What to change when doing A/B testing on Amazon

Part of new product development includes A/B testing or split-testing. This process is when you show two variants of the same product to ascertain which one get better results. This practice is time consuming though as it takes a minimum of 2 weeks to make sure that the results you receive are statistically significant yet the longer you test for, the more accurate results you will receive.

A/B testing requires the tester to switch between 2 variations of the same listing alternating them each day. After a designated time you can then look at the data and determine which variation got the most impressions, conversions or sales. When testing new products on Amazon there are several parts of an ASIN which you might want to consider performing A/B testing on to ensure the product is fully optimised. Below are the different areas of an ASIN you could decide to change to discover which variation provides the highest conversion rate.

  • Titles – As one of the 1st places to entice the consumer, discovering what draws a consumer in is important. Does a descriptive title work better than a concise one? Does a specific word to describe your product draw more people in than another word?
  • Images – Images show off your product where words cannot. However, certain images might work better than others. What about 360° view pics? Do they make more conversions or does a size comparison pic work better? Bear in mind that for fully objective results it is best only to change one of your images during split testing such as your hero image or secondary image, not all of them. Moreover, you might decide to add images instead of altering them which again would be a sensible variation to test.
  • Videos – Similar to images, videos explain your product without using up your character count and keep the viewer interested. By changing the video, you can check whether viewers are more likely to convert with a video focused on product specification or a more lifestyle-centric one.
  • Content – You may decide that what your product description or bullet points say about your product might be worth testing to see what alteration could garner better results. Like your title you might want to play with the keywords you use or find different words to explain the same thing. You could ultimately make the smallest detail difference such as moving around the order of your bullet points to see if this makes much difference.
  • A+ content – As a mixture of both content and images, you might decide to A/B test you’re A+ content. The same sort of questions asked for these areas can also be applied to A+ content instead. Depending on your reasoning you might decide to edit both image and content for example due to new branding, otherwise it might be better to be more specific and change only one of the parts of what makes you’re A+.
  • Pricing – Believe it or not, sometimes a lower price doesn’t always mean the best price. In our experience we have found that consumers on Amazon can’t be so easily psychoanalysed. Testing pricing options will provide you with a deeper analysis over the sort of pricing your product can be sold at. A too low price might suggest your product is cheap and tacky whereas an expensive price might deter consumers away as well.
  • Advertising – You can also test your ASINs by doing some A/B testing on Amazon’s advertising platform. Altering areas on a Sponsored Brand ad such as the title or slogan, and the chosen ASINs shown are valid variations. Depending on the keywords you are targeting, you may decide to play with both the title and the ASINs shown or equally just edit one of them to see which one brings in the highest CTR, conversion or impression.
Don’t know where to start with your split-testing?

In layman’s terms, the point of A/B testing on Amazon is to answer the question, what makes a better ASIN regarding the metric you want to analyse by? A/B testing or split-testing can be a tricky job to get right. One small mistake and all your hard work will be for nothing as the results will be skewed and won’t be reliable. Moreover, the data you do collect can provide you will little insight if you aren’t sure how to read the analytics. Our data team have the know how and experience to crunch the numbers to make sure you get as much out of A/B testing as you can. Contact us today to see how our account management and Amazon advertising teams can help you with new product development today.