A/B testing Through Amazon


I’m currently running A/B testing on my main image through Amazon. However, I’ve noticed that the key factor of click-through rate (CTR) hasn’t been taken into consideration.

Am I missing anything?
My main goal is to determine which main photo generates the highest number of clicks!

Thank you


Hi @Petunic, is this through Amazon Brand Registry or Amazon Advertising?

Remember, from an Advertising perspective, CTR is just views of the PDP from an ad–not conversions to sales.

Are you sure that CTR is the right metric for this test?

Clicks (or views) ain’t sales, just sayin’. :wink:

1 Like

Through " Manage Your Experiments: set up a test for listing content"

Correct, but If I’m focusing on testing the main image, my primary concern is solely the number of people who clicked specifically because of the main image.


CTR is exactly what we’ve focused on ever since the MYE Dashboard’s functionality was more-widely rolled out on 17Nov2019 (heralded by the New Headline "Quickly test new marketing strategies with “Manage Your Experiments” " (dead link) of that date, which was withdrawn on the 19th as being an example of a particularly-poor performance by the Editorial Team, both in the construction of grammar & syntax, AND in the direction to use links that are behind Amazon’s internal-only Midway Authentication Portal, AND further-still as being in direct conflict with the then-extant A+ Content rules, requirements, and regulations).

However, the recent extension of MYE functionality to the MYI Dashboard (“Manage Your Inventory” was the original choice of nomenclature for the “Manage Inventory” Dashboard, and the MYI acronym persists down to this very day), heralded in the 26Jun2023 News Headline “Create title and image experiments on the Manage Inventory page” (link, Seller Central), has apparently deposited a fly in the ointment (at least in ours, judging from a comparison of recent data points in comparison to the historical trends I’m used to seeing from well-targeted split-testing), so I’m wary about the dust not having yet settled @ this time from this latest expansion of the program.

Just in time for the first of 2023’s Prime Day Events, dadgummit.

1 Like

Which is unknowable - the CTR metric is only for advertising; the split test can’t account for CTR specific to image - BR experiments simply measures sales over an equivalent period of time - how exactly the listing is displayed and how the customer gets to the listing is a tad nebulous.

I think its safe to use the sample size and sales as the best rubric - which is why the results are presented as a probability and not as an absolute.