Cart and Checkout Product Images That Convert

The thumbnails in your cart and checkout do a different job than your product page hero — and most stores get them wrong.

|e-commerce conversion optimization e-commerce imagery image optimization

The average cart abandonment rate is 70.22%, and on mobile it climbs to 76.98%. Most teams respond by rewriting shipping copy or adding a progress bar. Almost nobody audits the product images inside the cart and checkout — which is strange, because those thumbnails are the only thing on the page still telling the shopper what they are actually buying.

Cart and checkout product images do a different job than the ones on your product detail page. On the PDP, images sell. In the cart, they confirm. A shopper scanning a checkout summary is asking one question: "is this the right thing, in the right colour, in the right size?" A 60px thumbnail that shows a generic packshot instead of the blue variant they picked creates exactly the hesitation that pushes a session into the 70% bucket.

This guide covers what to show at each step from cart through order confirmation, the variant-accuracy problem that quietly breaks most catalogs, and how to size and load these images without hurting checkout performance.

Why cart images are a conversion surface, not decoration

Baymard Institute's research puts the recoverable revenue from better checkout experiences at roughly $260 billion across the US and EU, and finds that fixing checkout UX issues can lift conversions by up to 35.26%. Imagery is a component of that UX, and it is one of the cheapest to fix because the assets already exist — they just need to be selected and rendered correctly.

The function of an image changes as the shopper moves down the funnel:

StageImage jobFailure mode
Category gridEarn the clickProduct unrecognisable at thumbnail size
Product detail pageSell and de-riskToo few angles, no scale reference
CartConfirm the selectionWrong variant shown
Checkout summaryReassure while payingImage missing or broken
Confirmation / emailReduce anxiety and WISMO ticketsHotlinked image blocked by mail client

Note the shift. Everything before the cart is persuasion. Everything from the cart onward is reassurance. Teams that reuse their hero-shot logic all the way through checkout are answering a question the shopper stopped asking two clicks ago.

Pro Tip

Open your own cart on a phone with three items in it, two of them variants of the same product. If you cannot tell them apart from the thumbnails alone, neither can your customers.

The variant accuracy problem

This is the single most common defect, and it is almost always a data problem rather than a photography problem. A shopper selects the sage green colourway, adds to cart, and the cart renders the product's default image — the charcoal one — because the cart template pulls product.featured_image rather than line_item.variant.image.

The shopper now sees a charcoal item in their cart. Best case, they re-check the PDP and lose momentum. Worst case, they assume the site got it wrong and leave. Categories where the variant is the purchase decision suffer most — which is exactly why Beauty & Personal Care carries the highest abandonment rate of any vertical at 82.51%, with shade matching a documented driver.

Cart abandonment rate by context
Beauty & personal care
82.5%
Mobile (all verticals)
77.0%
Overall average
70.2%
Desktop
64.8%

Fixing it takes two things. First, every purchasable variant needs its own image assigned at the variant level — not just present somewhere in the gallery, but explicitly attached to that variant. Second, your cart and checkout templates need to request the variant image with a fallback chain: variant image, then product featured image, then a branded placeholder. Never an empty box.

The gap is usually coverage. A 40-SKU catalog with six colourways each needs 240 variant-level images, and most brands shot 40. Generating the missing colourways from existing shots — rather than rebooking a shoot for a colour that already exists in the warehouse — is where AI variant generation earns its keep. Our guide to colour variant product photos covers the production side in detail.

What to show at each checkout step

Specifications that hold up across most storefronts:

SurfaceRendered sizeContent
Mini-cart / drawer60–80px squareVariant-accurate, tight crop, plain background
Full cart page100–140px squareVariant-accurate, product fully in frame
Checkout order summary50–64px squareSame asset as cart — consistency matters more than detail
Confirmation page80–120px squareSame asset, plus variant named in text
Transactional email120–200px, 2x for retinaAbsolute HTTPS URL, alt text, no lazy loading

Three rules apply across all of them.

Crop tight, but never crop the product. At 60px, a packshot with generous margins renders the product about 30px across. Use a crop that fills the frame while keeping the whole item visible — a cropped-off sleeve reads as the wrong item.

Keep backgrounds plain and consistent. A cart mixing white-background packshots with dark lifestyle shots looks like a broken page. Whatever your PDP does, cart thumbnails should be visually uniform. If your catalog is inconsistent, standardising backgrounds is the prerequisite fix.

Show the variant in the text too. The image confirms; the label proves. "Merino Crew — Sage Green / M" underneath the thumbnail removes the ambiguity a small image cannot fully resolve.

Typical cart

  • Default product image regardless of variant
  • Full-bleed lifestyle crop, product tiny
  • Mixed backgrounds across line items
  • Broken image icon when an asset 404s
  • Email images hotlinked from a CDN that blocks mail clients

Audited cart

  • Variant image with a defined fallback chain
  • Tight square crop, product fills the frame
  • Uniform plain backgrounds
  • Branded placeholder, never an empty box
  • Absolute HTTPS URLs with alt text in email

Performance: small images, big consequences

Checkout is the worst place on your site to ship heavy images. A cart with eight line items loads eight thumbnails, and every one competes with the payment form for bandwidth on a connection you do not control.

The most common waste is serving a 2000px PDP asset into a 64px slot. The browser downscales it and the shopper waits for roughly 1,000 times more pixels than they will ever see. Serve a purpose-built derivative instead.

