Online shops from the Moravian-Silesian Region: they look good, but customers can't find them

Online shop owners most often come to me saying the shop "somehow doesn't work" and asking for a new look. So on 25 October 2026 I measured 19 online shops whose operator is based in the Moravian-Silesian Region. The result is not what they would expect. Design was hardly ever the problem: the median design score came out at 4 of 5. The problem was whether a customer can find the shop and whether the product page answers their questions before buying.

For 12 national queries such as "toys e-shop" or "wine e-shop" (in Czech), the search an AI assistant draws on returned 105 results. Four of them were small online shops and none of them was one of the shops I measured. An online shop is therefore not finished on launch day. Without ongoing care for content and SEO, customers will not find it and will buy elsewhere, even from a shop with a weaker offer.

What I measured and how

This builds on my earlier measurement of 21 Czech online shop websites, which looked at technology. Here I was interested in what decides a sale: can a customer find the shop, do they trust it and does the product page tell them everything they need.

• When: 25 October 2026, a single measurement.
• Selection: online shops that appeared in search results for "e-shop Ostrava" (in Czech). I verified the operator by the company ID on the website in the Czech business register ARES. Of 23 measured shops, 19 had their registered office in the Moravian-Silesian Region; two were based elsewhere and for two I did not find a company ID.
• Categories: pet supplies 3, furniture 3, wine 3, toys 2, gifts 2, and clothing, sporting goods, jewellery, bicycles, health food and cosmetics 1 each. 10 of 19 ran on Shoptet.
• Findability: 12 national queries " e-shop" and a query for the shop's name, first 10 results. The tool is the web search used by an AI assistant (Claude), not Google.
• Shop: homepage and one product page rendered in a browser on desktop (1440 px) and mobile (390 px), with the cookie banner declined. A rubric of 7 areas with 0 to 5 points each: design, trust, product, shipping and payment, mobile, speed, SEO. Lighthouse in mobile mode.
• Design was scored from screenshots by a model against fixed criteria (hierarchy, photos, consistency, legibility, path to purchase). It is an expert judgement, not a user test.
• What I did not do: I bought nothing, added nothing to a basket and submitted no forms. Results are aggregated, without shop names.

Finding 1: design is not the problem

I could score the design of 17 shops; for two the screenshot was covered by a consent or privacy settings window. Six shops received the full 5 points, seven received 4. On mobile, content did not overflow sideways on 17 of 19 shops, and 15 of the 18 rated mobile screenshots showed no obvious problem.

So if you are considering a redesign because "the shop doesn't look good", first check whether that is really it. For most of the shops I measured, the weak spots lie elsewhere.

Finding 2: customers cannot find the shop

For 12 general queries " e-shop" (clothing, toys, bicycles, sporting goods, furniture, pet supplies, wine, jewellery, health food, gifts, cosmetics, coffee), the AI assistant's search returned 105 results.

I managed to run a query for the shop's own name for 16 shops. Its website did not appear in the top 10 even once. It was the same for service websites in my measurement of Ostrava businesses: this search barely sees small Czech websites and builds answers from directories, comparison sites and articles.

Google may look different, and I did not measure it. For an AI assistant, though, it means one thing: when a customer asks where to buy pet food or a gift, it answers from what it finds. A small shop with no content worth linking to and no presence elsewhere (comparison sites, directories, reviews, articles) will not be in the answer.

Finding 3: the product page does not answer pre-purchase questions

I found and measured a product page for 15 of 19 shops. All of them showed the price. The rest was weaker:

A customer on a product page asks: how much is it, is it in stock, when will it arrive and how much is shipping. All shops answered the first question. Only a third answered "when will it arrive". Product structured data was missing on a third of them, even though that is exactly where search engines and comparison sites read price and availability from. What a product page must not miss is covered in my article What a product page should contain.

