# Channel-Matched Testimonial A/B Test

Page: https://alidemirbas.com.tr/en/lab/ab-test-library/tests/channel-matched-testimonial-ab-test

## Hypothesis

If visitors see a customer testimonial that fits the need behind the channel they came from, they may more easily see that the product suits their situation. With a sound matching rule, the primary action completion rate within each channel could end up above the fixed testimonial's, while a poor match could reverse the effect.

## When to run this test

Conversion that varies clearly by source is the first reason to run this test; a high bounce rate for visitors from one channel and channel-level surveys with different expectations also tell you visitors arrive with different needs. When the match is right, visitors who see a matched testimonial move on to the primary action more often within their own channel. Check that every testimonial stays within the use its owner approved, and do not present it as "picked just for you," since that can make visitors feel tracked. If the channel gap comes from a mismatch between the ad promise and the page, fix that first.

## Primary metric

**Primary action completion rate**: Compare the same primary action between arms within each eligible acquisition channel.

## Other metrics

- **Conversion by source**: Review which channels show a gain and which do not.
- **Bounce rate**: Make sure an irrelevant testimonial does not increase exits.
- **Testimonial interaction rate**: Check whether visitors interact more with the testimonial.
- **Page load time**: Make sure personalization does not add delay.

## What to check before and during the test

- **Matching rule**: Set one fixed channel-to-testimonial match before launch.
- **Presentation**: Keep headline, testimonial count, placement and visual presentation unchanged.
- **Fallback**: Show the control testimonial when the source is not recognized.
- **Alternative rules**: Industry similarity and company size are separate matching strategies.
- **Evaluation**: Randomize within channel and review visitor groups with enough data.

## What to watch when reading the result

- A rise in testimonial interaction does not justify shipping if the primary action completion rate does not change within the channel. Also check that the bounce rate does not climb in any channel and that personalization does not add to page load time.
- Look at each channel on its own; a gain in one source is not enough to roll the matching out to every source. Keep visitors whose source was not recognized, and who therefore saw the fallback testimonial, separate, because they get the same experience in both arms.
- When the gap is unclear, first check that randomization worked evenly within each channel. If one arm gets more visitors than planned in a channel (a sample ratio mismatch), a difference in traffic mix can be mistaken for a variant effect. If the split matches the plan, enter each channel's numbers separately into the [A/B test significance calculator](https://alidemirbas.com.tr/en/calculators/ab-test) to see whether the gap is beyond chance.

## Frequently Asked Questions

### How should channel-matched testimonials be A/B tested?

Randomize within the same eligible traffic pool rather than assigning channels as arms. Balance by channel if useful and examine source-specific results without confusing source differences with treatment effects.

### How should channels map to testimonials?

Fix the mapping and fallback beforehand. Do not imply an individually selected testimonial, and stay within the customer’s usage permission. Industry or company-size matching is a different strategy.

### What if only one channel shows a gain?

Use completed main actions as the primary outcome, then assess reliable channel breakdowns. Do not roll the result across all sources automatically. Check mismatches and performance costs.