64pxTypical checkout thumbnail
128pxAsset to serve (2x retina)
<15KBSensible budget per thumbnail

Practical rules:

  • Serve at 2x the rendered size, no more. A 64px slot needs a 128px asset. Use your image CDN's resize parameters rather than uploading separate files.
  • Use modern formats. WebP or AVIF at these dimensions lands comfortably under 15KB with no visible loss. See format comparison for the tradeoffs.
  • Do not lazy-load cart thumbnails. They are above the fold in a short list. Lazy loading adds a visible pop-in that reads as instability.
  • Set explicit width and height. Unsized images shift the layout as they load, and a checkout page that jumps while someone is entering card details is a genuine conversion risk.
  • Pre-warm the derivatives. If your CDN generates sizes on first request, the first shopper to add a given variant pays a cold-start penalty at the worst moment.
Watch out

Many checkout platforms — Shopify's hosted checkout among them — render the order summary from their own templates. You control which image is attached to the variant and its filename and alt text, but not the markup. Get the source asset right, because you cannot patch it downstream with CSS.

Post-purchase: the images that reduce support tickets

Once payment clears, product images shift jobs again. Now they answer "did I order the right thing, and is it on its way?" — the two questions behind most "where is my order" tickets.

The confirmation page and the confirmation email should show the same variant-accurate thumbnail the shopper saw in the cart. Continuity is the point: an image that changes between cart and confirmation makes people re-read the order to check nothing went wrong.

Email introduces its own constraints, and they catch teams out regularly:

  • Absolute HTTPS URLs only. Relative paths resolve against the mail client, not your domain, and render as broken.
  • Alt text on every image. Many clients block images by default; alt text is what a meaningful share of recipients actually read.
  • Fixed pixel dimensions in the HTML. Email clients are inconsistent about CSS sizing, so set width and height attributes directly.
  • No lazy loading, no CSS background images. Support is patchy and the failure is silent.
  • Check your CDN's referrer policy. Hotlink protection configured for your storefront will happily block Gmail's image proxy.

Shipping and delivery notifications benefit from the same treatment. A delivery email showing the actual item is far more useful than an order number alone, especially for households receiving several parcels a week.

Auditing your own cart in under an hour

A concrete pass you can run today. Work through it on a phone, not a desktop — that is where three quarters of the abandonment happens.

  1. Build a mixed cart. Add four items: two variants of the same product, one item from your newest collection, one from your oldest. Old SKUs are where missing variant images hide.
  2. Check variant accuracy. Does each thumbnail show the exact variant selected? Log every mismatch as a data fix, not a design fix.
  3. Check the fallback. Temporarily point one line item at a nonexistent image. Do you get a branded placeholder or a broken-image icon?
  4. Measure the payload. In DevTools, filter the network tab to images and read the transferred size of each cart thumbnail. Anything over 30KB at 64px is a misconfigured derivative.
  5. Look for layout shift. Reload with throttling on. If line items jump as images arrive, your dimensions are unset.
  6. Complete a real order. Check the confirmation page and email on a phone, with images initially blocked, then unblocked.
  7. Log the gaps by SKU. You will typically find the problem concentrated in older products and long-tail colourways rather than spread evenly.

Step seven is usually the surprising one. Most catalogs are not uniformly broken — they have a specific cohort of SKUs that predate a process change. That makes the fix a bounded backfill rather than an open-ended project. Retouchable can generate the missing variant images from your existing photography and push them straight to the right product, with the filename and alt text already set. For the wider catalog view, see our catalog audit checklist.

Frequently Asked Questions

Should cart thumbnails show the variant image or the main product image?

<p>Always the variant image. The cart’s job is to confirm what the shopper selected, and showing a default image when they chose a specific colour or style creates doubt at the highest-intent moment in the session. Implement a fallback chain — variant image, then product featured image, then a branded placeholder — so a missing asset never renders as a broken box.</p>

What size should product images be in the cart and checkout?

<p>Serve at roughly twice the rendered size for retina displays and no larger. A 64px checkout thumbnail needs a 128px asset; a 120px cart image needs 240px. In WebP or AVIF, that lands under 15KB. Serving your full-resolution PDP image into a 64px slot wastes bandwidth on the page where it hurts most.</p>

Do product images in the cart actually affect conversion rates?

<p>They affect it indirectly but measurably. Baymard Institute finds that resolving checkout UX issues can improve conversions by up to 35.26%, and imagery is part of that surface. The mechanism is doubt: a thumbnail that does not match what the shopper selected sends them back to verify, and verification steps are where sessions are lost. The average abandonment rate is 70.22%, so small reductions in friction compound quickly.</p>

Why are my product images broken in order confirmation emails?

<p>Four usual causes: relative image URLs that resolve against the mail client rather than your domain; CDN hotlink protection blocking Gmail’s image proxy; lazy-loading attributes that email clients ignore or mishandle; and CSS background images, which have patchy client support. Use absolute HTTPS URLs, explicit width and height attributes, standard img tags, and alt text on every image.</p>

How do I find which products are missing variant images?

<p>Export your catalog and compare the count of purchasable variants against the count of variant-assigned images. The gap is almost never spread evenly — it concentrates in older SKUs created before a process change and in long-tail colourways. That makes it a bounded backfill you can scope and finish rather than an ongoing problem.</p>

Fill the gaps in your variant imagery

Retouchable generates missing colourway and variant images from photography you already have, so every line item in the cart shows the right product.

Try Retouchable Free No credit card required