Preset Amount Buttons vs Free Entry A/B Test

Users may be struggling to pick an amount when field analytics shows long pauses at the amount step, amounts are later corrected or cancelled, or drop-off clusters at that step. If the buttons make the choice easier, that arm sees faster choices and less drop-off; if free entry works better, users type their own amount without being pulled toward the suggestions. When drop-off comes from fees, currency or payment errors, fix those first. Before launch, confirm that the button version keeps a visible "Other amount" path and that the amount labels are easy to read on mobile.

Total amount per assigned user

Divide the total amount of completed transactions by all eligible assigned users, including those who did not complete a transaction.

Also track

• Transaction completion rateCompare completed transactions among eligible assigned users who start one.
• Amount per completed transactionCompare transaction amounts among completed transactions only; do not confuse this with total amount per user.
• Custom amount entry rateCompare visitors who enter their own amount.
• Amount correction and cancellation rateTrack transactions corrected or cancelled because the amount was wrong.

• InterfaceCompare free entry with predefined amount buttons and an “Other amount” option.
• Keep constantKeep currency, fees, amount limits and transaction flow unchanged.
• RecommendationsPredefine suggested amounts; results apply to this set of suggestions.
• SelectionDo not preselect a suggested amount.
• AudienceReview first-time and repeat users separately.

1. 1A higher transaction completion rate can come with a lower amount per completed transaction, so do not pick a winner from completion alone. Total amount per assigned user is the deciding measure; read the custom amount entry rate and the amount correction and cancellation rate to see whether the buttons push people toward amounts that do not fit them.
2. 2Whichever side wins has only been tried with this suggestion set. Try other amounts in a separate test where the buttons are already in place, so their presence and their values never change together. Count corrections and cancellations over the same follow-up window in both arms, and review first-time and repeat users separately.
3. 3Total amount per user can swing with a few large transactions, so before you decide on a narrow or unclear gap, make sure no single outlier is driving it and that the arms split as planned. Run the transaction completion rates through the A/B test significance calculator to see whether that gap stands out from chance. The calculator does not test averages, so treat the gap in amount per user with the same caution.