Images for Visual Search Guide

Someone photographs your competitor’s product with Google Lens, and Google serves up a link to buy it. That customer never typed a single keyword. They just pointed their phone at something they liked.

This is visual search, and it’s already massive. Google Lens handles over 12 billion searches a month. Pinterest Lens drives 80% of its users’ purchase decisions. If your product images aren’t optimized for this, you’re invisible to a growing chunk of your potential customers.

The good news: most of the optimization isn’t hard. The bad news: almost nobody does it properly.

What’s Actually Happening Under the Hood

When someone points Google Lens at a handbag, the system does a few things very quickly:

First, it figures out what objects are in the photo—“that’s a handbag, not a briefcase.” Then it extracts features: the color, the texture of the leather, the shape of the buckle, the pattern of the stitching. It compares those features against billions of indexed images. Finally, it checks the surrounding context—what does the page around the image say? What schema markup exists? What’s in the alt text?

That last step is where you have control. The visual recognition part is Google’s problem. The context you give the image? That’s yours.

Where Visual Search Happens

Google Lens

Google Lens is everywhere—baked into Google Search, Google Photos, every Android camera. Point it at a product and it shows shopping results. Point it at a plant and it identifies the species. Point it at foreign text and it translates live.

For e-commerce, Google Lens is the big one. If your product images are well-optimized, Lens can match them to shopping queries and surface your listings.

What Lens cares about: Structured data on your pages, high-quality product images, accurate Google Merchant Center feeds.

Pinterest Lens

Pinterest is where people go to discover products they didn’t know they wanted. Someone photographs a friend’s kitchen backsplash, and Pinterest shows them where to buy similar tiles, paint colors, and cabinet hardware.

It’s particularly strong in fashion, home decor, food, and anything visually driven.

What Pinterest cares about: Pin descriptions, board organization, rich pins with product data.

Bing and Amazon

Bing’s visual search powers image search across Edge and Microsoft products. Amazon’s app lets shoppers photograph products and find them on Amazon. Both matter for e-commerce.

Bing wants: Image sitemaps, schema markup, Bing Webmaster Tools submissions. Amazon wants: Perfect product listing images that follow their specs exactly.

Getting the Technical Basics Right

Image Quality Matters More Than You’d Think

Visual search algorithms need detail to work with. A blurry 400px product photo gives the algorithm almost nothing to match against. A sharp 1200px photo gives it plenty.

What You’re Shooting	Minimum Size	What I’d Actually Use
Product images	800×800 px	1200×1200 px or bigger
Blog/content images	600×400 px	1200×800 px
Infographics	800×1200 px	1500×2000 px

Beyond resolution, the image itself matters:

• Sharp focus on the subject (not the background, not the table it’s sitting on)
• Good lighting—even smartphone photos look decent with proper lighting
• Accurate colors, not overly saturated or filtered
• Clean backgrounds for products (white is standard for a reason)

Format and File Size

WebP gives you the best combination of quality and file size. JPEG works universally. PNG is for graphics with text or transparency.

Target under 200KB for standard images, under 500KB for hero images. Compress at 80-85% quality and most photos look identical to the original.

BulkImagePro handles batch compression while keeping images sharp enough for visual search to work with.

File Names: The Easiest Win Nobody Takes

This one drives me a little crazy because it’s so simple and so few people do it.

Search engines read file names. They extract meaning from them. And yet most product images are uploaded as:

IMG_2847.jpg
DSC_0042.png
product-1.jpeg

That tells Google exactly nothing. Compare that to:

red-leather-crossbody-bag-front-view.jpg
stainless-steel-kitchen-mixer-black.png
mens-blue-oxford-shirt-cotton.jpeg

Now Google knows what’s in the image before it even processes the visual content.

The naming rules that work:

• All lowercase (no case-sensitivity headaches)
• Hyphens between words (not underscores—Google treats hyphens as word separators)
• Include real attributes: color, material, style, product name
• Describe what’s in the image, not where it goes on your site
• Keep it under 50 characters when possible

Alt Text: Still Underrated

Alt text exists for accessibility—screen readers use it to describe images to visually impaired users. But search engines also lean on it heavily for context.

The trick is writing alt text that serves both purposes: descriptive enough for a screen reader, specific enough for a search engine.

