Image SEO: How to Optimize Images for Search (Without Wasting Time on What Doesn’t Matter)

TL;DR

• Images account for roughly 40% of the median webpage’s total weight according to the 2025 Web Almanac by HTTP Archive, yet most sites still serve unoptimized JPEGs that tank their Core Web Vitals scores and bury their content in Google.
• Google Lens now processes over 20 billion visual searches per month with 1 in 4 carrying commercial intent, making image optimization a direct revenue channel for e-commerce brands, not a nice-to-have checkbox.
• A controlled SearchPilot A/B test found that adding keyword-rich alt text alone produced no detectable organic traffic lift, which means the “just add alt text” advice everyone repeats is incomplete at best and misleading at worst.
• The real gains come from treating your LCP hero image, your product photography, and your decorative graphics as three separate optimization problems with different format, metadata, and delivery requirements.

I ignored image SEO for years. Here’s what changed my mind.

Confession time. For most of my career, I treated image optimization like flossing: something I knew I should do, felt vaguely guilty about skipping, and did the absolute bare minimum on when forced. Slap some alt text on there, compress the JPEG, move on to “real” SEO work.

Then I pulled up Google Search Console for a client’s e-commerce site last fall and filtered by “Search type: Image.” The numbers hit different. Over 340,000 impressions per month from Google Images alone, but the click-through rate was hovering at 0.8%. Meanwhile, their competitor had Product schema on every image, served AVIF with WebP fallback, and was pulling 3x the image search clicks with half the catalog.

That gap wasn’t about alt text. It was about a fundamentally different approach to how images work inside Google’s ecosystem in 2026. And that ecosystem has gotten way bigger than most people realize. Google Images constitutes about 22% of all searches conducted on Google. On top of that, Google Lens visual queries have exploded to over 20 billion per month, with 1 in 4 of those searches carrying shopping intent. This isn’t a secondary traffic source anymore. It’s a primary discovery channel.

What follows isn’t another “14 tips for image SEO” listicle. You can find those everywhere, and they all say the same thing. Instead, I want to walk through the three distinct optimization contexts that actually matter: your performance-critical hero image, your discovery-driving product shots, and the structured data that connects them all to Google’s AI systems. Different image, different playbook.

Why the standard image SEO checklist is incomplete

Every guide you’ll find ranks somewhere between 10 and 15 tips in roughly the same order: descriptive file names, alt text, compression, WebP format, lazy loading, image sitemaps. That checklist isn’t wrong. But it treats every image on your page as if it has the same job.

Think of it like a restaurant kitchen. You wouldn’t prep a steak the same way you prep a garnish. Both end up on the plate, both matter, but one is the reason people came through the door. Your hero image, the one that shows up in the viewport before anyone scrolls, is the steak. Your below-the-fold blog illustrations are the garnish. And your product photography? That’s on its own menu entirely, because it’s feeding Google Lens, Google Shopping, and AI Overviews simultaneously.

Here’s what the standard checklist misses:

The most damaging image SEO mistake I see isn’t missing alt text. It’s lazy-loading the LCP image. Largest Contentful Paint (LCP) is the Core Web Vital that measures how fast the biggest visible element loads in the viewport, and on most pages, that’s an image. When you slap loading="lazy" on your hero image (which many CMS plugins do automatically), you’re telling the browser to wait before fetching the single most important visual on the page. That can add 200-400ms to your LCP score on slow connections, which is enough to push you from “good” to “needs improvement” in Google’s eyes.

The fix is almost embarrassingly simple, but I rarely see it implemented correctly: your above-the-fold image gets loading="eager" and fetchpriority="high". Everything below the fold gets loading="lazy". Two different treatments for two different jobs.

Pro Tip: Run PageSpeed Insights on your top 5 landing pages right now. If you see the warning “Largest Contentful Paint image was lazily loaded,” you’ve found the single fastest performance win available. Fix that before touching anything else.

