Cox Automotive cut experiment analysis from weeks to minutes

When Sabrina Ho joined Cox Automotive, she didn't imagine she'd stay as long as she has. But as the company's product analytics matured, so did her mission: to elevate analytics and experimentation to a core part of business decisions.

From neuroscience PhD to startup analytics, Sabrina now sits as Sr. Director of Product Analytics at the global automotive services and technology provider.

It started with Sabrina and her team implementing Optimizely Experimentation at Kelley Blue Book back in 2018. From there, her responsibility gradually expanded — first to 3 brands, then 7 brands, and this year to over 10 brands, including Autotrader and still Kelley Blue Book.

Even with strong experiments and rising conversions, complex analysis was creating bottlenecks.

“Due to the complexity and interplay between our products, experiment analysis historically has been arduous because we have to consider potential downstream impact of dozens of key high-impact signals for intelligent trade-off discussions," explains Sabrina.

Supporting multiple brands requires continuous improvement, and Sabrina's team recognized the need to enhance their experimentation and analytics practices to gain faster insights and maximize the value of their products. To go from reactive to proactive.

By 2024, they had made significant progress. Cox Automotive's experimentation program was running strong — deeply embedded into multiple product workflows, with engineering teams using feature flags and analysts partnering with product managers across development cycles.

But Sabrina's team was also proactively looking for additional value from Optimizely. They knew Optimizely remained the industry leader in experimentation, but without more distinguishing value propositions, they anticipated other parts of the organization might push for different tools in the fast-changing technology landscape.

Then came the announcement: Optimizely had acquired NetSpring, a warehouse-native analytics platform.

"I couldn't raise my hand fast enough," she says. "I was looking for a speedier solution for statistical testing the huge number of value events needed for trade-off decisions in our business. The opportunity to work with the team through co-innovation was too good to pass up."

Just three months later, Sabrina and her team became the first co-innovation customer to go live with Optimizely Analytics, the newly integrated, warehouse-native platform.

While navigating internal processes required close coordination, the team was able to move quickly thanks to a solid technical foundation, along with the solution's advantages:

• A single source of truth in their centralized data warehouse
• No need to recreate data across multiple systems
• Streamlined data governance and lineage efforts

The team no longer had data spread across multiple disconnected systems. Optimizely Analytics allowed them to build on their strong foundations, as well as launch new reports and projects.

While the technology capabilities were impressive, what really won over leadership wasn't the Analytics tool itself — it was the partnership established between the Optimizely and Cox teams.

"My biggest win has been getting buy-in and rallying cross-functional teams to work on things together," Sabrina reflects. Cox Automotive's leadership team see that Optimizely isn't just an experimentation platform vendor; they're thought partners helping solve operational and cultural challenges beyond just technology.

Sabrina admitted she didn't realize the huge potential opportunities Optimizely's warehouse-native solution would mean. What started as a focus on speeding up statistical testing revealed much broader advantages.

With the analytics template built and initial validation complete on consumer brands, they're preparing to turn their focus to the dealership side of the business, where fragmented systems and legacy tools have historically made experimentation difficult.

"We're excited to explore data across some of our more neglected products," says Sabrina. "Optimizely Analytics gives us the visibility we've been missing, along with exploratory analysis capabilities for our B2B products using the GUI rather than SQL."

The team is already considering what AI and Opal can do to help them move even faster. For now, it's about scaling their proven approach with Optimizely Analytics and Experimentation across all products to tackle remaining cross-system challenges.