Back to work
Product MarketingPartner & Affiliate Programs
I ran a controlled test that raised conversion 21% on a partner landing page
The short version
- Company and role
- Discover Home Loans · Principal Digital Marketing Manager
- The problem
- A partner landing page could turn more visitors into mortgage applicants, but nobody could prove a change would work without a trustworthy test.
- What I did
- Turned the page idea into a specific, testable prediction.
- Designed a test-versus-control comparison with the Analytics team, including what could bias the result.
- Ran the test, reviewed the outcome, and turned it into a business recommendation.
- The result
- 21% more conversions than the unchanged control page, measured within one controlled test — not a before-and-after comparison.
- Who I worked with
- Analytics, partner publishers, and business stakeholders.
21%
More conversions vs. the control page in one controlled test at Discover Home Loans
Overview
Discover's affiliate and comparison-site channels were a meaningful source of mortgage leads. I owned the strategy and day-to-day management of these partnerships in a highly regulated lending environment, and identified an opportunity to improve the conversion performance of a key partner landing-page experience.
The Challenge
A customer-journey opportunity had the potential to improve conversion, but the result needed to be measured in a way that Analytics and business stakeholders could trust. Promotional-rate periods created a risk that normal market variation could be mistaken for test impact.
My Approach
- Framed the customer behavior or journey change as a specific, measurable prediction.
- Worked with Analytics to define test and control requirements and the metrics that would determine success.
- Identified promotional-rate-period bias as a potential confound and incorporated that consideration into the measurement plan.
- Aligned the landing experience and partner traffic treatment so the test compared the intended variables rather than unrelated execution differences.
- Reviewed the result with the relevant stakeholders and translated the finding into a business recommendation.
The Results
- Increased conversion by 21% versus the control page, supported by a trustworthy measurement method rather than a simple before-and-after comparison.
- Gave Analytics and business stakeholders confidence that the improvement was attributable to the landing-page treatment.
Key Skills
Test DesignWorking With Analytics TeamsBias AwarenessLanding-page ImprovementDecision-making
Want the full story?
I'm happy to walk through the details and how they could apply to your team.
Email me