Case: Yiwu Home Goods Seller on Ozon Lifts CTR 2x With New Cards

This is a representative story. The seller, the numbers and the timeline are generalized from scenarios we see regularly among cross-border sellers; no single company is named. If you are a home goods seller on Ozon shipping from Yiwu or Guangzhou, you will probably recognize most of it.

The starting point: good products, invisible listings

The seller ran a small trading company in Yiwu with a catalogue of about 40 SKUs: silicone kitchen tools, drawer organizers, collapsible storage boxes, drying racks and a few bathroom accessories. Prices were competitive, factory relationships were solid, and the team had already sold on Amazon and AliExpress for years.

They joined Ozon Global through the cross-border program and uploaded the same card images they used elsewhere: white-background product shots, a couple of English infographics made in a free online editor, and descriptions run through a translation tool.

After about six weeks the numbers were flat. Roughly 2,000 impressions a day across the catalogue, click-through rate under 1%, and conversion from card view to order of about 1.5%. That is 2 to 4 orders a day on 40 SKUs, which does not even cover the cost of restocking an overseas warehouse.

Diagnosing the problem

The team first blamed pricing and raised ad budgets. Impressions grew, clicks did not. When they finally opened competitor listings side by side, the gap was obvious:

Russian competitors used bright, dense infographics with 3 to 5 benefits written in natural Russian, visible even at thumbnail size.
Their own cards had English text that a Russian shopper simply skips, and machine-translated phrases like "high quality silicone material of food" that read as untrustworthy.
Dimensions were in inches or missing; Russian buyers expect centimeters and liters.
No lifestyle shots. Shoppers could not picture the organizer in a typical Russian apartment kitchen.

Ozon's search gives impressions based on relevance and sales velocity, but it is the main image that decides whether the shopper taps. With CTR under 1%, the algorithm had no reason to push the listings higher. The product was fine. The card was failing.

What they changed

Hiring a Russian-speaking designer for 40 SKUs was quoted at about $60 to $120 per SKU set, with a 3 to 4 week turnaround and several rounds of feedback through a translator. Instead the team tested AI card generation on five bestsellers first.

The workflow was simple. Upload one clean product photo, fill in the product basics (material, dimensions in cm, what is in the box), and let the service produce a full Russian-language set in about three minutes: a main image with a headline benefit, a lifestyle scene, a macro shot of details, a bundle view showing all pieces, a comparison card and a guarantee card. The interface was in English, so nobody on the team needed to read Russian to review the layout; a Russian-speaking freelancer spent an hour checking the copy and found nothing to fix.

Alongside the images they regenerated the descriptions with the SEO description tool, which writes Russian text around the search terms Russian shoppers actually type, such as «органайзер для ящика» instead of a literal translation of "drawer organizer".

For the pilot they kept prices and ad budgets unchanged so the card would be the only variable.

Results after 30 days

The five pilot SKUs moved first, and the rest of the catalogue followed once the team saw the numbers:

**CTR** rose from about 0.9% to roughly 2.1% within two weeks, which more than doubled clicks on the same impressions.
**Conversion** from card view to order went from about 1.5% to around 3.4%.
**Orders** across the catalogue climbed from 2 to 4 per day to about 12 to 16 per day by the end of the month.
**Ad cost per order** dropped by about 40%, because the same ad spend now produced more clicks and more of those clicks converted.
Returns stayed flat, which mattered: honest infographics with correct sizes set the right expectations.

These figures are typical rather than exceptional. In user tests across categories, conversion growth from a card rebuild ranges from modest double digits up to about +260%, and home goods tend to respond well because shoppers compare visually and buy on impulse.

Cost of cards: AI vs designer

For a 40-SKU catalogue the economics looked like this:

A designer: about $60 to $120 per set, so roughly $2,400 to $4,800, plus 3 to 4 weeks, plus translation overhead on every revision.
AI generation: from about $0.6 per card image, so a six-image set costs around $4 to $6. The entire catalogue came in at roughly $200, and the first drafts were ready the same afternoon.

The seller did not fire their designer. They now use AI sets as the baseline for every new SKU and bring a human in only for flagship products or seasonal campaign visuals. Tokens do not expire, so they buy a batch and spend it as new products arrive from the factory. See current rates on the pricing page.

Takeaways for cross-border home goods sellers

Treat the main image as a Russian ad, not a product photo. One benefit, large Russian text, high contrast.
Convert every measurement to centimeters, liters and grams before you generate anything.
Add at least one lifestyle scene that looks like a real Russian home, not a showroom.
Rewrite descriptions around Russian search queries; literal translations rarely match what shoppers type.
Test on your top five SKUs with everything else held constant, then roll out.

If you are new to the platform, our guide on how to sell on Ozon Global from China covers registration and logistics, and how product cards raise conversion explains the mechanics behind the numbers above.

Ready to run the same test on your own catalogue? Start a 7-day trial for $1 with full access and generate Russian cards for your first five SKUs today.