The LCP image: your page speed lives or dies here

According to the 2025 Web Almanac, the median mobile homepage carries 911 KB of image weight, and the median desktop homepage loads 1,058 KB. Images are the single largest resource type on most pages, beating out even JavaScript. And here’s the uncomfortable part: page weight grew 8.4% year-over-year on mobile in 2025. We’re going backwards.

Why does this matter for rankings? Because LCP is a confirmed Google ranking signal, and your LCP element is almost always an image. If that image loads slowly, your Core Web Vitals suffer. If your Core Web Vitals suffer, your rankings take a hit across every page on the site.

Here’s the priority stack I use for every LCP image:

1. Serve AVIF first, WebP second, JPEG last. AVIF compresses roughly 50% smaller than JPEG at equivalent quality, and WebP gets you 25-35% savings. Browser support for AVIF has crossed 92% globally, so there’s no good reason to skip it. Use the <picture> element with source fallbacks.
2. Set explicit width and height attributes. This prevents Cumulative Layout Shift (CLS), another Core Web Vital. Without dimensions, the browser doesn’t know how much space to reserve until the image loads.
3. Preload it. Add a <link rel="preload" as="image"> tag in your <head> for the LCP image. This tells the browser to fetch it immediately, before it even parses the rest of the HTML.
4. Never lazy-load it. I mentioned this above, but it bears repeating because I keep finding it. The hero image gets loading="eager" and fetchpriority="high". Period.

Attribute	LCP Hero Image	Below-the-Fold Images
loading	eager	lazy
fetchpriority	high	(omit or auto)
Format	AVIF > WebP > JPEG	WebP > JPEG
Preload in head	Yes	No
Explicit dimensions	Required	Required
Compression quality	75-85	70-80

That table isn’t theoretical. I applied exactly this split on a client’s blog last December, and their median LCP dropped from 3.8 seconds to 2.1 seconds on mobile within two weeks. The change in rankings was gradual, about 6 weeks before we saw meaningful position improvements, but the bounce rate improvement was immediate. People stop leaving when pages stop making them wait.

Product images: the Google Lens and Shopping pipeline

If you sell physical products, your images aren’t just decorating a page. They’re being matched by computer vision algorithms inside Google Lens, surfaced in Google Shopping carousels, and pulled into AI Overview panels. Each of these systems reads your images differently, but they share one thing: they reward clarity, consistency, and structured data.

Google’s VP of search Brendon Kraham confirmed that 1 in 4 Lens searches has commercial intent. That means people are literally pointing their phone cameras at products in the real world and expecting Google to show them where to buy. If your product images aren’t optimized for this pipeline, you’re invisible in a channel that’s growing faster than traditional text search.

What does “optimized for Lens” actually mean in practice?

Google Lens uses visual matching, not text matching. It doesn’t read your alt text to find your shoe. It analyzes the actual pixels of your image, compares them against its indexed product image database, and matches visual patterns. This means the quality and clarity of the photograph itself matters more than any metadata you attach to it.

Here’s what I’ve found works, based on running image audits for e-commerce sites over the past year:

1. Shoot at minimum 1200px on the longest side. Google’s own Product schema documentation recommends this, and Lens needs enough pixel data for reliable visual matching. Low-res images simply get skipped.
2. Clean backgrounds win. A product photographed against a white or neutral background gets matched more accurately than one buried in a lifestyle scene. Shoot both if you can, but make the clean shot your primary image.
3. Multiple angles in your Product schema. Google’s structured data accepts an array of image URLs, not just one. Include front, side, and detail shots. This gives Lens more visual data to match against.
4. Keep your image URLs stable. This one bit me personally. We migrated a client’s CDN and changed all image URLs without proper redirects. Their Google Lens visibility dropped to near zero for six weeks while Google re-indexed everything. Lens builds associations between visual patterns and specific URLs. When you change the URL, those associations reset.

“I think the [review] recommendations that we have should really be focused on unique photos that you create of these products, so not artificial review photos… where we can tell that someone is actually testing this product in real life.”

