A/B Testing
Samantha avatar
Written by Samantha
Updated over a week ago

Optimize your referral campaign by A/B testing various elements of your widget such as the imagery or the CTA. You can easily do this in Friendbuy by setting up an A/B test.

Setting up an A/B Test

Step 1: Create a widget variant

From your list of widgets, select the one that you want to test. Then click on the Cog button under the Overview tab, and select Duplicate - this creates a variant, a virtual 'carbon copy' of your existing Share Button and Share Widget pair.

Step 2: Decide what to test and edit the variant

Put on your thinking cap and get creative; try new ideas by changing the copy, colors, images, etc. We recommend changing just one element, or a small handful of elements at a time.

You might want to begin your A/B test by isolating the share button or ribbon to see what Call to Action creates the most engagement. Next, you could test the Call to Action in the widget Header and Sub Header. If your program offers a double sided incentive, test the offer in the Shared Content. Here is a quick rundown of โ€˜low hanging fruit.โ€™

TEST SHARING RATE

  • Header โ€“ the primary call to action to the sharer (referrer)

  • Sub Header

  • Background Hero Image

TEST REFERRAL VISITS

  • Preview Title for Facebook - the primary call to action to the friend (referred visitor)

  • Preview Image for Facebook

  • Preview Description for Facebook

  • Subject Line, Copy and Image for Email Message

  • Copy for Tweet

SHARE BUTTON or RIBBON

  • Size

  • Position on the page

  • Color

  • Call to action

Step 3: Assign weights

When you are done editing the variant, choose what percentage of traffic will see the Control and Variant. Click on the Settings dropdown in the menu, then click Update Weights.

You will be promote to enter in percentages for each variants - they must always add up to 100. When you are finished, click on Save.

Step 4: View real-time stats and promote a Champion

You can designate which variant you want to compare all other variants against by promoting it to Champion. THIS IS KEY: In A/B testing, all the reporting metrics among the variants are relative to the champion variant (percent gain or percent loss). Reporting makes a lot more sense when you keep this important fact in mind.

Did this answer your question?