Everyone agrees that images matter on Amazon. Far fewer sellers can say how much their images matter, which ones are underperforming or whether last month's redesign actually helped. Without numbers, image decisions come down to taste, and taste is a poor guide to what shoppers click and buy.
This guide is about measurement. It explains the two jobs images do in the shopping funnel, which reports show each one, how to calculate the metrics that matter and how to read changes without fooling yourself. It does not cover how to run a controlled split test; that belongs to our guide on A/B testing images with Manage Your Experiments. Here we focus on the everyday numbers every seller can watch.
The Two Jobs Your Images Do
Images influence two distinct moments, and they are measured differently:
- Winning the click. In search results and ad placements, the main image is the most prominent visual, shown beside the title, price, rating and delivery promise. Its job is to earn a click. The metric is click-through rate: clicks divided by impressions.
- Winning the sale. On the product page, the full gallery, including secondary images and any video, helps the shopper decide. Its job is to convert a visit into an order. The metric is conversion rate, which Amazon reports as unit session percentage in Business Reports.
Keeping these separate is the most useful habit in image analysis. A new main image should be judged mainly on click-through. A new set of infographics should be judged mainly on conversion. Mixing the two leads to wrong conclusions, such as blaming a main image for a conversion problem caused by a confusing size chart.
Business Reports: Sessions and Unit Session Percentage
Business Reports in Seller Central are available to all professional sellers. The most useful view for image work is the detail page sales and traffic report by child item, which shows, for each ASIN over a chosen date range:
- Sessions: visits to the product page, counted per shopper within a period of time.
- Page views: total views of the page, which can be higher than sessions.
- Units ordered: the number of units sold.
- Unit session percentage: units ordered divided by sessions, Amazon's main conversion measure.
Unit session percentage is the number most directly affected by your gallery. If sessions are healthy but unit session percentage is weak compared with your other products or your own history, the product page is not persuading. Images are one of the first things to check, alongside price, reviews and the buy box. Note that multipacks can push unit session percentage up because one order contains several units, so compare like with like. Amazon has adjusted these reports over time, so field names may differ slightly in your account.
Brand Analytics: Search Query Performance
Brand owners enrolled in Brand Registry can access Brand Analytics, which includes the Search Query Performance dashboard. For each search query that led shoppers to your brand, it shows the funnel from impressions to clicks, add to carts and purchases, both in total for the query and for your brand's share.
This is the closest thing Amazon offers to an organic click-through rate for your products. Divide your brand's clicks by your brand's impressions for a query to get your click-through rate, and compare it with the overall rate for that query. If your share of impressions on an important query is reasonable but your share of clicks is much lower, shoppers are seeing your product and choosing others. The main image, along with price and rating, is a prime suspect.
Brand Analytics also includes views at the ASIN level, such as search catalogue performance, which show a similar funnel per product across all queries. Availability and naming of these dashboards vary by marketplace and change from time to time, so explore what your account shows.
Advertising Data as an Image Signal
Sponsored Products reports show impressions, clicks, click-through rate, orders and conversion rate for each campaign, ad group and search term. Because your main image appears in sponsored placements, advertising data offers a fast read on how well it earns clicks, often with more volume than organic data alone.
Images also affect advertising costs. In an auction system, an ad that earns more clicks for the same impressions is usually more efficient, and a page that converts better turns those clicks into sales at a lower cost per order. Weak images can quietly make every campaign more expensive. Our article on the best Amazon PPC strategies for 2026 covers the advertising side; the point here is that image quality shows up in your advertising cost of sales as well as your organic numbers.
When reading ad data, keep bids, budgets and targeting stable around an image change if you can, or changes in placement will muddy the comparison.
Diagnosing Which Images to Fix
Put the metrics together and a pattern usually emerges. A simple diagnostic grid:
- Low click-through, healthy conversion. Shoppers who reach the page buy, but too few click. Focus on the main image. Our guide to creating a main image that wins the click covers what to change.
- Healthy click-through, low conversion. The listing attracts interest but fails on the page. Look at secondary images, video, the order of the stack and the questions shoppers ask that the gallery does not answer.
- Low on both. Check fundamentals first: price, reviews, stock, buy box and relevance of the keywords you rank for. Images may be part of the problem, but not all of it.
- Healthy on both. Protect what works. Make changes cautiously and one at a time.
Compare against sensible benchmarks: your own best-performing products, the same product's past performance and the query-level averages in Search Query Performance. Avoid comparing against figures you read online, which vary widely by category and price point.
Reading Before and After Changes Honestly
The most common way sellers misjudge an image change is a quick before-and-after look that ignores everything else going on. To read changes fairly:
- Record a baseline. Note click-through, sessions and unit session percentage over a stable period, ideally several weeks.
