Ali Demirbaş - Lab

Home & Landing

Highlighting the Money-Back Guarantee A/B Test Example

Making an existing refund guarantee easier to notice may reduce purchase hesitation.

  • Home & landing
  • Net revenue per visitor

Control vs Variant

How hard the same message is pushed is the variable.

Changed
Emphasis
Page
Home & landing
Difference
changed

Hypothesis

Additional orders can also bring more refunds and processing costs. This test changes the guarantee’s visibility, not its terms, and needs enough follow-up time for refunds to occur.

How to run this test

Primary metric

Net revenue per visitor

Compare revenue after returns per assigned visitor over a fixed period; assess transaction costs separately.

Other metrics

  • Purchase conversion rate (CR)Check whether purchase rate changes.
  • Return rateReview returns after an adequate follow-up period and assess their effect on profitability.
  • Return processing costMake sure the operational workload remains manageable.
  • Repeat purchase rateCompare later purchases over the same follow-up period in both arms.

What to check before and during the test

  • EmphasisChoose one way to make the existing guarantee more prominent.
  • TermsKeep coverage, duration, eligibility and claim process unchanged.
  • WordingUse the same accurate terms; a conditional versus unconditional guarantee is a separate test.
  • EconomicsRead net revenue alongside return-processing costs.
  • Customer historyReview first-time and repeat purchasers separately.

Setup mistakes to avoid

  1. 1Do not promise a guarantee that cannot be honored.
  2. 2Do not emphasize the guarantee while making its terms too small to read.
  3. 3Do not change guarantee visibility and duration together.
  4. 4Wait for return outcomes before declaring a revenue winner.
  5. 5Do not present a statutory minimum return right as a special marketing benefit.