The 3x3 Experiment

Suppose you want to test your homepage. Pick two areas: your strapline and your “why shop” box. Then indicate those regions in the HTML source with invisible JavaScript snippets. This takes a few minutes.

Next, load different versions of each region into GWO. We’ll assume you’ve picked three different straplines (the slogan beneath your logo) and three different graphical “why shop” presentations.

Strapline Variations:
“ Merchandise breadth: “The Widget Superstore”
“ Established and trustworthy: “Your Trusted Widget Source Since 1984”
“ Post-sale support: “Helping You Choose, Install and Enjoy Widgets”

”Why Shop” Variations
“ Guarantee: “110% Satisfaction Guarantee — Your Order and Your Shipping Fully Refunded If You’re Not 110% Satisfied”
“ Upgraded shipping: “All Orders Shipped Two-Day Air No Additional Charge”
“ Accessories with order: “Accessory Gadgets FREE With Every Widget Order”

After loading these regions and versions into GWO, you turn on the test. Each visitor will randomly experience one of the nine different homepage combinations (3 x 3). GWO will keep each visitor in that specific treatment for that entire visit and subsequent visits so users won’t experience any inconsistencies while they surf. For visitors without JavaScript, the testing platform serves the current version of the page. GWO tracks which tested treatments produce the highest conversion rates. It provides statistical confidence intervals on its results. It also parcels out the contribution from each tested region.