Image Budget Cut in Half — Now What? What the Numbers Say About Fashion and E-Commerce

Anyone working in marketing for fashion, apparel or beauty brands is living a difficult contradiction: the budget for image production has been cut, but the demand for visual content has never been higher. More channels, more formats, more product variations, faster publication cycles. The math doesn't add up — at least not the way production used to work.

The immediate reaction of many brands was to bet everything on generative AI. Fast, cheap, no studio or crew needed. It looked like the perfect solution to the problem. But the numbers now emerging tell a different story — and a more expensive one than expected.

More channels, more formats, more variations — demand for visual content in fashion brands has grown while production budgets have been cut. Solving this equation is the central challenge of product marketing today.

The cost no one is putting in the spreadsheet

There is a cost most brands don't directly link to product image quality: the return rate. And the numbers are eye-opening.

According to data from Red Stag Fulfillment, fashion e-commerce records return rates of 30% to 40% on average — reaching up to 88% during heavy promotional periods. And according to a ClickPost study, 22% of consumers directly cite the gap between the online image and the product they received as their reason for returning.

Run the math: if a brand does $1M a month in e-commerce with a 30% return rate, and 22% of those returns originate from inadequate imagery — that's $66,000 per month evaporating from a problem that starts in visual production. Every month.

This cost rarely appears in the marketing budget. It ends up in operations, in reverse shipping, in returned inventory. But the origin is in the image — and specifically in the gap between what the customer saw and what arrived in the box.

Why cheap AI can be the most expensive investment

The logic seems straightforward: cut image production costs using generative AI. Less studio, no photographer, faster turnaround. The result appears quickly and budget is freed up.

The problem lies in what generative AI does with the product when it's the hero of the image. It interprets — it doesn't reproduce. The fabric tone the client approved in the showroom may appear with a subtle shift the consumer notices on delivery. The garment's drape may differ from the real product. The texture that justifies the premium price point can disappear in the generated image.

And when that happens, the consumer doesn't call to complain about AI. They return the product and don't buy again. According to ClickPost, 84% of UK consumers say they would not buy from a brand again after a poor return experience. The cost isn't just logistical — it's relational and reputational.

Texture, drape and exact color tone are elements that drive purchase decisions — and that generative AI interprets rather than faithfully reproduces.

What the data says about image demand — and where the budget went

Image production budget didn't disappear. It was redistributed. Understanding where it went is what separates brands that are growing from those cutting costs and losing margin.

According to the Business of Fashion, European e-commerce is expected to grow 11% annually between 2022 and 2030, with the US at 7% in the same period. More e-commerce means more product pages, more image variations per SKU, more content for social channels, more material for targeted campaigns. Demand for images hasn't dropped — it has multiplied.

Italian luxury fashion brand Etro is a concrete example of what happens when this equation is resolved correctly. According to CEO Fabrizio Cardinali at BoF VOICES 2025, the brand adopted AI in its e-commerce content production while keeping the human element at the center of the process — and grew its e-commerce business by 46% in 12 months. This isn't about cutting costs. It's about producing more, at the same quality, directing budget toward what generates return.

What genuinely needs photography — and what AI can handle

The decision isn't binary. It's not "photography only" or "AI only." It's understanding, product by product and channel by channel, what each approach delivers — and where the mistake will cost more than the saving.

Needs real photography: any image where exact color, texture, drape or packaging typography are purchase decision elements. Fashion in general, jewelry, cosmetics with visible packaging, premium pieces where quality perception is built visually. In these cases, AI-generated images will interpret the product — and the gap between that interpretation and reality will show up in the return rate.

Can be generated or complemented with AI: scenes, environments, background variations, ambient color adjustments, multiple formats from the same photographed product, skin and fabric retouching, upscaling for large formats. All of this AI handles well — with speed and cost that physical production can't match.

The hybrid workflow resolves both simultaneously: the product is photographed in studio with full technical control. The scene is generated by AI. Post-production integrates both — equalizing light, perspective and grain. The client receives images with real product fidelity in environments that would be impossible or unviable to produce physically. This is the model being adopted by brands that grew in e-commerce over the last 12 months.

The hybrid workflow combines what each approach does best: real photography for product fidelity, AI for scenes and variations. The result is more images, faster, at significantly lower production cost.

The question that changes the budget

Most brands are asking the wrong question. "How do we produce cheaper?" leads to cuts that create larger problems. The right question is: which image is costing more — the one that was expensive to produce, or the one that was cheap and is generating returns?

An e-commerce operation that produces 500 SKUs per collection with generative AI and has 5% higher returns than it would with product-faithful imagery is paying an operational cost that likely exceeds whatever it saved in production. This calculation is rarely done — but it should be the first step in any image budget decision.

The AI market applied to fashion is growing at 40% annually and is expected to reach US$ 6.99 billion by 2029, according to Research and Markets. That growth isn't about replacing photography — it's about using AI where it delivers real value, in combination with technical processes that guarantee fidelity where it's still necessary.

How brands that are growing are solving this in practice

The pattern appearing among best-performing fashion e-commerce brands is consistent: they didn't choose between photography and AI. They chose to understand what each one solves — and built a workflow that uses both in the right place.

For collections where the product changes constantly and SKU volume is high, they photograph the product in studio and use AI to create the scene variations, formats and applications that channels require. For campaigns where the environment is the central creative element, they generate the scene with AI and integrate the photographed product into the composition. For large-format material — catalog, billboard, window display — they maintain high-resolution photography with precise technical post-production.

The result is more content, in less time, with a budget that goes further — without sacrificing the fidelity that makes the consumer trust the image and, when the product arrives, not return it.

If your brand is navigating this equation and wants to understand which combination makes most sense for your volume, segment and budget, talk to Kado. Over 20 years producing commercial imagery for fashion, beauty and e-commerce — and the answer for your specific case starts with a conversation.