If you want to know whether changing your Amazon listing images could improve your sales, the normal way to find out is to make the change and wait for actual Amazon performance data.
The problem is that a real experiment on Amazon can take a long time to run its course. You may be waiting weeks or even months before you feel confident about what happened.
There’s another way to get useful feedback before you make that decision.
I use PickFu to put different versions of an image in front of real people and ask which one they would rather buy from. You can create a poll in just a few minutes, get people to explain why they chose one option over another, and then use that feedback to improve the next version.
In the example I’m walking through here, I’m testing a very simple Amazon image change. I’m not rebuilding the entire listing. I’m changing one image, putting the two versions against each other, and asking shoppers which product they would rather buy and why.
That makes the feedback much easier to use.
What I’m Trying to Learn From the Poll
The question I use is very simple:
“If you were shopping on Amazon, which product would you rather buy and why?”
I like wording the question this way because I’m not just asking somebody which image they think looks nicer.
I want them to think about the decision the way an actual shopper would.
That second part — “and why?” — is also really important.
The vote tells you which version people preferred. The written responses tell you what specifically caused them to make that choice.
That can be much more useful than simply seeing that Option A received more votes than Option B.

Image above: PickFu lets you get feedback from real people on product and creative decisions before making the change live.
Step 1: Start a PickFu Mini Poll
For this demonstration, I log into PickFu and use the free daily mini poll available through my PickFu subscription.
There are larger poll options you can pay for as well. Those can make sense if you need more responses or want to get more specific about the audience you’re surveying.
But I really like the mini polls for quick creative questions.
They’re simple enough that you can actually use them regularly rather than turning every image decision into some huge research project.
For this test, all I need to know is which of two Amazon image sets people would rather buy from.
Step 2: Use Your Existing Amazon Listing as the Starting Point
The easiest way to set this up is to create an image-set comparison and enter the ASIN from your Amazon listing.
I grab the ASIN from the Amazon URL and put it into PickFu.
PickFu can then pull in the images that are already on the listing.
That saves a lot of time because I don’t have to manually rebuild the entire original option from scratch.
On the listing I’m testing, the images are displayed as a vertical set, so I set up the poll to match that same general presentation.
I want the original option in the poll to represent what shoppers are currently seeing as closely as possible.
Step 3: Change Only the Image You Actually Want to Test
This part matters.
For this particular poll, I only want to test one image change.
I’m not changing five different things and then asking people which set they like better. If I did that, I could see which version won, but it would be harder to understand which individual change was actually responsible for the preference.
So I take the original image set and replace one image with the new version I want to test.
In this example, one version is the original listing and the alternative version contains an image where I added the product box.
Everything else stays as close to the original as possible.
That gives me a much cleaner comparison.
If shoppers prefer the version with the box, I have a much better idea of what they are reacting to.

Image above: This is the PickFu mini-poll builder where I enter the shopper question and set up the two image options I want people to compare.
Step 4: Upload the Replacement Image
Once the original image set is loaded, I upload the new image I want to compare against it.
In the video, the replacement graphic is one I had created in Canva, so I simply download it and upload it into the PickFu poll.
At this point I have my two alternatives:
- The original Amazon image set
- The same basic image set with the one image change I want feedback on
That’s really all I need for this test.
You do not need to overcomplicate this.
A useful poll can be built around one very specific decision.
Step 5: Preview Exactly What Respondents Will See
Before starting the poll, I click Preview and Start.
PickFu shows me the poll from the respondent’s perspective so I can check the question, the options, and how the images are going to appear.
In my recording, the images briefly weren’t loading correctly in my browser during the preview.
That doesn’t change the process. The important part is that the preview gives you a chance to make sure the poll is set up correctly before you publish it.
You don’t want to collect feedback and then realize afterward that you uploaded the wrong image or that the respondent was seeing something different from what you intended.

Image above: This is the respondent preview. Before publishing, I can see the question and image options essentially the way the person taking the poll will see them.
Once everything looks right, I start the free mini poll.
Quick PickFu Image-Testing Workflow
| Step | What You Do | Why It Matters |
|---|---|---|
| 1 | Start a PickFu mini poll | Gives you a quick way to get outside feedback on a specific creative decision. |
| 2 | Ask which product someone would rather buy and why | Frames the choice around purchase preference instead of simply asking which image looks better. |
| 3 | Enter the ASIN and pull in the existing listing images | Makes your current Amazon listing the baseline for the test. |
| 4 | Change only the image you want to test | Helps isolate the specific change you’re trying to evaluate. |
| 5 | Upload the replacement image and preview the poll | Lets you confirm the alternative creative and verify what respondents are about to see. |
| 6 | Start the poll and review the feedback | Shows both which option people preferred and why they preferred it. |
How Long Does a PickFu Poll Take?
