The rubric
What the audit actually checks
Every image is scored across nine categories. Each finding carries a severity, the reason it costs you conversions, and a specific fix — written against what's actually in your image, not a generic checklist.
The image type sets the rules
Main / Hero
The image that wins or loses the click in the search row. Held to Amazon's policy rules, with compliance risk reported separately from the conversion score.
Nine categories
Every image, scored the same way
- 01
Amazon compliance
What a compliance choice costs you in attention — whether a badge adds clutter, whether a prop pulls the eye off the product. Suppression risk is a different question and gets its own section, so it never quietly drags down the score you came for.
- 02
Keyword & text placement
Whether your most important benefits sit where the eye naturally lands first, or are buried in a corner. On hero images — where no added text is allowed — this becomes about whether the product angle itself communicates the value proposition.
- 03
Font legibility at thumbnail size
Search results render your image at roughly 100–160px. Specific text elements that are too small or too thin to survive that shrink get called out individually, not as a general warning.
- 04
Render & image quality
Lighting, shadow grounding, sharpness, compression artefacts, and the specific tells that make a render read as amateur or as untouched AI output.
- 05
Product prominence
How much of the frame the product actually occupies, and whether empty gutters are costing you visual weight against competitors in the same row.
- 06
Contrast & colour
Text-to-background contrast measured against legibility thresholds, plus whether the palette supports the product or competes with it.
- 07
Clutter & hierarchy
Whether there's one clear focal point and a readable eye path, or several elements fighting for the same attention.
- 08
Emotional appeal & value prop
Whether the core benefit registers within the first second of viewing — the window you actually get in a search grid.
- 09
Differentiation
Whether the image looks distinct in its category, or like the same stock template as everyone else on page one.
The honest limit
What it can't tell you
It also won't confuse two different questions. Breaking one of Amazon's main-image rules is a suppression risk, and it gets reported as one — separately, with a note on how common the pattern is. It never touches the score, because a rule most of your category breaks is a business decision, not a design fault.
This audits your image against Amazon's published policies and conversion best practice. It has no access to your listing's click-through or conversion data, so it can identify why an image is likely underperforming — it can't confirm that it is. Treat it as a design review with teeth, not an analytics report.
Run an audit