How to Split Test Your Amazon Image
Setting up an Amazon store can be overwhelming. Sourcing products, figuring out shipping, learning marketing — it’s a lot. And the risk is that after all that work, your product may not sell.
Fear not: I’m going to teach you an effective method to get the results you need: split testing. Specifically, I’ll show you how to split test your Amazon main image, because that one photo has a profound effect on your product’s clickthrough rate.
What is Split Testing?
Also called A/B testing, split testing is a method to gather data by comparing one version of your listing with another.
A traditional split test directs some of your website traffic to one version of a web page, and another portion of your traffic to a different version of that page. Doing this helps you compare how website visitors in each group behave. You might send visitors to these pages via Google or Facebook Ads, through email segmentation, or using landing page software.
On Amazon, however, you have less control with split testing. Instead of sending traffic to different versions of listings simultaneously, you have to rotate the versions over time. So to split test your Amazon main image, you would update the listing with Option A for a set period of time (for example, one week), then rotate Option B as your main image for the same period of time.
Several fantastic tools on the market will help make split testing easier. I’ll mention some of them as I outline the four steps to successfully split-test your Amazon main image.
Step 1: Decide What to Test
Since we’re talking about testing your Amazon main image, this step is easy! Just make sure you take excellent Amazon product photos and choose the best ones to test. And know that you can split test other elements of your Amazon listing, such as the product title, description, and A+ content.
For a more reliable split test, there should only be one difference between the two photos. Take a look at the two main images below.
The only difference between these photos is the treatment of the product itself. In the first option, the hair removal device is positioned on the left side of the package. In the second, it’s on the right and has been sized slightly bigger. The text, packaging, and brand name are all the same on both.
One slight difference can make or break the presentation of your product…
After a round of split-testing, the second image won by a large margin. Since the images were tested on PickFu, which I’ll talk about more in a moment, we can know why.
Respondents liked the second image because the actual hair removal tool is bigger. They can see it in better detail.
This shows that even one slight difference can make or break the presentation of your product as conveyed through the main image.
Though the results will be clearer if you only test one thing, you can also experiment with photos that are completely different. Let’s take a look at some secondary images that were recently tested.
The first image is an infographic that shows what comes in the set. The second is a lifestyle image that shows the product in use. The lifestyle shot performed better during a split test because customers found it more appealing. This would make a perfect second product image for the seller.
Once you’ve figured out which photos you want to test, it’s time to find the means of split testing, and any tools you want to use.
Step 2: Choose Your Testing Method
There are two types of split testing tools:
- Those that test inside the Amazon marketplace
- Those that allow you to test listings privately outside of Amazon
Both have their place in your marketing strategy.
Live Split Testing
Here’s how a live test within the Amazon marketplace works.
A program like Splitly, Cashcowpro, or Listing Dojo automates the rotation of different photos. It can then help you measure the impact of the change by collecting customer data and insights from both variants.
Otherwise, your test will not collect enough data to tell you much. If you’re just starting out with Amazon selling, traditional split testing is likely not your best route — especially if you’re deciding on a main image for your very first listing.
But if you do meet the sales threshold, the neat thing about live split testing is that it gives you cold, hard data about how your listing is performing, including sales and search ranking.
A downside of running split tests within Amazon is that data is all you get. You won’t hear the why or know which image shoppers would choose if they saw them side by side. That’s why it’s better to keep testing variances to a minimum, so that you’re assured that the one thing you tested was behind whatever effect you see.
Another downside is that live split testing can take days if not weeks to gather data. Splitly, for example, runs each test for at least a week. The length of your test depends on your product’s sales volume, how many variants you test, and how many visitors your listing gets.
Most seriously, if your new image variants perform poorly, you risk losing sales while that image is up and running, which could hurt your search ranking. Tread carefully when you perform live tests.
Split Testing Outside of the Amazon Marketplace
Unlike live split testing, you can run some tests outside of the Amazon marketplace. This helps you to know which main image will work better before you mess around with your live listing.
