Experiments on Amazon FAQ

Split tests are a fairly new feature and you may have many questions. Here are the answers to the FAQ about experiments on Amazon.

Read on or watch our video!

Why did the validation of my Amazon experiment fail?

Apparently the content you added in the test does not meet Amazon’s requirements.

The content you test follows the same rules as the active listings on Amazon.

The title must be a certain length. Usually it cannot be longer than 200 characters, but some categories have specific additional limits.

The photo as one of the most important elements of the listing must fulfil a number of requirements. It must have a certain size (min. 500 pixels, recommended 1600 pixels), the product must cover a minimum of 85% of the image space, the background must be solid white (RGB 255, 255, 255).

In the same way, A+ content must meet the requirements, images must be of high quality, texts must not violate Amazon’s rules and policies.

If your experiment is rejected Amazon should provide a reason for this. Fix the element that Amazon had objections to and run the experiment again.

Can I edit my experiment?

Yes, you can edit your experiment at different stages and to various degrees.

If the experiment has not yet started, you can edit all of its elements. However, if it has already been accepted, it will be reviewed again after you have made changes.

If the experiment has already started, you can only change the hypothesis or the duration of the experiment.

If the experiment has ended or has been cancelled, you can no longer make any changes. You can then start a new experiment.

Why are the results inconclusive?

It may happen that the results do not show a difference between version A and version B, or the result has low confidence.

This may be for the following reasons:

the difference between version A and version B was too small to have a significant impact on customer behaviour

the traffic was too low to show a difference

the difference between versions A and B do not have a significant impact on customer behaviour or their impact is equally strong

However, such results can also be useful to you. It all depends on the hypothesis you set up at the beginning of the test.

Like our article about experiments on Amazon FAQ?

Follow our blog and channel and stay up to date!