- Change one thing at a time. If you change the main image, price and title in the same week, you cannot tell which one mattered.
- Allow enough time. Give the change a few weeks to collect data, longer for lower-traffic products.
- Watch for outside factors. Seasonality, deals, stock-outs, new competitors, review changes and advertising shifts all move the numbers.
- Compare equal periods. Same number of days, same days of the week, and avoid periods that include big sales events.
When stakes are high and you have Brand Registry, a controlled experiment is more reliable than a before-and-after comparison because both versions run at the same time.
A Worked Example: Finding the Weak Link in a Lunch Box Listing
A seller of a stainless steel bento lunch box sees sales flatten while advertising costs rise. In Business Reports, sessions are stable but unit session percentage has been similar to their other products. In Search Query Performance, their brand's share of impressions for the main query "metal lunch box for kids" is solid, but their share of clicks is well below their share of impressions. Sponsored Products data for the same term shows a click-through rate lower than on their other ads.
The diagnosis points to the click, not the page. Looking at search results on a phone, their main image shows the closed box from directly above, and it looks like a plain silver rectangle. Competing listings show the box open with compartments visible. The seller reshoots the main image with the lid propped open at an angle so the compartments are clear, keeping a pure white background with no added text and the product filling most of the frame. They leave price, title and bids unchanged. Over the following weeks, their click share on the main query moves closer to their impression share, and ad click-through improves. Unit session percentage holds steady, which confirms the page itself was not the problem.
Building a Simple Image Performance Tracker
A spreadsheet is enough. For each important ASIN, record each week or month:
- Sessions and unit session percentage from Business Reports.
- Impressions, clicks and purchases for your top few queries from Search Query Performance, if available.
- Advertising click-through and conversion for your main campaigns.
- A notes column logging every image change, price change, promotion or stock issue with its date.
The notes column is the most valuable part. Months later, it lets you see what actually happened after each change, and it becomes the evidence base for deciding when images need attention again. Our guide on when to refresh listing images explains how to use those signals.
Image Metrics Checklist
- Separate click-through (main image) from conversion (gallery) in every analysis.
- Check sessions and unit session percentage in Business Reports by child ASIN.
- Use Search Query Performance to compare your click share with impression share.
- Use Sponsored Products click-through and conversion as a fast secondary signal.
- Diagnose with the four-way grid before deciding what to change.
- Record a baseline before any image change.
- Change one element at a time and hold other variables steady.
- Allow several weeks and compare equal periods.
- Log every change in a tracker with dates.
- Use controlled experiments for high-stakes changes where available.
Images are one of the few levers you control fully, and measuring them turns design into a commercial decision. Desnora designs listing images with these metrics in mind through our Amazon listing image design service, and works alongside our Amazon advertising team so creative and campaigns support each other. To review your numbers with us, contact us.
Frequently Asked Questions
Which Amazon image affects click-through rate most?
The main image, because it is the visual shown in search results and most ad placements. Secondary images mainly affect conversion once shoppers reach the product page.
Where can I see my conversion rate on Amazon?
In Business Reports, the detail page sales and traffic report by child item shows unit session percentage, which is units ordered divided by sessions.
How do I find my organic click-through rate?
If you have Brand Registry, Search Query Performance in Brand Analytics shows impressions and clicks per search query for your brand, so you can calculate click-through for each query.
What is unit session percentage?
It is Amazon's conversion measure: the number of units ordered divided by the number of sessions on the product page over the same period.
Can I use PPC data to judge my main image?
Yes. Sponsored Products reports show click-through rate for your ads, which carry your main image. Keep bids and targeting stable around an image change for a fair read.
Do better images lower advertising costs?
They can. Ads that earn more clicks and pages that convert better generally produce sales more efficiently, though price, reviews and competition also matter.
How long should I wait after changing images?
Allow several weeks, longer for low-traffic products, and compare equal periods that avoid major sales events and stock problems.
What is a good conversion rate for Amazon listings?
It varies widely by category, price and traffic source. Compare against your own best products, your history and query-level averages rather than general figures.
Why are my sessions high but sales low?
Shoppers are reaching the page but not buying. Review your secondary images, video, price, reviews and whether the gallery answers the questions shoppers ask.
Can I see how each individual image performs?
Amazon does not report metrics per secondary image. You infer impact from listing-level changes, or test versions with Manage Your Experiments where available.
Should I change several images at once?
Only if you accept that you will not know which change made the difference. For learning, change one element at a time and log the date.
Do multipacks affect unit session percentage?
Yes. Orders containing several units can raise unit session percentage, so compare multipack listings with each other rather than with single units.