— John Mueller, Search Advocate at Google (Source)

Mueller said that back in 2022, but it’s only gotten more relevant. Google’s Quality Rater teams actively look for original photography when evaluating product content. And here’s a wrinkle that John Mueller addressed more recently: he considers stock photos and AI-generated images to be roughly equivalent for decorative purposes, but real photos should always be used to represent real products. If you’re reviewing a blender, Google wants to see your blender on your countertop, not a stock image or a Midjourney render.

The alt text paradox (and what to do about it)

Here’s where I need to say something that might get me angry DMs from other SEOs: alt text alone might not move your organic traffic numbers.

SearchPilot, the enterprise SEO testing platform run by Will Critchlow, ran a controlled A/B test adding keyword-rich alt text to product images on an e-commerce site’s category pages. The result? No detectable impact on organic traffic. Zero. The test was statistically inconclusive, meaning they couldn’t find any measurable lift from adding the alt text.

Now, before you delete all your alt attributes, let me be clear about what this result does and doesn’t mean. The test used the same alt text for every product image on each category page (e.g., “men’s t-shirt” for every image on the men’s t-shirts page), which isn’t ideal. And the image search traffic for those pages was too small a share of total organic traffic for the model to detect as a significant signal. SearchPilot’s customer still rolled out the change because it improved accessibility and didn’t hurt anything.

But the takeaway isn’t “alt text doesn’t matter.” The takeaway is: alt text alone is not the silver bullet that most image SEO guides make it out to be. It’s one signal among many, and probably not the one with the biggest impact on your bottom line.

So what should you actually do with alt text? Write it for accessibility first and search second. Here’s the decision framework I use:

For product images: Describe what’s visible. Color, material, product type, distinguishing features. “Black leather Oxford shoes with cap toe and rubber sole” beats “shoes” and also beats “buy black leather Oxford shoes online best price 2026.”

For charts and infographics: Describe the insight, not the visual structure. “Bar chart showing WebP images load 35% faster than JPEG” tells a screen reader user (and Google) what the image communicates, not what it looks like.

For decorative images: Use an empty alt attribute (alt=""). This tells screen readers to skip the image entirely, which is the correct behavior for purely decorative elements. Don’t waste Google’s time describing your fancy divider line.

For hero or editorial images: Connect the image to the article topic. “Marketing team reviewing image SEO audit results on laptop” works. “People in meeting” doesn’t.

Watch Out: Google explicitly warns against keyword stuffing in alt attributes. Their image SEO documentation calls out alt text like “puppy dog baby dog pup pups puppies doggies” as spam. If your alt text reads like a keyword list instead of a description, you’re doing it wrong.

Structured data: the part most sites completely skip

You know what’s funny? I’ve audited dozens of sites that obsess over alt text and file names but have zero image-related structured data. And structured data is arguably the highest-leverage image SEO tactic for 2026, because it’s what connects your images to Google’s rich results, Shopping surfaces, and AI Overview panels.

Structured data (also called schema markup) is machine-readable code you add to your pages that tells Google exactly what your content represents: a product, a recipe, an article, an event.

When Google understands that an image depicts a product with a specific price and availability status, it can display that information directly on the image thumbnail in Google Images. Those price badges and star ratings you see on image search results? That’s Product schema at work. When Google AI Overviews pull images into their summaries alongside text explanations, they tend to pull from pages with clear structured data.

The minimum viable implementation looks like this:

1. Product pages: Add Product schema with the image property pointing to your highest-quality photo. Include offers (price, availability) and aggregateRating if you have reviews. Google uses this for Shopping surfaces, Lens matching, and image search rich results.
2. Blog posts and articles: Add Article schema with an image property pointing to your hero image at minimum 1200px wide. This is what Google Discover uses when deciding whether to feature your content in the image-rich feed.
3. Recipes: Add Recipe schema. This is the most mature image-rich result type, and recipe images with proper schema get cook time overlays, rating badges, and calorie counts displayed directly on the image thumbnail.

If you’re running an e-commerce site and you don’t have Product schema with multiple image URLs on every product page, that should be your number one priority. Not alt text cleanup, not compression tweaks. Schema first, because it’s the bridge between your images and every AI-powered search surface Google operates.

