Split testing or A/B testing, on Amazon, is an excellent way of optimizing your listings through experiments with product content, images, and pricing. There are four crux elements of the listing that affect conversion rate: Image, Title, Product Features and Description, and Price.
Seller Central A/B Testing
Access the Amazon Seller Central page by clicking the URL: https://sellercentral.amazon.com/
-
Once there you will be asked to log in. Log in using the client's Amazon login. (Can be found on the Credentials Spreadsheet).
-
Once you are logged in, the landing page should look something like this:
Hover over the Brands tab, and select Manage Experiments.
On the experiments page, click Create a New Experiment.
You can run an experiment for A+ Content, Bullet Points, Main Image, Product Description, and Product Title.
Main Image A/B Testing
Search the ASIN for A/B testing. It will display if the ASIN is Eligible for the experiment, then click Select Reference ASIN.
If the ASIN is part of a variation, you will need to select first the whole parentage as the reference ASIN.
Enter a descriptive test name, duration, and desired start date. You may also enter an optional hypothesis around what you wish to learn from the test.
-
Experiment Name
-
Hypothesis - Brief description of the experiment and the desired outcome.
-
Duration - Minimum of 4 weeks to a maximum of 10 weeks.
-
Start Date will be automatically set 1 week after you created the experiment, but you can still set it on a later date.
-
End Date will be automatically set depending on the duration of the experiment.
-
You can tick To significance which automatically ends the experiment when there is enough time data to declare the winner.
-
There’s an option now to Automatically publish the winning version.
Uncheck all other Child ASINs that are not included in the experiment.
-
When you run an experiment on a variation family, you can submit an experiment for one or more of the child ASINs, but you always have to submit for the parent ASIN.
-
Upload the 2nd or B image you are putting up against the current main image.
-
Once all information has been entered, click Schedule Experiment. You will be routed back again to the experiments landing page.
-
Click View Details to check the experiment's current status.
