Back to all SEO reportsImage performance
Audit images before they become scenery
More images do not make a site more credible. This audit separates file duplicates from design repetition and retains only visuals that explain content or product state.
Review date: 28 August 2026Scope
Audit scope
- SHA-256 comparison of every public raster asset
- Conservative perceptual difference hash and runtime references
- Format, dimensions, alt text, lazy loading and mobile crops
Evidence
What the audit found
The retained image audit records no byte-identical duplicates.
The audit records file formats, dimensions and source-code references.
Ten guides previously shared only three image sets. Each topic now has a distinct hero, with supporting visuals built from its own content.
Implementation
What was implemented
- Product captures remain WebP and use contained presentation without content-damaging crops.
- Images below the fold load lazily; explicit dimensions or aspect ratios reserve stable space.
- Brand marks may repeat. Project and guide imagery is not multiplied to imply a larger body of work.
Boundary
Evidence boundary
The file audit measures repository and rendering properties. Without real-user data it claims no Core Web Vitals field score or ranking improvement.
The right replacement is sometimes not another photograph but a precise typographic element built from the real content.Research