading to 312 more orders (+3.1%) in the variant group.
- Average order value rose from $120 to $123, a 2% lift.
- Revenue increased by $65,871 (+5.5%) during the test period.
- Orders using at least one credit increased by 2 percentage points, from 76% to 78%.
The team learned that the progress bar encouraged more users to complete purchases, spend more, and take advantage of the member credit program. These changes contributed to higher revenue and deeper customer engagement, indicators of increased customer lifetime value.
Beyond individual experiments, LaunchDarkly delivered significant, organization-wide benefits for Savage X Fenty. With all teams working from a single, trusted data source, data, product, and engineering teams significantly improved their ability to communicate and deliver new features and functionality quickly.