You need brutal honesty!
For example, you could run a Facebook or Twitter poll on two potential images. Just be aware that your friends aren’t usually the best people to ask. They likely aren’t your target market, and (I hope) they’ll try to spare your feelings and encourage you to think your hunches are probably right.
When it comes to your business, you need brutal honesty! It’s not always the easiest to hear, but it’s better than the sound of crickets when your product is for sale on Amazon.
This is where a polling service like PickFu can help. Using PickFu, you can test your Amazon images with unbiased respondents from around the United States, and most importantly, you can make sure they’re within your target market.
For example, if you’re selling a high-end wrinkle cream, you could use PickFu to ask 50 women over the age of 45 who make at least $60,000/year to tell you which photo they like better. Neat, huh?
The best part about testing privately is the results won’t affect your live listing on Amazon, so there’s less risk. It’s cheaper than buying ads, and it won’t take two weeks to get inconclusive results.
Step 3: Split Test Your Amazon Main Image
Now that you’ve chosen your tool (or tools), it’s time to get testing.
How to Split Test Your Amazon Main Image in a Live Test
Programs like Splitly and Cashcowpro require you to sign up for a monthly payment plan to start your first split test. To avoid the cost, you could manually the images or try a free tool such as Listing Dojo.
Any of these software solutions links directly to your Amazon seller account and rotates your image variants until a winner is determined. Each tool has its own method for declaring the winner.
On Splitly, if a statistically significant winner is never found (because of low sales volume, for example), it will stop the test and recommend you try another split testing route.
How to Split Test Your Amazon Main Image with PickFu
Setting up your first poll is simple. First, input a simple question such as, “Based on the image, which product would you buy?”
Then, upload two to eight photos that you are considering for the main image. As discussed, minimize variances in these photos. Keep any text (such as the product name) the same across all options.
Next, choose your audience. Note how you can select a targeted audience. Below, “Amazon Prime Membership” and “Audiobook Listener” have been selected, meaning that the poll will only go to respondents who match both criteria.
The list of responses keeps going, and it’s worth your while to read both sets of comments, not just those in favor of the winning photo.
One PickFu ran a poll to testimages for a back support band. The product itself was aimed at women, but the user did not select a target audience of females. Instead, they tested a mixed-gender audience.
The photo of the woman with more skin showing won the poll with a score of 64 to the less racy option’s 36 (out of 100).
But if you look at the comments, you see that women actually preferred the losing option.
Since the product is intended for women, it’s worth reading their input. They said things like, “I need to know how it fits. I don’t need to see all that skin.”
For this seller, it would be worth going back and retesting with an audience of women to see which one they chose, and why.
How Does Polling Compare to Traditional Split Testing?
If you want to run an interesting and helpful experiment, try split testing your Amazon main images on PickFu. Take note of the winning and losing images. Then, using the same two images, run a live split test.
Many PickFu users have done this. They report back that the PickFu results reflect the same winner and loser as the traditional split test.
It can definitely be worth it to use both live and outside split testing methods. Once you’re confident that polling is predictive of what happens on Amazon, polling with PickFu can save you time and money for the same results.
Step 4: Continue to Optimize
Whichever type of split testing you decide to do, update your listing to feature the winner.
If you used PickFu, make sure to read all the feedback and digest the graphs so you can make an informed decision about other areas of improvement.
Continue to optimize other parts of your listing, such as the product title, description, secondary content. Then watch your sales go up.
Manny Coats, founder and CEO of Helium 10, has this to say about PickFu:
“PickFu.com is where I go when I want to quickly split test Amazon product images to actual Amazon Prime customers, and know which images customers prefer, before I make my product listing live. I get results within one hour with detailed ‘reasons why’ they voted a certain way. This has the potential to save thousands of dollars in lost sales from choosing the wrong image.”
If you haven’t tried split testing yet, start experimenting with your Amazon main image and see how it moves the needle for your products.