Product Photography and Conversion: What the Data Really Says
Product photography lifts conversion and cuts returns, but not by one miracle percentage. The effect is the sum of better information, less doubt and fewer surprises when the box is opened. This article shows which numbers on that hold up when you trace the source, and which do not.
There is a lot of noise around e-commerce photography. Numbers fly across the table: so many percent more conversion, so many percent fewer returns. Many of those figures do not hold up once you trace the source. For this update we did that again. Two figures from the previous version, about a retail chain that lifted conversion with 360-degree photos and a DIY chain that saw fewer returns, came from a 2020 article in which a vendor of image software passed them on from unnamed reports. Too old and too far from the source. They are gone.
What remains is smaller, but it holds up. With sources and concrete implications for your store on bol.com, Amazon or your own Shopify site.
If you take e-commerce photography seriously, you need numbers that hold up. These are those numbers.
Why Product Photos Determine Your Conversion
Buying online is buying without holding. Your customer cannot pick up your product, feel it or smell it. The only bridge between doubt and purchase is what that customer sees on screen, and that is often what returns are about.
In Klarna research cited by Thuiswinkel.org, "looks different in real life" accounts for 19% of return reasons. Not because it was broken. Not because the size was wrong. Because the product looked different in reality than in the photo.
That gap between expectation and reality is a direct result of weak or incomplete product photography. And returns cost more than shipping: handling, customer service, inspection, repacking, sometimes write-offs. According to the ACM figures published by Thuiswinkel.org, the national return rate in the Netherlands was 12.3% of all parcels, for 2022. We did not find a more recent national figure. In fashion, return rates of 30% are not exceptional according to the same source. There, photography that shows reality honestly is a direct lever on your margin.
Professional product photos close that leak in two ways. First, they give an honest and complete picture of the product: colour, texture, proportions, material. Second, they build trust. A sharp, consistently lit product photo says: this is a serious brand that invests in quality.
This is not about "pretty photos". It is about photos that carry information. A main image on white tells the customer what the product is. A mood shot shows how it fits into someone's life. And a detail shot answers the question the customer would otherwise put to your customer service.
Scale and size are an underrated problem here. In its product page benchmark (updated March 2026), Baymard Institute measures that 42% of users try to work out a product's size from the product images. Yet 37% of the sites studied show no image at all that conveys scale. If the estimate fails, the chance of a wrong purchase rises, and with it the chance of a return.
Multi-Angle Photography and Return Reduction
More angles mean fewer returns, because every extra angle answers a question the customer would otherwise only answer after delivery. How thick is the material? What does the back look like? Does it fit in my hand? How does it reflect light?
One photo cannot answer those questions. Five photos can.
Stores that switch to richer product visualisation often report noticeably fewer returns. Figures like that usually come from individual client cases, not broad research across hundreds of retailers. Treat them as illustration, not as a hard rule. The direction holds: a more complete picture, fewer surprises, fewer returns.
For bol.com and Amazon this applies even more strongly. Customers compare directly with competitors there. A listing with three photos next to a listing with eight photos loses. Not because the competitor's photos are prettier, but because more information is available.
What does this mean for your shoot? Plan at least five unique angles per product: front, back, side, detail and a scale shot. For products with texture or material differences, add close-ups. For products used in context, add a mood shot that shows the product in action.
A scale shot does not have to be an expensive photo, by the way. Baymard names a person, a computer-generated model or an object of known size next to the product as a full solution. As long as the proportions are right, the technique does not matter.
360-Degree Product Photography: The Conversion Impact
360-degree photography and other interactive presentation raise the chance of a purchase mainly for more expensive products. The customer rotates the product, views it from every angle and decides on a complete picture instead of a selection you made.
The most recent research we found, covered by Twinkle in February 2026, shows that interactive product presentation (zoom, 360 degrees) lifted sales by up to 54 percent in controlled tests. The effect was strongest for more expensive products, because that is where the financial risk for the buyer is greatest. Interactivity mimics picking the product up in a store.
| Type of presentation | What the research found | Where it pays off most |
|---|---|---|
| Interactive (zoom, 360 degrees) | Up to 54 percent more sales in controlled tests | More expensive products, where the buyer carries more risk |
| Contextual (product in a setting) | In one case 94 percent more conversion than without context | Home decor and sites where context fits the brand |
| Consistent with the platform's norm | On Amazon, white images did slightly better | Marketplace search results |
Source of the table: Twinkle, February 2026.
That last row is the nuance missing from most stories about 360-degree photos. What works depends on what the platform expects. On your own store you have every freedom for interaction and context. On a marketplace you first win by sticking to the norm, and then stand out in the additional images.
The investment is higher than for static photography. You need a turntable, more shots per product and software to assemble the view. Whether it pays depends on your price point: for a twenty-euro product it rarely does, for a product of a few hundred euros it often does.
On Amazon the main image has to meet strict requirements: pure white, at least 1,000 pixels for zoom and a product filling 85% of the image. The additional images leave room to stand out, and most sellers do not fill that room well. The full requirements are in our guide to Amazon product photo requirements.
Professional Photography vs. AI Product Photography
AI product photography wins on speed and variation, photography on proof. The right approach is not a choice between the two but a split: photograph what has to provide proof, generate what is variation.
