# Auto-Applied vs Manual Discount Code A/B Test

Page: https://alidemirbas.com.tr/en/lab/ab-test-library/tests/auto-applied-vs-manual-discount-code-ab-test

## Hypothesis

If the eligible code is applied to the cart automatically, shoppers do not have to find and type it and may move toward purchase more easily. If the extra orders outweigh the added discount cost, that could raise post-discount revenue per visitor (RPV).

## When to run this test

Repeated invalid code attempts, shoppers leaving the cart to look for a code, or support tickets asking why an eligible code was not applied are signs that the discount code slows shoppers down in the cart. When the code applies itself, shoppers with an eligible code should move on to checkout more easily. If the hesitation is about price or shipping rather than the code, look at that first. Make sure before launch that the discount shows up in the cart total promptly and clearly, with its amount and source, and that no second coupon field appears on the same screen.

## Primary metric

**Revenue per visitor (RPV)**: Compare post-discount revenue per assigned visitor and track gross margin separately.

## Other metrics

- **Purchase conversion rate (CR)**: Read purchase rate alongside RPV.
- **Code redemption rate**: Check whether automatic application changes redemption.
- **Average order value (AOV)**: Track changes in order value.
- **Gross margin**: Make sure automatic discounts do not erode margin.

## What to check before and during the test

- **Application**: Apply the same eligible code manually in the control and automatically in the variant.
- **Checkout flow**: Measure checkout start and completion time using the same start and end events.
- **Confirmation**: Show the same discount confirmation information in both arms.
- **Keep constant**: Keep discount amount, eligibility, promotion terms and manual code entry unchanged.
- **Separate test**: Test hiding the coupon field behind a link or changing its message separately.

## What to watch when reading the result

- Higher conversion alone is not a win, because more orders can come with more discount. Let post-discount RPV make the call; if gross margin erodes, do not make automatic application permanent despite a higher RPV.
- Expect the result to hold only for the code, discount amount and promotion terms you tested; validate it again before applying it to another promotion. Hiding the coupon field behind a link or changing its message is a separate test.
- Promotion terms must stay the same for the whole test, so start by getting the visitors needed per arm for conversion rate from the [sample size calculator](https://alidemirbas.com.tr/en/calculators/sample-size-calculator). Divide that by each arm's daily traffic to see whether the test can finish before the promotion ends; RPV is noisier, so treat that run length as a minimum. If the difference stays unclear because time ran out, do not take that as evidence that automatic application fails.
- Code redemption is expected to rise when the code applies itself, so treat it as a diagnostic rather than success, and read it with RPV, conversion and average order value. If checkout starts or completion time get worse, use session recordings to see whether the discount confirmation is confusing or creating doubt.

## Frequently Asked Questions

### How might automatically applying a discount affect conversion?

It can reduce the work of locating and entering a code. It may also discount orders whose buyers would not have entered one. Preserve eligibility, manual entry and the same discount conditions.

### Does higher conversion make automatic discounts a winner?

Not necessarily. Use revenue per visitor after discounts as the primary outcome. Monitor gross margin as a guardrail because more discounted orders can increase conversion without creating sufficient economic value.

### How should an automatically applied discount be displayed?

Show its amount and source promptly in the cart total. Make the remaining payable amount clear. Avoid multiple code fields that compete or suggest contradictory discount states.

### How should increased code usage be interpreted?

Automatic application can increase redemption by design. Treat it as diagnostic evidence, not success. Use consistent events for checkout entry and completion time, and interpret redemption against the predefined revenue outcome.