Travis Johnson has witnessed the full lifecycle of analytics pain. As Director of Product Analytics Capabilities at Cox Automotive, with over eight years at the company and a career spanning consulting and multiple enterprise organizations, he's lived through every migration nightmare — from Adobe to Google Analytics, to homegrown SQL-only solutions, to Amplitude.

Each transition left scars on his team, forcing analysts to manually pull data offline, run t-tests in Excel, and field constant stakeholder requests asking if results were statistically significant yet.

Travis and his team discovered something transformative: a native analytics solution that eliminated the questions, concerns, and manual workflows plaguing every experimentation program he'd worked with.

With hundreds of tests running annually and 40-50 metrics per test, the impact was immediate. Tests that took hours or days now take minutes, but it's not all about time savings. The team's newfound trust in the numbers is just as important.

So, would he recommend Optimizely Analytics? "In a heartbeat. What are you waiting for?"

Cox Automotive's 8-Year Partnership with Optimizely

• Travis Johnson is Director of Product Analytics Capabilities at Cox Automotive, where he's spent over eight years building experimentation and analytics programs
• Cox uses the full Optimizely suite: Web Experimentation, Feature Experimentation, Personalization, and Analytics — with Opal AI on the horizon
• Cox became Optimizely's first Analytics customer and received the Analytics Customer Award at Opticon 2025
• The relationship started over eight years ago and has evolved into what Travis calls "an extension of a partnership" — working daily with the CSM team and getting pulled into early releases and beta projects for co-innovation

The Breaking Point — Years of Analytics Scars

• Cox's analytics team lived through multiple painful migrations over the years — Adobe Analytics, Google Analytics ("not happy"), a homegrown SQL-only solution with no UI, then Amplitude — creating what Travis calls "scars" on the team as they jumped from system to system trying to get answers for stakeholders
• Before Optimizely Analytics, every single test required 4-5 hours of manual work: pulling data offline, running t-tests in Excel, fielding constant "Is this stat sig yet?" requests — multiplied by hundreds of tests per year and 40-50 metrics per test
• The breaking point came when leadership demanded increased velocity, but the team had to raise their hands and say, "We can't. We are at capacity" — they couldn't scale without either skipping analysis or adding significant headcount

What is Warehouse-Native? Okay, We Can't Live Without That

• When Travis first heard the term "warehouse native analytics" at Opticon, most people in his organization would nod along without understanding what it actually meant
• The moment it clicked: Optimizely Analytics sits on top of Snowflake, bringing the analytics system directly to where Cox's data already lives — eliminating the need to send data anywhere
• For a large organization like Cox, implementation took from October to January — not because of technical complexity, but because of internal approvals: seven to eight internal tickets that needed to be prioritized, reviewed, and signed off by senior leaders across different groups
• The irony? Cox was already sending data to other systems with existing NDAs — Optimizely Analytics was actually safer because the data never leaves their environment

From Hours to Minutes with Templates and a Single Source of Truth

• With Optimizely Analytics, anyone on the team can create a dashboard for a new test in two minutes — something that used to take hours of manual work across multiple systems
• Cox built templates for their standard test pages with all metrics, user flows, and statistical significance pre-configured — teams just open the template, select the test from a dropdown, hit run, and everything's there
• If a stakeholder requests an additional metric mid-test, there's no need to go back to engineering for new tagging — because Analytics sits natively on their data warehouse, teams can add any metric they want on the fly
• The game-changer: it's the same data for both standard reporting and experimentation. Before, different systems with different tagging would show a 97.4% conversion in one system and 95% in the test results — creating trust issues with stakeholders and wasting time investigating phantom discrepancies that were just system differences
• For critical flows like Kelley Blue Book's vehicle valuation path (six pages, constantly updated), Cox now runs tests on the exact same data source that powers their standard analytics — eliminating the questions, the trust issues, and the wasted cycles

I could have you log in and you could pull a report and create another dashboard for a new test in a matter of two minutes.

Travis Johnson — Cox Automotive

Time Saved, Trust Restored, and the Power of Saying Yes

• Time savings vary by test complexity, but what used to take 4-40 hours of manual work now takes 1-2 minutes to set up — eliminating repetitive manual effort and ongoing frustration for analysts
• The operational game-changer: Travis can now say "yes" more often. When teams ask "Can we do this?", the answer is simple — plug in a dataset and get them set up quickly
• Because Analytics sits natively on the warehouse, Cox can rapidly scale and expand their testing program the right way — connecting any data source, creating or adjusting datasets, and even helping teams without BI tools get up and running fast
• Testing velocity increased not just from speed, but from trust — teams no longer waste cycles debating data discrepancies or questioning results

It's important to point out that now once we build the dashboard, it's our single source of truth with all of the data. Our analysts are confident, the product team is confident, we're confident. There is a new found level of trust and belief that the result is real.

Travis Johnson — Cox Automotive

"Never Seen a Solution Like This"

• Travis has been in the experimentation industry for 15 years and has never seen a solution like Optimizely Analytics — it solves a fundamental problem that testing teams face: proving their value to senior leadership
• The issue: When experimentation teams report results from a different system than the one leadership uses for OKRs and base metrics, leaders don't fully believe it — they'll say "good job" but won't connect testing to real business impact
• Now that Cox's testing program runs on the same source of truth leadership uses for their OKRs, the value creation applies directly to the metrics that matter — changing the game for how experimentation is perceived and trusted
• Looking ahead: With Opal and agentic workflows on the horizon, Travis believes warehouse-native analytics will be "transformational for the entire industry" — and Cox plans to be early adopters once again

In a heartbeat. What are you waiting for? It removes so many questions and so many concerns... the concerns of data alignment with what they think is the truth. It just removes all of that and it automates everything.

Travis Johnson — Cox Automotive

Rapid Fire with Travis and Career Wisdom

• Travis takes on the quickfire round, with some unexpected answers — plus, some AI tips when asked about best purchases and time-saving hacks
• Career advice: Don't lie or make up answers when you don't know something — being truthful about gaps in knowledge is the most important thing for moving forward, and it requires setting your ego aside

15 Years of Experimentation Problems, Solved

Cox Automotive's journey with Optimizely Analytics isn't just about switching platforms — it's about finally solving problems that have plagued experimentation teams for 15 years. After surviving multiple painful analytics migrations and countless hours of manual Excel work, Cox found a solution that does what others couldn't: provide a true single source of truth that leadership actually trusts.

The impact is measurable. What used to take 4-40 hours of manual work per test now takes 1-2 minutes. Hundreds of tests per year multiplied by 40-50 metrics each — all automated. Templates built once, reused constantly. Metrics added on the fly without engineering tickets. And most importantly, the same data powering both standard reporting and experimentation, eliminating the trust issues and wasted cycles that came from system discrepancies.

But the real transformation goes deeper than time savings. When experimentation runs on the same source of truth that leadership uses for their OKRs, testing teams finally get credit for the value they create. Travis can now say "yes" more often, plug in datasets quickly, and help teams scale their testing programs the right way. The result: increased velocity, restored trust, and a testing culture that's proven its worth to the business.

With Opal and agentic workflows on the horizon, Cox is ready to be early adopters once again. Because when you've spent 15 years searching for a solution like this, you know exactly what's worth the investment.