1. Filenames Google can read
Google can't fully "see" your image — it reads the text around it, and the filename is the
first hint. Describe the subject in 3–6 lowercase words separated by hyphens: schnitzel-with-potato-salad.jpg. No spaces, no umlauts or special characters
(they turn into escape codes in URLs), no IMG_ numbers. everyimg cleans filenames
automatically on every download — "Schneiderin Müller nimmt Maß.JPG" becomes
schneiderin-mueller-nimmt-mass.webp.
2. Alt text for people first
The alt attribute is read aloud by screen readers and displayed when images fail
to load — write it for those users and Google gets exactly what it needs too. One natural
sentence describing what's visible. Skip "image of" (it's implied), skip keyword stuffing
(it reads terribly and Google discounts it), and leave alt empty only for purely decorative
flourishes.
3. The right pixel size
Google Discover requires images at least 1200 px wide — a real traffic source for blogs, and the reason 1200 × 675 became the featured-image standard. Beyond that, size is about speed: an image displayed at 800 px shouldn't ship 4000 px. Match the size to the display spot.
4. Lean files on fast pages
Page speed is where image SEO and technical SEO meet: Core Web Vitals are a ranking factor, and images are usually the heaviest thing on the page. Content images under 150 KB, heroes under 200–500 KB, WebP format — those numbers keep the speed report green.
What doesn't matter
DPI (screens ignore it), EXIF keywords (Google doesn't index them), and quality-100 exports (invisible improvement, triple the bytes). Put the effort into the four points above — they're quick, mechanical and they compound with every image you publish.