Most photographers obsess over where their images rank in the main Google results and ignore the surface that was built specifically for their work. Google Images is its own search engine with its own results page, and it sends meaningful traffic to sites that take it seriously. A single well optimized photo can pull visitors for years, long after a blog post has stopped ranking for its keyword.
The catch is that search engines cannot actually see your photo the way you do. They infer what an image shows from the text around it, the words inside the file, and the technical signals the page sends. Image SEO is the practice of feeding those signals deliberately instead of leaving them blank. This guide walks through every signal that matters, from the filename to the image sitemap, with concrete examples you can apply today.
Google Images is one of the largest search surfaces on the web, and for visual topics it is often the first place people look. Someone planning a kitchen renovation, identifying a bird, shopping for a jacket, or researching a travel destination frequently starts in the image tab, then clicks through to the page that hosts the photo they liked. That click is a real visitor who never typed your target keyword into the main search box.
For photographers and visual businesses this is undervalued territory. Text-heavy competitors usually treat images as decoration and leave every signal empty: generic filenames, no alt text, no captions, oversized files. That means the bar to rank is low. A portfolio, product page, or photo essay that handles the fundamentals well can outrank far bigger sites in the image results simply because almost nobody else bothered.
The same page can rank in both web search and image search, and the two often surface for different queries. Web search might find your page for "how to photograph waterfalls," while image search surfaces the same photo for "long exposure waterfall." Optimizing images is not a substitute for page SEO, it is a second, parallel path to the same content.
Google has gotten remarkably good at recognizing objects in pictures, but it still leans heavily on text to decide what an image is about and which queries it should appear for. Five signals do most of the work, and they reinforce each other. When all five point at the same subject, Google is confident. When they conflict or go missing, the image gets ignored.
The first words Google reads about an image. A descriptive name is a small but genuine ranking and relevance signal.
The text alternative in your HTML. It is the single clearest description you give of the image, and it doubles as accessibility text for screen readers.
Visible text directly tied to the image. Readers actually look at captions, so they carry both ranking weight and human attention.
The heading, paragraph, and topic of the page the image sits on. Context tells Google whether a photo of a "crane" is a bird or a construction machine.
Machine-readable signals: ImageObject schema, IPTC fields embedded in the file, and your image sitemap. These confirm and enrich what the visible signals already say.
None of these work in isolation. The goal is alignment: the filename, alt text, caption, and page copy should all describe the same subject in natural language. That consistency is what turns a photo from invisible into rankable.
These are the signals you control completely and can fix in an afternoon. Getting them right is most of the battle.
Camera and phone exports land on disk as IMG_4821.jpg or DSC_0093.jpg, which tells Google nothing. Rename the file to describe what it shows before you upload it. Use lowercase words separated by hyphens, not underscores or spaces, because Google treats hyphens as word separators and underscores as joiners.
Weak: IMG_4821.jpg
Strong: red-fox-snow-banff-national-park.jpg
Keep it honest and specific. Two or five real words beat a stuffed string like fox-wildlife-animal-nature-photo-best-hd-stock.jpg, which reads as spam to both Google and people.
The alt attribute is the most important on-page image signal. Write it to describe the image plainly for someone who cannot see it. That single rule keeps you honest for both accessibility and SEO, because the same plain description is exactly what Google wants.
Weak: alt="photo" or alt="fox image hd"
Strong: alt="Red fox standing in fresh snow at the edge of a pine forest"
Aim for roughly one accurate sentence. Do not begin with "image of" or "photo of," since the element is already known to be an image. Do not keyword-stuff, and do not repeat the exact same alt text across many images on the same page. If an image is purely decorative, such as a background texture or a divider, give it an empty alt="" so screen readers skip it. For a deeper treatment of phrasing, length, and edge cases, see the guide on writing good alt text.
People confuse captions and alt text constantly, so here is the clean distinction. Alt text is invisible on the page and exists for screen readers and crawlers. A caption is visible to everyone and sits next to the image in a <figcaption> inside a <figure> element.
Because captions are among the most-read text on any page, they earn extra weight as a relevance signal and they keep visitors engaged. They should not simply repeat the alt text. Use the caption to add context, a story, a location, or a name: "A red fox pauses on a frozen creek in Banff at first light, late winter." The alt text describes, the caption explains.
An image rarely ranks well on a page that has nothing to say about it. Surround your photos with relevant text: a heading that names the subject, a paragraph that discusses it, and a page topic that matches. A wildlife photo on a page titled "Photographing Foxes in the Rockies" with supporting paragraphs gives Google far more confidence than the same photo dropped onto a bare gallery page with no words at all.
IPTC metadata is text embedded inside the image file itself: title, caption, keywords, creator, copyright, and credit. It travels with the photo no matter where it goes, which is its great advantage over alt text and captions that live only in one page's HTML.
