For most of digital commerce’s history, “Is this a good product image?” was answered by whoever had the most seniority in the room. Brand managers had taste. Creative directors had instincts. Nobody had a number, so nothing shipped against a standard. It shipped against an opinion.
That’s starting to change, and the data behind the shift is worth paying attention to, especially for teams that have spent years treating copy and specs as product data while treating imagery as something closer to art direction.
Three Years, 100 Million Images, One Score
Vizit, one of Sitation’s AI and Commerce partners, spent three years building what it calls the Content Conversion Standard: a model trained on more than 100 million images and over 10 trillion retail performance data points across 20-plus countries. The goal was to answer one question with a number instead of an opinion: will this image convert a shopper at first impression, on a phone, in a fraction of a second, on a retailer page they didn’t choose?
The research team started with 1,140 candidate signals, things like where the product sits in the frame, how much of the pack is legible at thumbnail size, color temperature, and logo prominence. Every signal had to earn its place by predicting real conversion outcomes on data the model hadn’t already seen. By year two, the list was down to 184 signals. By launch, 87.
What Actually Predicts Conversion
The signals that survived are less about art direction and more about basic legibility. Pack-front clarity at thumbnail size carried the heaviest weight. Product centrality, foreground-to-background contrast, claim hierarchy, and cropline geometry all made the final cut.
What didn’t survive is the more interesting result. Brand color saturation, logo size as a percentage of the frame, image resolution above 1080 pixels, and how “premium” the background looked all got cut. Most notably, photographic style, the dimension creative teams argue about the most, turned out to be a near-zero predictor of conversion once pack-front legibility was accounted for. As Vizit’s research team put it, the shopper doesn’t care whether the product is lit moodily. The shopper cares whether they can tell what’s in the package.
That’s a meaningful reframe. A lot of creative review time gets spent debating lighting, mood, and composition when the far bigger lever is whether a shopper can identify the product and its core claim within the first fraction of a second of seeing it.
The Cost of Getting This Wrong
The stakes here aren’t abstract. A separate eye-tracking study Vizit ran across nearly 15,000 shoppers and six categories found that mobile hero images lose 23% of shoppers in the first half-second of viewing, before a shopper has consciously processed anything about the product. That’s a substantial share of potential buyers lost to a visual decision made in production, often without anyone realizing a decision was being made at all.
On the flip side, brands that treat imagery as something to systematically test and optimize see real movement in the numbers. Vizit’s own published case studies point to a baby products brand that saw a 29% increase in Amazon conversion rates after optimizing its imagery, and Mars Petcare, which reported a 30% conversion lift on SKUs where pack-front cues were specifically optimized based on the scoring model.
Treat Visual Content Like Product Data, Not Decoration
Most PIM and content strategies are built to enforce rigor on the things that are easy to measure: is the title within the character limit, are the required attributes populated, does the description meet the minimum word count. Imagery usually escapes that same discipline, reviewed on instinct rather than scored against a standard.
That gap is worth closing. If pack-front legibility, product centrality, and claim hierarchy can be measured and scored the same way a spec sheet can be validated, there’s no reason imagery should be the one part of a product page still governed by taste instead of data. The brands seeing double-digit conversion lifts aren’t the ones with the most artistic photography. They’re the ones treating every image the way they’d treat any other piece of structured product content: something that gets tested, scored, and improved on evidence.
As a Vizit partner, Sitation helps brands bring that same rigor to visual content that we already bring to specs, titles, and claims. If your product pages are winning on copy but you’ve never actually measured whether the imagery is converting, that’s a gap worth closing before your next content refresh.
Ready to find out where your product imagery actually stands? Contact us to see how our digital merchandising and applied AI services, alongside our partnership with Vizit, can help you score, test, and improve the visual content already sitting on your product pages.
