Photos, infographics and video covers on Ozon: what converts

As one seller puts it: the buyer is blind, he buys with his eyes. Everything below comes from seller videos — independent sellers and agencies selling card design — plus two official Ozon materials: an interview with its product-upload manager and a high-season webinar. Transcripts collected mid-2026; videos from 2024 to 2026. Limits change often, so read every figure as "true when the video was recorded".

The first frame has one job

The Ozon manager describes the main photo precisely: the product from the front, in full, large enough to examine, on a neutral background — ideally white in some categories, an interior shot in others. The test: at a glance the buyer must grasp what you sell, its quality and whether it suits him, because he has to fall into the card rather than keep scrolling.

An infographics trainer splits that job in two: catch attention in search, and sell — show the product in use, close objections. Regular shoppers, he argues, develop "banner blindness" and miss a product because everything looks alike. Unnamed research he cites says over 75% of users look at photos first; the source is not identified, so it is not hard data.

The sharpest advice on standing out comes from a long seller course: colours nobody else uses, one promise carried onto the main photo, and in niches like electronics the figure buyers actually read, such as wattage. In softer niches, emotion: for an insulated gilet, skip the studio, take the model outdoors, add snowflakes. Conversion rises sharply, he says, and was even measured — but names no source. The technical part is not negotiable: highest resolution, frames from every side, the product at work.

The four visual layers of a card

One training video splits the media into four and answers "which matters most?" with "all of them". Period limits: photos cut from 30 to 15, videos raised from one to five.

Layer	Where it appears	Limit in the video	Its job
Photos	Top of card; first in search	Up to 15	Earn the click
Videos	After the photos	Up to 5	Win the basket
Video cover	Replaces the main photo	Plays 8 seconds	Catch the eye in the listing
Rich content	Just above the reviews	Visual editor	Warm up the undecided

Photos must differ in scale, style and meaning — a "photo funnel". Ozon's webinar fills it in: a close-up, the product in use, an emotional frame, the texture in macro, frames with and without packaging, and the size chart as a photo of its own, since the chart in card settings must be clicked, and not everyone does. On video, the product should fill most of the running time; in rich content, the commonest mistake is reusing the card's photos.

The video cover: eight seconds — and when it plays

The video cover is an eight-second preview replacing the main photo; upload longer and only the first eight seconds play. Two rules from the webinar: two or three seconds to hook the viewer, and a first frame carrying as much information as possible, because connections stall and the buyer may see nothing else. The product must be in it; an infographic is allowed, but let text appear piece by piece. You can cut eight seconds from an ordinary video, let Ozon's neural network build one free when the panel says the cover is ready, or generate one externally with glare and zoom.

When the cover plays is a point on which the sources contradict each other — critical for anyone testing covers:

• Ozon's webinar: it plays when any promotion tool is attached — display ads, search promotion, the brand shelf, reviews for points. And the reverse: sellers already running ads forget to upload a cover.
• A seller course: it only runs with the display-ad campaign.
• Another seller: it starts as soon as the buyer opens the card; the first photo decides the click, the cover the basket.

The direction is agreed, the trigger is not.

Rules, formats, and who makes the infographics

Moderation is strict: no watermarks, no links to third-party sites, no contact details of your own — such products never pass and get banned, whether new or updated. About 80% clear moderation in under ten minutes.

The white background is period-bound: its last stronghold was electronics, and varied photos should be allowed there soon; an infographic on the main photo is permitted. Ozon's caveat follows — do not overdo it: they often see cards with more infographic than product, when the product must be visible first. The webinar says the same of text: it has its own block.

On numeric requirements the videos disagree; verify them in the panel.

What	What the videos say
Card image ratio	One seller: "Ozon's format, 9:16"; another works at 3:4 and 4K
Rich content, Ozon	640x640 px mobile, 10 MB ceiling (desktop figure unintelligible)
Rich content, Wildberries	900 px wide, 5 MB ceiling
Description text	500-2,000 characters; the title fits three lines

Who makes them? One video's premise: you need no designer at the start. The workflow: download a weak banner, strip the background in a neural network, let it identify the product, then pick a template or use the prompt editor — product name in English, a setting ("in a light bathroom"), generate. Six variants in fifteen minutes; 100 free generations a month, about ten dollars for the paid tier, text overlaid afterwards elsewhere. His framing is honest: not skip designers, but that 15-20 minutes with a generator beats a dull banner published for want of budget.

Did anyone measure anything, and where it goes wrong

The firmest figures come from a slide in Ozon's own webinar: a video cover, with promotion attached, raises display-campaign click-through by a single-digit percentage (the number is garbled, coming out as "2 and 9 per cent"), and plain video content lifts add-to-basket conversion by 16%. Another cites Ozon research that a unique description, tidy infographic, quality photos and video plus completed attributes can raise purchase conversion by over one and a half times.

Seller numbers are weaker and unverifiable: "according to research", photos from every side lift conversion 35% and video can double it; a cover raises sales conversion 15-20%, per "Ozon's 2025 statistics and my own research"; an agency reports goods with infographics sell far better — its own test, and it sells card design.

One measurement points the other way, and may be the most instructive: a team added extra pictures to a card, watched conversion fall, deleted them and saw it recover. Another seller names the real problem: nobody counts which infographic gives a higher click-through. Ozon's answer is candid — no single metric exists in the panel, so do it by hand: put a photo up, note clicks against impressions, swap, compare. The seller version: run each variant through display ads for at least 1,000 impressions. Shortest rule: if click-through is low, change the main photo instead of raising the bid; conversion below 5% signals a problem. And an infographic alone brings no sales — it works with advertising.

The mistakes that keep recurring:

• Reusing the card's photos in the rich content block.
• More infographic than product — Ozon's own complaint — and photos buried under text.
• Watermarks, phone numbers and site addresses on media: they never pass moderation.
• Cramming everything into eight seconds, leaving no legible first frame.
• Running ads and forgetting to upload a cover; or looking like everyone else.
• Never refreshing content; Ozon recommends seasonal updates, and a change for the better costs nothing.
• Adding images without measuring the result.

In short

• The first photo has one job: let the buyer see what you sell, and stop scrolling.
• The cover plays eight seconds; the first two or three decide it, so frame one must stand alone.
• Sources disagree on when it plays: any promotion tool, only display ads, or as soon as the card opens.
• Ozon's one firm warning: never let the infographic outweigh the product; watermarks and contacts are banned.
• The only official measured figures, from Ozon's webinar: video content lifts add-to-basket conversion 16%, a cover lifts click-through a single-digit percentage.
• No panel metric shows a visual change's effect; measurement is manual — and adding images has lowered conversion in real cases.

Commbot tracks image and content state per card, so a list shows which cards lack a photo, cover or rich content.