Most brands discover the same thing at the same moment: right around image 47 of a planned 300, somewhere between the third round of AI re-generations and the second export settings adjustment.
It takes way longer than they thought.
Product image optimization sounds like a background task — something to knock out in an afternoon. In practice, it’s one of the more time-consuming corners of ecommerce operations. Here’s an honest look at why.
The Work You Can See — and the Work You Can’t
The visible part of image optimization is the editing: cropping, color correction, background cleanup, sizing for different placements. Most brands have a rough sense that this takes time.
What catches people off guard is the invisible layer underneath — the technical work that actually affects page speed and search performance.
Every image needs a compressed version that doesn’t sacrifice visible quality. It needs a descriptive file name (not IMG_4832.jpg). It needs alt text that’s specific enough to be useful for accessibility and search. It needs to be sized correctly for mobile and desktop — ideally served as different files, not the same one scaled down by the browser.
For one product, that’s manageable. For a catalog of 500 SKUs, it’s a significant project. And if that catalog has seasonal updates, it’s ongoing.
Where AI Fits — and Where It Doesn’t
AI tools have made parts of this faster, and that’s genuinely useful. Background removal that used to take ten minutes per image now takes ten seconds. First-pass compression is nearly instant. Batch processing means you’re not manually exporting one file at a time.
But here’s the part the demos don’t show: the back-and-forth.
AI output for product images is inconsistent in ways that matter. A background removal tool that handles 90% of your catalog cleanly will also flatten a reflective surface, clip a fine texture, or add an artifact along a curved edge — and you won’t catch it until you’re looking at a customer-facing page.
So the workflow becomes: run the AI pass, review every output, flag the ones that need correction, re-run or manually fix, review again. For a large catalog, that review loop is the majority of the time. The AI handles the easy cases quickly. The hard cases — which are often your best-selling or most photographically complex SKUs — still need human attention.
This isn’t a complaint about AI tools. They’re a real efficiency gain. But “AI does it for you” is a different thing from “AI does it well enough to ship without checking,” and the gap between those two is where the hours go.
The Numbers Add Up Faster Than You’d Expect
Here’s a rough sense of what a mid-size catalog actually requires.
A brand with 200 SKUs and two images per SKU is looking at 400 images. If each image needs compression, resizing for mobile, alt text, and a quality check — even at an optimistic five minutes per image after AI assistance — that’s over 33 hours of focused work. Add in file naming conventions, folder organization, re-uploading to the platform, and clearing cache, and you’re closer to a full work week.
That’s for a one-time cleanup. If the catalog updates quarterly, that clock starts again.
Most ecommerce teams don’t have a dedicated person for this. It lands on whoever is available — usually a designer who has other priorities, or a developer who finds the work tedious. Either way, it gets deprioritized, done partially, or done inconsistently. And inconsistent optimization is almost as bad as none at all, because the slowest images still drag down page performance.
The Compounding Cost of Letting It Slide
Here’s what makes this worth taking seriously: the impact isn’t static.
A product page with unoptimized images doesn’t just load slowly today. It loads slowly for every visitor, every day, until someone fixes it. That’s conversion rate drag that quietly accumulates across every session.
Portent’s research found that a site loading in one second converts at nearly three times the rate of a site loading in five seconds. Images are typically the single largest contributor to that gap. A 1.4MB hero image and an 180KB version of the same photo look identical to your customer — but one of them is quietly costing you sessions every time someone bounces before the page finishes loading.
What to Do About It
The practical answer is to treat image optimization as a discipline, not a one-time task — and to be realistic about the resource it requires.
For small catalogs, building a solid checklist and running it on every new product upload prevents the problem from compounding. For larger catalogs that have accumulated years of unoptimized images, a dedicated cleanup project — often best handled by a team that specializes in exactly this — is usually the faster path to results.
The goal isn’t perfect. The goal is consistent: every image loads fast, looks crisp, and has the metadata it needs to be found.
If your team is hitting that back-and-forth loop and the backlog keeps growing, that’s a signal — not a personal failing. It’s just a harder job than it looks.
Want a quick read on where your catalog stands? Send us three product URLs and we’ll come back with a specific assessment.
To better visuals,
Lily at Melon Tree