Logo
Statsig’s unified experimentation platform empowers Bed Bath & Beyond, Inc. to standardize testing across brands, accelerating innovation, improving data trust, and retiring legacy tools.

How Bed Bath & Beyond, Inc. unified experimentation across brands

How Bed Bath & Beyond, Inc. unified experimentation across brands

1 week

time to first experiment

3 developers effort saved

by sunsetting an internal testing app

Bed Bath & Beyond, Inc. is the parent company behind some of the most recognizable retail brands in the U.S., including Bed Bath & Beyond, Overstock, and buybuy BABY. With multiple sites serving millions of customers, Bed Bath & Beyond Inc.’s digital product teams are accountable for building seamless online experiences that directly impact conversion, funnel progression, and revenue.

Experimentation is a cornerstone of their product culture. Having originally launched as Overstock.com in 1999, the company has more than two decades of operating history with a deeply ingrained experimentation culture. That ethos continues today across Bed Bath & Beyond, Inc.’s expanding portfolio of brands.

Life before Statsig: a fragmented experimentation culture

Before adopting Statsig, Bed Bath & Beyond, Inc. relied heavily on an internal experimentation platform built and maintained by their development and analytics teams. This in-house system powered thousands of experiments on Overstock.com when it was hosted on their internal infrastructure and became central to product decision-making.

But as Bed Bath & Beyond, Inc. expanded and launched new sites, including a reimagined Overstock.com on a third-party e-commerce platform, cracks began to show. The internal platform only worked for sites hosted on Bed Bath & Beyond, Inc.’s internal tech stack, leaving teams on the third-party platform unable to run tests. Supporting multiple systems fragmented their experimentation culture.

Even for the sites where the internal tool was available, running tests often came with challenges. Engineering resources had been reduced, leaving the application without a dedicated team to support or debug it. Traffic splits were frequently inaccurate—sometimes veering 60/40 on tests designed to be 50/50—creating a black-box experience for product managers and analysts. The result: some brands could experiment with difficulty, while others were stuck without the ability to test at all.

The path to Statsig: statistical power, flexibility, and fit

When it came time to evaluate vendors, Bed Bath & Beyond, Inc.’s product and analytics teams ran a side-by-side comparison of three platforms. Statsig quickly rose to the top as the only solution that not only matched the rigor of their internal system but surpassed it with more advanced capabilities and scalability.

The decision came down to a few critical factors:

  • Warehouse-native control: Direct BigQuery integration gave Bed Bath & Beyond, Inc.’s data scientists full control over segmentation and KPI analysis, without compromising flexibility.

  • Advanced statistical methods: Statsig’s engine supported sequential testing and confidence interval corrections with multiple KPIs, capabilities that went beyond what their internal platform could provide.

  • Server-side experimentation: Bed Bath & Beyond, Inc. often tests algorithms and backend changes. Statsig’s mature SDKs and proven server-side support made it a natural fit for those workflows.

  • Third-party e-commerce platform support: At the time, Statsig had more direct experience supporting modern e-commerce storefronts, enabling Bed Bath & Beyond, Inc. to get started quickly.

  • Pricing and accessibility: Transparent, flexible pricing ensured teams across Bed Bath & Beyond, Inc. could use the platform without bottlenecks or restrictive seat costs.

“Statsig didn’t just meet the bar set by our internal platform—it raised it. The statistical engine was more advanced and rigorous than anything we had seen elsewhere, giving our data scientists the confidence that tests would run faster and results would be more reliable.”

Nate Dee

Senior Director of Product, Beyond

Fast time-to-value: running experiments in the first week

Statsig’s first proving ground at Bed Bath & Beyond, Inc. was Overstock.com on their third-party e-commerce platform. Product teams drove discovery, analytics defined KPIs and set up experiments in Statsig, and development integrated the SDKs. Within one week of signing the contract, they were up and running their first experiment.

“Within a week of signing the contract, I was able to connect Statsig to BigQuery, define KPIs, and even set up some initial tests on my own. The fact that I could do all of that so easily, without extra overhead, gave us confidence that running experiments in Statsig would be straightforward and powerful.”

Nate Dee

Senior Director of Product, Beyond

Though the team only had a month of active testing before making a business decision to replatform Overstock off the third-party solution to internally host the site, the early outcomes were clear. In that short period, Bed Bath & Beyond, Inc. ran 5–6 experiments that delivered meaningful improvements in user engagement and conversion.

These experiments were not radical redesigns—they were targeted, customer-experience optimizations with strong hypotheses based on prior learnings:

  • Category navigation: Reintroducing category navigation at the top of product listing pages improved exploration and engagement.

  • Sticky checkout button: Making the checkout button persist while customers scrolled the cart page reduced friction and increased completion rates.

“Within just one month, we were able to run a handful of experiments on site experience—and those alone drove measurable improvements.”

Nate Dee

Senior Director of Product, Beyond

Scaling experimentation with Statsig

For Bed Bath & Beyond, Inc., Statsig has proven to be a unified platform for experimentation that is both powerful and easy to use. Teams can connect directly to their warehouse, define KPIs with precision, and run nuanced tests without unnecessary overhead.

One of the biggest benefits has been the shift toward self-service testing. With metrics already defined, data connections in place, and backend integrations set up, product teams can go directly into the Statsig UI to launch experiments themselves. Instead of funneling every request through a single data scientist, experimentation is now accessible across teams—removing bottlenecks and accelerating the speed and efficiency with which Bed Bath & Beyond, Inc. can test and learn.

What’s next: unified experimentation across brands

Despite only running tests briefly on Overstock, Bed Bath & Beyond, Inc. has already made a major decision: they are sunsetting their internal testing software and migrating entirely to Statsig across all brands. This decision saved the company the cost of rehiring three new full-time developers to support the infrastructure of the internal testing application.

A couple weeks ago, the Bed Bath & Beyond, Inc. site launched its first Statsig experiment: a proof-of-concept test of a third-party AI-powered shopping assistant, designed to surface review-based insights for customers directly on product pages. Additional high-priority tests—such as redesigning the featured product landing page for paid traffic—are queued up next.

With Overstock and Bed Bath & Beyond live on Statsig and buybuy BABY next in line, Bed Bath & Beyond, Inc. is on track to unify experimentation across its portfolio and enable faster, more confident product decisions.

About Bed Bath & Beyond, Inc.

Bed Bath & Beyond, Inc. is the parent company of well-known retail brands including Bed Bath & Beyond, Overstock, and buybuy BABY. Together, these brands serve millions of customers across home goods, furniture, baby, and lifestyle categories. With a broad portfolio and a strong presence in e-commerce, Bed Bath & Beyond, Inc. continues to shape how consumers shop for everyday needs and big life moments alike.

We use cookies to ensure you get the best experience on our website.
Privacy Policy