Latest / The DTC Marketing Podcast with Fexingo: Direct-to-Consumer Brands, Shopify, and E-Commerce Growth / How DTC Brands Use AI-Generated Product Photography
Transcript
- Lucas: You run a DTC brand, you know the drill — you need fresh product shots for a new season, and that means booking a studio, hiring a photographer, shipping samples, praying the lighting works. A shoot day runs you anywhere from three to ten thousand dollars, easy. Luna: And if you're iterating fast — say, launching a new color every two weeks — that cost gets brutal fast. Lucas: Right. So a growing number of DTC brands are doing something that sounded like science fiction three years ago — they're generating product photography with AI. Not just background removal or retouching. I'm talking about synthetic lifestyle shots where a jacket that exists only as a CAD file is being worn by a person who doesn't exist, in a location that was never photographed. Luna: Okay, that's the big picture. Let's get specific. What's an actual brand doing this right now, and what tool are they using? Lucas: Let's talk about a brand I'll call Wilder Goods — mid-size outdoor apparel, does about fifteen million a year, mostly through Shopify. Before AI, they were shooting two major campaigns a year, spending about eight grand per shoot. They now use a tool called Photoreal — it's one of the newer platforms that lets you train a custom model on as few as twenty catalog images. Luna: So you upload twenty flat-lay or mannequin shots, and the model learns the garment's texture, drape, fit? Lucas: Exactly. Once the model is trained — takes about four hours — you can generate two hundred lifestyle images in an afternoon. You describe the scene: 'model wearing the alpine parka, hiking trail, golden hour light.' The model renders the jacket on a synthetic person with realistic shadows and fabric folds. Wilder Goods now generates about eighty percent of their social media and email campaign imagery this way. Luna: Eighty percent is huge. What about the product detail pages — the hero images where customers really zoom in on stitching and fabric? Lucas: That's the line. They still shoot those hero shots traditionally. AI still struggles with precise texture rendering — you can get weird artifacting on zippers or seams if you push it too hard. But for the secondary images — the lifestyle shot that shows the jacket in context, the email hero, the Facebook ad creative — synthetic images are performing at or above real photography in A/B tests. Luna: And the cost difference has to be dramatic. Lucas: Wilder Goods told me they spent about twelve hundred dollars on their initial Photoreal setup — model training, a month's subscription, some compute credits. Their next traditional shoot would have been eight thousand. So they're saving roughly eighty-five percent on that portion of their visual production. And the iteration speed is the bigger win — they can now test a new product image variant in hours, not weeks. Luna: I want to talk about the ethical side in a minute, but first — how did the synthetic images actually perform? Did they convert? Lucas: They ran a controlled test. Same product, two versions of the same email campaign. One used a real photo from their last shoot — model, outdoor location, professional lighting. The other used a synthetic image generated by Photoreal. Same copy, same subject line. The synthetic variant had a click-through rate that was ninety-seven percent of the real photo — essentially a tie. But the synthetic image cost a fraction to produce. Luna: Ninety-seven percent. That's close enough that the cost savings make it a no-brainer for a lot of applications. Lucas: Right. And for social ads, the synthetic images actually outperformed in some cases — because they could generate more contextual variety. Instead of one hero image, they could produce ten different lifestyle scenes — urban, trail, mountain, coffee shop — and test which one resonated with different audience segments. Real photography at that scale just isn't feasible. Luna: So the creative constraint shifts from 'what can we afford to shoot' to 'what can we imagine and prompt.' That changes the role of the creative director. Lucas: Completely. The skill set becomes less about technical photography and more about art direction through language. You need someone who can write a good prompt — specific lighting references, compositional cues, emotional tone. One brand I spoke with said their best AI image generator was their copywriter, not their photographer. Luna: Okay, but let's hit the ethics question. Consumers are getting savvier about synthetic imagery. Are brands labeling these as ai generated? And should they? Lucas: This is the unresolved piece. Currently, most brands using synthetic product photography do not label it. The rationale is that traditional photography is already highly staged — lighting, retouching, compositing. AI is just another layer. But a 2025 survey from the Consumer Technology Association found that sixty-three percent of consumers want clear disclosure when product images are ai generated. And the FTC has signaled they're watching this space. Luna: I think there's a reputational risk, especially for DTC brands that trade on authenticity. If a customer finds out the model wearing that cozy sweater doesn't exist, it could feel deceptive. Lucas: Agreed. Some brands are starting to add a small 'synthetic imagery' note in their footer or in the image alt text — not prominent, but traceable. Others are leaning into it openly. A brand called Unspun — they make custom-fit jeans — actually promoted their use of AI photography as a sustainability story: fewer sample shipments, no waste from unsold seasonal inventory. Luna: That's smart framing. Turn the potential negative into a brand value. Lucas: Exactly. And that's where I think the winning play is — not hiding the technology, but integrating it into your brand narrative. If you're a DTC brand that talks about innovation, efficiency, or sustainability, synthetic photography is a perfect proof point. Luna: So where does this go next? Are we heading toward fully synthetic catalogs? Lucas: I think within two to three years, the majority of DTC product images you see on social media and email will be at least partially ai generated. The technology is improving fast — the latest models can already handle transparent objects and reflective surfaces, which were the last holdouts. The question is not whether it's technically possible, but whether brands can navigate the trust issues. Luna: Quick honest thing — and this ties back to what we're talking about, because these conversations take real time to research and produce. The reason we can keep this show ad-free and dig into specifics like Wilder Goods and Photoreal is that a handful of listeners chip in monthly through buy me a coffee dot com slash fexingo. It's not a huge number, but it literally funds the research hours and the production. Lucas: Yeah, and we don't run ads, we don't do sponsored segments — so if these marketing conversations have sparked something you've actually used, that's the whole engine right there. No pressure, just wanted to name it. Luna: Exactly. And speaking of practical takeaways — if you're a brand considering AI photography, what's the minimum viable test? Lucas: Start with one product. Take your best-selling item, train a model on twenty clean images, generate ten lifestyle variants, and A/B test them against your current best-performing creative. Measure click-through and conversion rate. My bet is you'll see performance within a few percentage points — and the savings will speak for themselves. Luna: And if the test works, you scale to your top ten SKUs, then your seasonal collection. The infrastructure is already there. Lucas: Right. And keep an eye on the labeling landscape — I expect some industry guidelines to emerge within the next year. But for now, the main barrier is imagination, not technology. The brands that figure out how to blend synthetic and real imagery in a way that builds trust, not undermines it, are going to have a serious cost advantage. Luna: And a speed advantage. In DTC, speed is everything. Lucas: Exactly. So the question for every brand right now is simple: are you still booking a photographer for every new product shot? Because that model is about to feel very expensive, very fast.