# Bundles From Similar or Mixed Price Tiers A/B Test

Page: https://alidemirbas.com.tr/en/lab/ab-test-library/tests/bundles-from-similar-or-mixed-price-tiers-ab-test

## Hypothesis

If judging the bundle as a whole makes the decision easier, the similar-price bundle may bring more purchases; if shoppers want one bundle to cover different needs, the mixed-price bundle may lead to larger orders. The balance between those two effects decides which composition could raise revenue per visitor more.

## When to run this test

When shoppers pick single items over the bundle, remove items from it, or give return reasons tied to its contents, the way bundles are put together may be costing sales. If composition is what matters, one bundle brings either more purchases or larger orders, and revenue per visitor reflects it. If bundles are being missed because of the price explanation or placement, look at those first. Before launch, fill both bundles with in-stock, good-quality, related products, and if you show savings, calculate them from actual standalone prices.

## Primary metric

**Revenue per visitor (RPV)**: Compare revenue per visitor for the two defined bundle compositions.

## Other metrics

- **Bundle selection mix**: Report which bundle or single-item option each visitor selects.
- **Return rate**: Compare returns over the same follow-up period and record bundle contents and return reasons.
- **Average order value**: Calculate per completed order, using the same revenue definition for discounts and cancellations.

## What to check before and during the test

- **Composition**: Define one bundle made of similarly priced items and one mixed-price bundle.
- **Product set**: Keep quantity and use case as similar as practical; record unavoidable product differences.
- **Presentation**: Keep price explanation, placement and any savings calculation unchanged.
- **Flexibility**: Keep item-removal and single-item purchase rules the same.
- **Audience**: Review predefined buyer groups.

## What to watch when reading the result

- Let revenue per visitor decide; use the bundle selection mix to see whether revenue comes from bundles or single items, and average order value to see the effect per order. Keep the single-item option equally visible in both arms; a layout that pushes shoppers into the full bundle inflates revenue, so do not ship on that kind of gain. If revenue rises along with bundle-related returns, question the result.
- Because the products differ, the result compares these two specific compositions; do not credit price range alone, and note unavoidable product differences in the report. Do not let the bundle's total price change along with its composition, or you cannot tell which one caused the difference. Read predefined buyer groups separately, and compare returns only after the follow-up period has run its full length in both arms.
- A few large orders can move revenue per visitor, so treat it as a noisy metric. The [A/B test significance calculator](https://alidemirbas.com.tr/en/calculators/ab-test) tests rates, not averages, so use it on the share of visitors who buy in each arm and treat the revenue gap with the same caution. If both stay unclear, report the result as inconclusive.

## Frequently Asked Questions

### What should stay fixed when comparing bundle compositions?

Define each composition and its pricing before launch. Preserve the single-item route and record product differences so they can be considered when interpreting the result.

### Why use revenue per visitor instead of conversion?

Composition can change both purchase frequency and order value. Revenue per visitor captures both. Calculate order value consistently for discounts and cancellations, and report bundle versus single-item selections.

### Would a mixed bundle win because of its price range?

Not necessarily. Different products mean the test compares two specific compositions. It does not isolate price range. Document the product differences.

### Which products should not be used to fill a bundle?

Avoid unavailable, poor-quality or unrelated items added just to reach a price. Calculate savings from genuine separate selling prices. Track returns by bundle contents and reason.