The difference between lazy and useful alt text:

Image	Lazy Version	Useful Version
Product shot	”bag"	"Brown leather messenger bag with brass buckles and adjustable strap”
Data graphic	”chart"	"Bar chart comparing 2025 e-commerce sales by category”
Lifestyle photo	”woman shopping"	"Woman browsing vintage dresses at an outdoor flea market”

What good alt text looks like:

• Specific and descriptive (80-125 characters is the sweet spot)
• Includes relevant keywords without forcing them
• Mentions color, material, and style when they matter
• Doesn’t start with “image of” or “photo of” (screen readers already say “image”)
• Doesn’t stuff keywords—Google penalizes this, and it sounds terrible for accessibility

Structured Data: Telling Search Engines Exactly What They’re Looking At

Schema markup gives search engines structured information about your images and the products they show. This is especially important for e-commerce visual search.

Product Schema

If you’re selling something, this is non-negotiable:

{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Classic Leather Briefcase",
  "image": [
    "https://example.com/images/briefcase-front.jpg",
    "https://example.com/images/briefcase-side.jpg",
    "https://example.com/images/briefcase-open.jpg"
  ],
  "description": "Full-grain leather briefcase with laptop compartment",
  "brand": {
    "@type": "Brand",
    "name": "Your Brand"
  },
  "offers": {
    "@type": "Offer",
    "price": "299.00",
    "priceCurrency": "USD"
  }
}

Multiple images in the array gives Google more visual data to match against. Three to five images is the minimum; eight or more is better.

ImageObject Schema

For non-product images (blog graphics, infographics, editorial photography):

{
  "@context": "https://schema.org",
  "@type": "ImageObject",
  "contentUrl": "https://example.com/images/guide-infographic.jpg",
  "name": "Visual Search Optimization Infographic",
  "description": "Step-by-step guide to optimizing images for visual search engines",
  "width": "1200",
  "height": "2400"
}

Test your schema with Google’s Rich Results Test before moving on.

Image Sitemaps: Helping Search Engines Find Everything

Your regular sitemap tells Google about your pages. An image sitemap tells it about your images specifically—including context it can’t get any other way.

<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
        xmlns:image="http://www.google.com/schemas/sitemap-image/1.1">
  <url>
    <loc>https://example.com/products/leather-bag</loc>
    <image:image>
      <image:loc>https://example.com/images/leather-bag-front.jpg</image:loc>
      <image:title>Brown Leather Messenger Bag - Front View</image:title>
      <image:caption>Handcrafted full-grain leather messenger bag with brass hardware</image:caption>
    </image:image>
    <image:image>
      <image:loc>https://example.com/images/leather-bag-detail.jpg</image:loc>
      <image:title>Brown Leather Messenger Bag - Interior Detail</image:title>
    </image:image>
  </url>
</urlset>

Submit through Google Search Console and Bing Webmaster Tools. Most people forget Bing—don’t. It powers a surprising amount of visual search through Edge and Microsoft products.

Product Image Strategy

E-commerce images need extra attention for visual search. For the complete pipeline from photography to upload, see our e-commerce image optimization guide. Our product image SEO guide covers ranking specifically in Google Images.

Shoot More Angles Than You Think You Need

Every angle you provide is another opportunity for visual search to match your product. At minimum:

• Front view (this is your primary—make it count)
• Back view
• Both sides
• Detail shots showing texture, stitching, hardware
• Context shots showing the product in use or to scale
• 360° views or video if your platform supports it

A customer photographing your competitor’s product from the side? If you only have a front view, you won’t show up in that match.

Keep Everything Consistent

Visual search engines learn patterns across your catalog. Consistent product photography—same background, same lighting, same framing—helps them understand your brand as a whole.

White or neutral backgrounds are standard for a reason: they let the algorithm focus entirely on the product. Consistent dimensions mean your catalog looks professional and your theme renders predictably. Use BulkImagePro’s resizer to standardize dimensions across your full product catalog.

Photograph Every Variation

If a shirt comes in blue, red, and green, you need separate photos of all three. Don’t just photograph the blue one and describe the others in text.

