Which main image wins the click? Test your Amazon eye

Why the main image carries the click

On a search results page a shopper compares a handful of things in about a second: the image, the price, the rating and the first words of the title. Of those, the main image is the largest element and the only one they read without reading. On a phone it is roughly the size of a coin, and it sits next to a dozen competitors that sell almost the same thing.

That is why I start every listing review with the main image, and why I look at it on a phone at thumbnail size, never full screen on a monitor.

The rules come first

Some of what wins attention is not allowed, and a non-compliant main image is not a small risk: Amazon says it may remove a listing from search until a compliant main image is provided.

Category rules people miss

• Footwear: a single shoe, facing left, at a 45-degree angle. A mirrored photo is the most common slip.
• Adult clothing: on a model who is standing, not sitting, kneeling or leaning. No mannequins.
• Packaging: not on the main image, unless it is an important feature of the product. Amazon's example is a gift basket, so a gift set's keepsake box belongs in the shot.
• Variants: each colour gets its own main image, showing only that colour. Three colours in one image reads as a three-pack.
• Accuracy: colour, scale and quantity as the shopper will receive them; sharp, no pixelation, and at least 1,000 px on the longest side so zoom works.

Principles worth testing (not rules)

• Angle: a gentle three-quarter view shows volume and shape; straight-on flattens a product into a label.
• Edge definition: light products on white need a soft natural shadow, or they vanish from the grid.
• What's in the box: for kits and bundles, showing every included piece answers the shopper's first question before they click.
• Geometry: long, thin products use a strip of a square frame. On the diagonal the same product can cover far more pixels, still whole and on white.
• Show the differentiator: if the reason to buy is inside (compartments, a feature), an open view can say it before the title does.
• Lead with contrast: when variants sell equally, the one that reads best against a white page usually makes the stronger parent image.
• Thumbnail legibility: if the variant, size or count matters, it has to be readable at coin size — through the product itself, not added text.

These are patterns, not guarantees. The same change can win in one category and lose in another, which is exactly why they are worth an experiment.

How I review a main image

1. Open the search results for the main keyword on a phone and screenshot the first two rows.
2. Put the listing's thumbnail among them. Does it read as the product, at a glance?
3. Check every rule above. Anything non-compliant gets fixed first, whatever it does for clicks.
4. Write one hypothesis per change: “a three-quarter angle will lift CTR because it shows size”.
5. Test one variable at a time and read CTR and conversion together — then check the result with the A/B test calculator.

If the main image is fine and the listing still underperforms, the problem is usually further down: start with the listing auditor or a free mini-audit.