This is one of my favorite parts of the process.
When I run these mini polls, I will often get the results back extremely quickly.
I said during the setup that I often see results in around five minutes.
In this particular example, I went back to doing a couple of other things before leaving the office, checked the poll again, and it was already finished.
The entire process took about 15 minutes.
That does not mean every PickFu poll will always finish in five or 15 minutes.
If you run a much larger poll or get very specific about the audience you want to reach, it can take longer.
But even then, this is the kind of thing where you can start the poll, go get lunch or work on something else, and come back later to see what people said.
That is very different from making a change on Amazon and then waiting a long time to gather enough real sales data to know whether you made the right decision.
The Winning Version Got More Votes, But That’s Not the Most Useful Part
When the results came back, the version where I added the box received quite a few more votes.
One person still preferred the original version.
I actually like seeing that.
Real customers are not all going to agree with each other, and the goal isn’t necessarily to get some artificially unanimous result.
What I care about is understanding the pattern in the feedback.
And this is where PickFu becomes much more useful than a simple A-versus-B vote.
I don’t have to look at the result and say, “Okay, Option B won. I guess I’ll use Option B.”
I can actually look at what people wrote and figure out why they liked the different options.
Why did they prefer the new version? What did they notice? Why did the person who voted for the original prefer that version? Is there something about the winning image that could still be improved?
Those written explanations can give you ideas for the next version that you would never get from the vote count alone.
What to Look at in Your PickFu Results
| Part of the Result | What It Tells You | What to Do With It |
|---|---|---|
| Vote totals | Which option more respondents preferred | Use it as one signal rather than treating the winner as the entire result. |
| Written explanations | Why people chose one version over another | Look for repeated reasons, confusion, benefits people noticed, and objections. |
| Responses favoring the losing option | Why some people still preferred the alternative | Identify tradeoffs or things from the original version that may still be worth preserving. |
| PickFu AI analysis | Themes and recommendations based on respondent feedback | Use the analysis to help identify potential improvements for the next version. |
| Your next creative version | How you applied what you learned | Test it again if you want additional feedback before making the Amazon change. |
Use the Written Feedback to Create a Better Version
The poll result doesn’t have to be the end of the process.
Once I see that the new version performed better and understand why people chose it, I can use those comments to create an even better version.
That means the workflow can look something like this:
- Start with your current Amazon images.
- Create one new variation.
- Ask real people which version they would rather buy from.
- Read why they made that choice.
- Use those comments to improve the image again.
- Create another version if the feedback gives you a good reason to do it.
- Test again if necessary.
Now you’re not simply making creative decisions based on what you personally like.
You have outside feedback guiding the next iteration.
If you’re still deciding what kinds of creative variations to build, you can also look through our Amazon listing image design examples. That shows examples of main images, comparison graphics, infographics, lifestyle images, and other types of creative that can be tested.
PickFu Can Also Analyze the Feedback for You
Another useful part of PickFu is that its AI can analyze the respondent feedback and provide recommendations based on what people said.
So instead of manually trying to find every common theme across the comments, you can also use that analysis to help identify what respondents are reacting to.
That gives you another way to turn the responses into specific ideas for your next creative version.
You can also bring AI tools into the workflow to create additional variants.
For example, you can connect AI tools with PickFu and use them to help create more variations based on the feedback you’re receiving.
That can make the entire testing cycle much faster:
Create → test → read feedback → create another version → test again.
The important part isn’t using AI just for the sake of using AI.
The useful part is that you can combine fast creative iteration with actual feedback from people instead of creating ten different versions in a vacuum and guessing which one is best.
PickFu Feedback Is Different From an Actual Amazon Experiment
There is an important distinction here.
A PickFu poll tells you what the respondents prefer and why.
That is not exactly the same thing as measuring actual conversion performance from shoppers on your Amazon listing.
| PickFu Poll | Amazon Experiment |
|---|---|
| Measures what respondents say they prefer | Measures actual shopper behavior on Amazon |
| Can provide written explanations for why someone chose an option | Gives you performance data based on what shoppers actually do |
| In this example, the results came back in about 15 minutes | Can take much longer to gather enough actual Amazon performance data |
| Useful before making a creative change | Useful for validating performance on the live Amazon listing |
| Can help you improve an image before publishing it | Helps determine what actually happens after the change is live |
After getting the PickFu results, I still have options.
I could take the improved image and go through the process of testing it on Amazon.
Or I could decide that the feedback is strong enough that I’m comfortable publishing the improved version directly.