Name them accordingly: mens-oxford-shirt-blue.jpg, mens-oxford-shirt-red.jpg. Include the color in alt text and schema markup. Someone photographing a red shirt expects to find red shirts.

The Stuff Around the Image Matters Too

Visual search doesn’t just analyze the image file—it reads the page the image lives on.

Put Images Near Relevant Text

This sounds obvious, but I’ve seen product pages where the image is in one section and the description is three scroll-lengths below it. Search engines use proximity to establish context. A product image right next to its description gives much stronger signals than an image floating alone.

Use Descriptive Headings

The heading above an image tells search engines what to expect:

<h2>Handcrafted Leather Messenger Bags</h2>
<img src="leather-bag-collection.jpg" alt="...">

Much better than a generic <h2>Our Products</h2>.

Add Captions When They’re Natural

Captions provide yet another text signal. Don’t force them where they’d be weird, but for editorial content and product showcases, they add real value:

<figure>
  <img src="image.jpg" alt="...">
  <figcaption>2025 collection featuring Italian full-grain leather</figcaption>
</figure>

Mistakes That Kill Visual Search Performance

Blurry, dark, or tiny images. If the algorithm can’t identify objects in your photo, it’s useless. Minimum 1200px on the longest edge, properly lit, properly focused.

Missing or garbage alt text. alt="" or alt="image" or alt="product123" gives search engines nothing. Every image needs descriptive, specific alt text.

Watermarks and text overlays. Big watermarks confuse object recognition. Promotional text baked into product images interferes with feature extraction. If you need overlay text, do it with CSS so the underlying image stays clean.

Inconsistent product photos. A catalog where every product has different lighting, different backgrounds, and different angles looks unprofessional to humans and confuses visual search algorithms.

Desktop-only optimization. Most visual searches happen on mobile. If your images aren’t responsive, load slowly on cellular connections, or look terrible on small screens, you’re failing where it matters most. Use srcset for responsive images and test on actual phones.

Measuring What’s Working

Google Search Console

This is your primary dashboard:

1. Go to Performance → Search Results
2. Filter by Search type: Image
3. Track impressions, clicks, and CTR for image searches over time

A rising trend in image impressions means your optimization is working. Flat or declining means something needs attention.

Google Analytics 4

Look for referral traffic from images.google.com. Monitor traffic from Pinterest. Camera-initiated searches sometimes show up as direct traffic, so watch for unexplained spikes in direct visits to product pages.

Pinterest Analytics

If you’re active on Pinterest, track which image styles get the most saves and impressions. Visual search appearances are a specific metric in Pinterest analytics—use it.

The Checklist

Before you call it done:

Technical foundations:

• Images at 1200px+ on longest edge
• File sizes under 200KB (use BulkImagePro if needed)
• WebP format with JPEG fallback
• Descriptive file names with hyphens
• Images not blocked by robots.txt

Metadata and context:

• Specific alt text on every image
• Product schema markup implemented
• Image sitemap submitted to Search Console
• Images placed near relevant text content
• Captions where they make sense

Performance:

• Core Web Vitals passing
• Responsive images with srcset
• Lazy loading for below-fold images
• Mobile-tested and working

FAQ

Does file size affect visual search rankings?

Not directly, but large images slow your pages down, which hurts rankings across the board. Compress everything. BulkImagePro keeps quality high while cutting file sizes.

JPEG or WebP for visual search?

WebP for serving to visitors—better compression, universal browser support. For products, JPEG is fine too. For graphics with text, PNG. The format matters less than the quality.

How many images per product?

At least 3-5: front view, alternate angles, and a detail shot. Eight to twelve performs noticeably better because you’re giving visual search more data to match against.

Do captions actually help?

They provide another text signal that search engines can associate with the image. They also increase user engagement. Worth adding where they feel natural, not worth forcing where they’d be awkward.

Should I put text on product images?

No. Text overlays interfere with object recognition. If you want promotional text, use CSS overlays so the source image stays clean for visual search.

How do I check if my images show up in visual search?

Google Search Console’s image search filter gives you data. You can also manually test—open Google Lens, point it at your products, and see what comes back. If your competitors show up and you don’t, you have work to do.

---

Ready to optimize your images? Try BulkImagePro — batch resize, compress, and format your images for visual search. Process up to 50 images at once, no signup required.