Virgin Money is a pioneer in social lending. Formerly CircleLending, in 2007 they were purchased by The Virgin Group, extending their services from social lending to credit, mortgages, wholesale loans, and more.
Virgin Money realized early on that website optimization was crucial to their online marketing efforts. They wanted to use Google Website Optimizer (GWO) to increase conversion rates from paid search traffic, but were unsure of where to start.
A guided multivariate test, performed by Virgin's internal team, and guided by VKI Studios! A test of 12 redesign options combined with a complete walkthrough of the testing procedure.
Key lead generation page chosen. This page provided a clear metric (leads) which was also a key function of the website.
Three section multivariate testing. The page was based on a three column design, splitting easily into three testable sections.
All three sections redesigned. Headlines and copy for first section were re-written to explain and promote the service. Sign up form length within the second section was reduced so as to promote sign ups. Images and copy were changed within the third section to relate to the theme of "changing the face of money". Each section was redesigned in as many as three ways, providing a total of 12 possible re-combinations. We predicted that the shortened form length would play the major role in increasing conversions.
Construction and implementation of new CSS and HTML set up on local servers. This illustrated how the actual GWO test would function, as well as highlighting noteworthy aspects of the system, such as the role of browser cookies.
12 variations fed to samples of users and results managed. Virgin's internal dev team was lead through the process of setting up the test, and instructed on how to prune and manage results.
Virgin's internal team was educated on setting up and implementing a test that ran for 42 days, producing a 90% increase in conversions. The test concluded that the best combination was original content with the shortened form, as we had predicted.