The current generation of image models is a step beyond a year ago. Google's image models accept up to 14 reference images to place objects recognisably in a new image, and every generated image carries an invisible SynthID watermark. For background swaps and colour variants of a product that has already been photographed, that is a serious acceleration. You test ten variants in the time a photographer makes one.
But there is a limit. The model reconstructs your product from references and fills in what it cannot see. A seam, a label, the grain of ceramic or the way light falls on a specific fabric: there it invents something plausible. A handmade ceramic mug looks like a neat mug in a generated image. In a real photo it looks like a handmade ceramic mug.
Customers often cannot tell the difference. 71% of Dutch people find it hard to tell AI images from real photos, according to Norstat (November 2025). That is exactly why the risk sits with you: show an image that makes the product nicer or different than it is, and it comes back as a return or a bad review.
Transparency also costs you less than you might think. Research agency Validators tested five commercials with groups who were or were not told beforehand that AI had been used (Adformatie, September 2026). The group with the AI disclosure scored slightly lower on almost every attribute, but the differences were small. "The differences are small and not convincingly significant," says Tycho Slors, insights consultant at Validators (translated from Dutch). Openness is therefore not a commercial risk; concealment is.
Use AI for variants, backgrounds and quick iteration. Use professional photography for your main images, your mood shots and every image that has to build trust. We worked out the trade-off between photography, CGI and AI separately in CGI versus AI product visualisation, and what you actually buy from an AI tool in AI product photos: a licence, not an image.
What you should not do: hand your entire visual identity to AI without human direction. Without a briefing, without brand context, without someone who knows what your customer expects. That produces images that look like your competitor's AI images.
At Oase we combine both. We shoot on location and in the studio for the images that have to feel real, and we use AI where it is faster and cheaper without the customer seeing anything that is not true. The situation decides, not a principle. See our product photography service for a full overview of what we deliver.
Amazon and bol.com: Platform Requirements and Opportunities
Amazon sets strict requirements for the main image, while bol has not enforced its image rules since July 2025. On both platforms the opportunity lies in the additional images, not the main image.
According to its own photo guide for sellers, Amazon asks for a pure white background (RGB 255, 255, 255), a product filling at least 85% of the image and a size of 500 to 10,000 pixels on the longest side, with zoom from 1,000 pixels. The main image must be a professional photograph of the actual product, without text, logo or watermark. Since 22 July 2026 you must also flag any image containing a photorealistic AI-generated person in the metadata (CNBC, July 2026).
Amazon allows up to nine images per listing. Many sellers use three or four, often the same white shot from slightly different angles. That is a missed opportunity.
The additional images are where you tell your brand story. Infographics that highlight your strengths. Mood shots that place the product in context. Comparison tables that set you apart. Dimensions that prevent returns. How to make the most of that space is covered in our guide to Amazon A+ content.
On bol it works differently. Bol states on its partner platform that since 1 July 2025 it no longer enforces its image rules, with self-promotion in images as the exception. Nobody rejects a weak main image there anymore. You only notice it in your sales figures, which is a much later and more expensive signal. What that means when you send the same file to Amazon and Google as well is covered in why your white is not 255.
A complete image set per product includes at least: a main image on white, two to three extra angles, a mood shot, an infographic, a scale shot and a detail shot of material or texture.
Discover how we design product listings that use every available image position.
The Business Case: Calculating Photography as an Investment
Product photography pays for itself as soon as a better image set improves conversion or returns even slightly on a product that sells regularly. How much exactly, you can only estimate in advance. That is why we calculate below with assumptions you can adjust yourself.
According to their own price lists, Dutch studios charge 8 to 15 euros per standard packshot. For a product with eight images (main image, three angles, mood, infographic, detail, scale) you therefore pay 64 to 120 euros in packshots, plus whatever you add in mood shots and infographics. Those rates and their sources are in product photography costs 2026.
| Worked example | Now | After a better image set | Assumption |
|---|---|---|---|
| Visitors per month on the listing | 1,000 | 1,000 | Same traffic |
| Conversion | 3% | 4% | A third better, no guarantee |
| Sales per month | 30 | 40 | |
| Revenue per month at 35 euros | 1,050 euros | 1,400 euros | Plus 350 euros per month |
| Returns at 30% (fashion) | 12 | 9 | A quarter fewer returns |
| Return costs at 10 to 15 euros each | 120 to 180 euros | 90 to 135 euros | Own estimate of handling costs |
With these assumptions an image set of 64 to 120 euros pays for itself within a month on the extra revenue alone. In fashion the return side adds another 30 to 45 euros per month. If you sell in a category with fewer returns, that saving is smaller, but the mechanism stays the same.
Read the table as a worked example, not a promise. A one-third rise in conversion is an order of magnitude for listings currently running on too few or weak images, not a measured average. The point is the ratio: the cost of an image set is one-off and small, the gain is monthly.
Fill in your own numbers: visitors, conversion, price and return rate. If your listing already runs on a complete, strong image set, the gain is small. If it runs on three phone photos, it is large.
The question is not whether professional product photography is worth it. The question is how much revenue you miss without it. Unsure about the approach for your range? Get in touch and we will look at it together.