Google reads certain IPTC fields and uses them in the image results. When IPTC Creator and Copyright Notice fields are present, Google can display a "Licensable" badge and surface image credit and license details in the image viewer. That metadata helps establish provenance, gives you attribution when your photo spreads across the web, and signals to Google that the image is professionally managed rather than a random screenshot.
Set IPTC fields once, at the source, using your photo software. In Lightroom Classic you fill the Metadata panel and bake the fields into exported files, or apply a metadata preset across a whole import for copyright and contact details. Tools like Photo Mechanic and ExifTool write the same fields in bulk. If you are filling IPTC keywords at scale, the principles in the dedicated photo keywording guide carry over directly.
Alt text and captions are tied to one page. Embedded IPTC data is tied to the file. When someone downloads, re-hosts, or shares your image, the IPTC title, creator, and copyright go with it. That makes embedded metadata the most durable identity your photo carries, so it is worth setting properly even when you also write good on-page text.
The relevance signals above decide whether your image is a match for a query. The technical signals below decide whether Google trusts the page enough to rank it well and whether visitors stay once they arrive. A slow, bloated image hurts both.
Use the format that fits the content. JPEG remains the safe default for photographs. WebP and AVIF deliver the same visual quality at much smaller file sizes and are widely supported, so they are worth adopting for photo-heavy pages. PNG suits graphics, screenshots, and anything needing transparency, but it is wasteful for photographs. SVG is for logos and icons, not photos.
Do not serve a 6000 pixel wide camera original where a 1600 pixel wide image will display. Resize to the largest size the layout actually uses, then compress. For most web photos, aiming for under roughly 200 KB keeps pages fast without visible quality loss. Serve responsive sizes with the srcset attribute so phones download a small version and large screens get a sharp one.
Always include width and height attributes so the browser reserves space and the page does not jump as images load, which protects your layout stability score. Add loading="lazy" to images below the fold so they only download when the visitor scrolls near them. One important exception: do not lazy load your main above-the-fold image, because that delays the content visitors see first.
| Format | Best for | Notes |
|---|---|---|
| JPEG | Photographs | Universal support, reliable default |
| WebP | Photographs, graphics | Smaller files than JPEG at equal quality |
| AVIF | Photographs | Smallest files, excellent quality |
| PNG | Graphics, transparency | Heavy for photos, avoid for them |
| SVG | Logos, icons | Vector, not for photographs |
Once the basics are solid, two machine-readable layers help Google find and trust your images.
Adding ImageObject schema in JSON-LD tells Google explicit facts about a photo: its URL, the creator, the copyright holder, the license, and a credit line. This is what powers image license features in the results, and it pairs naturally with the IPTC data embedded in the file. If you sell or license images, the license URL and acquire-license URL in your structured data can turn a plain image listing into one with a clear path to purchase. A minimal block looks like this:
{
"@context": "https://schema.org",
"@type": "ImageObject",
"contentUrl": "https://example.com/red-fox-snow.jpg",
"creator": {"@type": "Person", "name": "Your Name"},
"creditText": "Your Name",
"copyrightNotice": "© Your Name",
"license": "https://example.com/license"
}An image sitemap lists the images on your pages so Google can discover photos it might otherwise miss, especially images loaded through scripts or galleries. You can add image entries to your existing XML sitemap using the image namespace, or maintain a dedicated image sitemap and submit it in Search Console. This is particularly valuable for portfolios and galleries where the images are the content and you cannot afford to have them go undiscovered.
Uploading DSC_0093.jpg straight from the card throws away the easiest signal you have. Rename before upload, every time.
Empty alt attributes make images invisible to crawlers and unusable for people who rely on screen readers. Identical alt text repeated across a gallery is nearly as bad. Write a unique, accurate description for each meaningful image.
A multi-megabyte camera original slows the page, hurts your performance scores, and frustrates mobile visitors. Resize and compress before publishing.
Words rendered inside a JPEG or PNG are invisible to search engines and to screen readers. Headlines, prices, and key information set as image pixels cannot be indexed or read aloud. Keep important text as real HTML and let the image carry the visual.
An image hosted on a path disallowed in robots.txt, or served only through a script Google cannot follow, will never appear in image results. Make sure your important images sit on crawlable URLs.
Run every important photo through this short list before it goes live.
You do not need to do all of this at once. Filenames, alt text, and file size give you most of the benefit and take the least effort, so start there. Layer in IPTC metadata, structured data, and a sitemap as your library grows. Every image you handle properly keeps working in the image results long after you publish it.
Keep building on these signals with the rest of the PhotoScanr library:
PhotoScanr analyzes your images and writes descriptive titles, accurate alt text, captions, and keywords that line up with everything image search rewards
Try PhotoScanr FreeFree to use . No sign-up required . Instant results