The point is that I now have a lot more information than I had when I was simply looking at the graphics myself.
I’m not completely guessing anymore.
Testing is also something we use as part of broader Amazon listing optimization and A/B split-testing work. The important thing is understanding which type of test answers which question instead of treating every kind of customer feedback as interchangeable.
Why Your Own Opinion Usually Isn’t Enough
This is especially useful when you are the person who created the graphics.
If you’ve been staring at the same image for hours, you have a very different perspective from somebody seeing it for the first time.
You know why every element is there.
You know what you were trying to communicate.
You remember the previous eight versions.
A customer doesn’t have any of that context.
They see the image and react to it.
That outside perspective is incredibly useful.
There have been plenty of times where something makes complete sense to the person creating the listing because they already understand the product, but the customer sees it differently.
A poll gives you a way to catch some of those differences before you make a bigger decision.
You can see an example of the broader process of improving and testing Amazon creative in our skincare brand Amazon case study, where listing images were improved and then titles and images were split tested.
Why I Like Having PickFu Available All the Time
This is also why I like having a PickFu Plus account rather than treating customer research as something you only do for major projects.
Once you have access to the tool and get accustomed to using it, you start finding more situations where getting real feedback would be useful.
You stop thinking:
“We need to be absolutely certain this decision is important enough to justify research.”
Instead, when you have a meaningful creative decision, you can simply ask people.
That can open up a lot of opportunities to improve what you’re doing on Amazon or on your own website.
You may realize there are many more decisions where customer feedback would be useful than you originally thought.
A Simple PickFu Workflow for Amazon Image Testing
If you want to try this yourself, here’s the basic process I demonstrated:
- Log into PickFu.
- Start a mini poll or another poll type that fits what you need.
- Ask: “If you were shopping on Amazon, which product would you rather buy and why?”
- Create an image-set comparison.
- Enter the ASIN for the existing Amazon listing.
- Let PickFu pull in the existing listing images.
- Keep one option as the original.
- Replace only the image you want to test in the second option.
- Upload the new graphic.
- Preview the poll from the respondent’s perspective.
- Verify that the question and image sets are correct.
- Start the poll.
- Come back after the responses are complete.
- Compare the vote totals.
- Read the written explanations instead of relying only on the winner.
- Review PickFu’s AI analysis if it’s useful.
- Use the feedback to build a better version.
- Decide whether you want to test that version further on Amazon or publish it with the additional confidence you now have.
The actual setup is incredibly simple.
The value comes from what you do with the responses afterward.
Frequently Asked Questions
Can I test my current Amazon listing directly in PickFu?
In the workflow I demonstrated, I entered the ASIN and used the existing Amazon listing images as the starting point for the poll. I then replaced one image in the alternative version so I could compare the current listing against the proposed change.
Should I change several images at once?
For this test, I deliberately changed only one image. If your goal is to understand whether a specific image change improves shopper preference, keeping the rest of the comparison consistent makes the result easier to interpret. You can always take what you learn and test another change afterward.
How quickly will I get results?
My mini polls often come back very quickly. During this demonstration, the completed results were available after roughly 15 minutes, and I mentioned that I have sometimes seen them come back in around five minutes. Larger polls or polls with more specific audience targeting can take longer.
Should I only look at which option gets the most votes?
No. The vote is useful, but the written explanations are one of the most valuable parts of the poll. Read why people chose each option. Even responses from people who preferred the losing version can help you understand how different customers are interpreting your images.
Can PickFu replace testing on Amazon?
I look at it as a different type of information. PickFu lets you get preference feedback and written explanations very quickly. You can then decide whether you want to run an actual test on Amazon or use the feedback to make the change with more confidence.
Can AI help with the process?
Yes. In the workflow I discussed, PickFu can analyze respondent feedback, and AI tools can also be used to help create additional image variants for another round of testing. The combination can make it much faster to go from customer feedback to a new creative version.
Get Real Feedback Before You Guess
You do not necessarily need to wait months to learn something useful about a proposed Amazon image change.
In this example, I was able to take the current listing, create a version with one different image, ask real people which product they would rather buy, and get detailed feedback back in about 15 minutes.
The winning option was useful.
The explanations behind the votes were even more useful.
And now, instead of deciding what to do based entirely on my own opinion after staring at the graphics for hours, I have feedback from people seeing the options with fresh eyes.
That gives me a much better starting point for the next version.
If you want a walkthrough, the video above covers this step-by-step. And if you’ve got a quick question about your situation, leave a comment on the video and I’ll try to point you in the right direction. If you want to try PickFu yourself, I also have an affiliate signup link in the video description.