Here are the results of our first A/B test from our 2015 End of Year fundraising campaign.
Three page flow (our control)
In our control (above) credit card donations are processed (via Stripe) from within our user interface, in a form we build, test and maintain.
One page flow (our challenger)
- This test was most likely to impact on donors who choose to pay by Credit / Debit card. Paypal donors complete their transaction off-site on paypal.com so never see step 3 of the three-page flow.
- We ran this test for our en-US and DE audiences to look for any initial bias by locale. We’re trying to better combine testing and localization, but it adds complexity, and our en-US audience is where we can get the quickest results so it’s a balancing act.
- Switching to the one-page form does not have revenue impact of any statistical significance. We should consider these two flows equal.
- Switching to the one-page form does have a positive impact by simplifying our development effort and codebase by significantly reducing the amount of code we have to test and maintain. This is because the overlay checkout flow is built, tested, localized and maintained by Stripe.
Further test considerations
- This test variation combined many small changes to the user experience. Some of these changes may have had positive or negative impacts on conversion rate, even though the sum of these changes was not significant.
- After switching to the Stripe hosted checkout flow, we can further test and tweak this user experience. One possible test is a 2-page flow: Step 1) Choose Amount > Step 2) Choose Payment type.