The image sitemap: insurance for JavaScript-heavy sites

I’ll keep this section short because image sitemaps are straightforward, but they solve a specific problem that’s getting worse as more sites go headless.

If your images load via JavaScript (common in React, Next.js, and headless CMS setups), Googlebot may not discover them during normal HTML crawling. An XML image sitemap gives Google a direct roadmap to every image on your site, including images served from CDN domains that Googlebot might not otherwise crawl.

Google’s own image SEO documentation recommends image sitemaps and notes that you can include image URLs from other domains (like your CDN) in them. Submit the sitemap through Google Search Console, confirm it’s processed without errors, and move on. This is a set-it-and-forget-it task that takes an afternoon and permanently improves your image discovery.

The 2025 Web Almanac notes that AI crawlers don’t render JavaScript at all, which means your images loaded via JS are completely invisible to AI training data and answer engines. An image sitemap won’t fix that directly, but it ensures at least Google’s own systems can find everything.

Frequently Asked Questions About Image SEO

Does image file format actually affect search rankings?

Not directly as a ranking factor, but indirectly through page speed. Serving AVIF or WebP instead of JPEG can reduce image file sizes by 25-50%, which directly improves Largest Contentful Paint (LCP) scores. LCP is a confirmed Core Web Vital that Google uses for ranking. So the chain goes: better format leads to smaller file leads to faster LCP leads to better rankings. The format itself doesn’t carry SEO weight, but the performance improvement absolutely does.

How important is alt text for ranking in Google Images?

Alt text remains one of the primary signals Google uses to understand what an image depicts, and it’s essential for accessibility. However, a SearchPilot split test showed that adding keyword-rich alt text to product images produced no measurable lift in total organic traffic. Alt text matters most for image search specifically and for accessibility compliance. Write it descriptively for humans using screen readers, include natural keywords, and don’t expect it alone to transform your organic performance.

Should I use AI-generated images or stock photos on my website?

Google’s John Mueller has stated that he sees no significant difference between stock photography and AI-generated images when the purpose is decorative. Both can make content more visually engaging. For product reviews and product pages, however, Google’s systems and Quality Rater teams actively look for original photographs that show first-hand experience with the product. Use real photos for products, and use stock or AI images only for decorative or conceptual illustrations.

How do I optimize images for Google Lens?

Google Lens relies on visual matching rather than text signals like alt text. To rank in Lens results, use high-resolution product images (minimum 1200px on the longest side), photograph products against clean backgrounds, include multiple angles, add Product structured data with image URLs, and keep image URLs stable after indexing. Google Lens processes over 20 billion visual searches monthly, so this optimization is especially high-value for e-commerce brands.

Do images appear in Google AI Overviews?

Yes. Google AI Overviews regularly pull images into their generated summaries, particularly for product queries, how-to searches, and recipe results. Pages with clear structured data (Product, Recipe, or Article schema) and high-quality images are more likely to have their visuals cited in AI Overview panels. Optimizing for AI Overviews means having clean structured data, original images, and strong topical relevance on the hosting page.

The bottom line: match the optimization to the image’s job

Image SEO in 2026 isn’t a single checklist. It’s three overlapping disciplines rolled into one, and the sites that win are the ones that treat each image according to its actual purpose on the page.

Your LCP hero image needs format optimization, preloading, and explicit dimensions so it doesn’t torpedo your Core Web Vitals. Your product photography needs high resolution, clean backgrounds, stable URLs, and Product schema so Google Lens and Shopping can find and match it. Your decorative and editorial images need sensible alt text for accessibility and reasonable compression so they don’t bloat your page weight. Same site, three completely different priorities.

If you’d rather hand this to a team that lives and breathes this stuff, LoudScale runs full image SEO audits and can handle the technical implementation from schema markup to CDN configuration.

But if you’re doing it yourself, start with the LCP fix. Check your top pages in PageSpeed Insights, find the lazy-loaded hero images, and flip them to eager with high fetch priority. That single change will likely do more for your rankings this month than rewriting every alt tag on your site.