Finding 4: shipping and trust

• Shipping: 15 of 19 shops linked to a shipping and payment page. I found the shipping cost or a free-shipping threshold directly on the homepage or product page on only 8 of 19. Other customers learn the shipping cost on another page or in the basket.
• Operator: I found a company ID for all 19, but only 2 had it directly on the homepage. The others have it under contact details or terms and conditions. That is fine, the customer just has to click through to it.
• Terms and complaints: 14 of 19 linked to both terms and conditions and complaints on the homepage.
• Reviews: I found a mention of reviews, ratings or Heureka on 11 of 19.
• Cookie banner: I captured one on 17 of 19. It could be declined on the first layer on 5 of 17, on 7 only via settings, and on 5 my script could not decline it automatically at all.

Finding 5: they are slow on mobile

Lighthouse in mobile mode could be measured on all 19 shops. The median performance score was 44 out of 100, 12 of 19 scored below 50 and none scored 90 or more. The largest content element (LCP) rendered in the lab test in a median of 13.5 s, and in more than 4 s on 18 of 19. 6 of 19 had an accessibility score of 90 or more.

This is a single lab measurement with the default network and CPU slowdown Lighthouse simulates for mobile, not real-customer data. On a good connection the shop will be faster. But the big gap compared with design shows where online shops get the least attention: design is visible immediately, speed on a phone only in the numbers. What to do about it is in my article How to speed up a website.

An online shop is not finished on launch day

The operators of the measured shops are not beginners. The median company age in ARES was 14.4 years and 17 of 19 have been trading for five years or more. The shops look fine and mostly run on a sensible platform. What they lack is work that cannot be done once: adding information to products, writing content worth linking to, collecting reviews, watching which queries the shop shows up for.

That is exactly the part owners put off after launch, and then they look for the problem in the design. Yet a new design alone will not bring a single extra customer if nobody finds the shop and the product does not answer basic questions. How to bring more people to an online shop is in my article How to increase online shop traffic, the most common mistakes in SEO mistakes of online shops.

How to check your own shop

This study does not tell you where your shop ranks on Google. You can find that out yourself, for free and more precisely. Answer two questions: which queries does the shop show up for and in what position, and does AI cite it.

According to its announcement, Google made the generative AI performance reports available to all websites as of 31 August 2026; they show impressions in AI Overviews and AI Mode. Since 10 February 2026 Microsoft has offered AI Performance in Bing Webmaster Tools as a public preview. How Czech online shops are set up for AI crawlers I measured in AI crawlers and 30 Czech online shops.

What to do first

On Shoptet most of the changes in the second row can be done without a new shop. If you need help, I handle it as part of SEO for online shops and website management. A new shop only makes sense when the current platform genuinely cannot do what is needed; then have a look at a custom online shop.

Limitations

• Small, non-random sample. 19 online shops from one region, selected from search results. This is not a representative survey.
• A single point in time. Everything is from 25 October 2026.
• An AI assistant's search is not Google. I did not measure Google rankings and they cannot be inferred from this study. I did not get to the shop-name query for 3 of 19.
• Automated text checks. A delivery time or shipping cost shown in an image, a pop-up or only after a click is invisible to the check. The numbers are therefore more of a lower bound.
• Design is a judgement. A model scored it against fixed criteria from screenshots, not a test with people. For two shops the screenshot was covered by a consent window.
• Lighthouse is a single lab measurement in mobile mode, not data from real visitors.

Frequently asked questions

Does this mean these shops are not on Google?

No. I did not measure Google. The study shows what the search an AI assistant draws on sees. Where your shop ranks on Google is in the Performance report in Search Console.

Is a redesign of my online shop worth it?

Only if you know what it should solve. For the measured shops, design was the strongest area. Check findability and product pages first; that tends to be cheaper and faster.

Why does product structured data matter?

It describes price, availability and other details in a form that search engines and comparison sites can read. It does not guarantee a position, but without it price and availability are less likely to make it into results.

Which shops were in the sample?

I do not publish names or domains. The aim is to describe what is common, not to point at specific shops. I excluded my own clients and large chains from the sample in advance.

Is being on a comparison site enough?

A comparison site helps with the visibility of specific products and reviews. But without your own content and good product pages you compete mainly on price.

Can you help me with this?

Yes. I will go through Search Console, your products and the path to an order, list what is missing and order it by impact. Get in touch via the contact page.

Measured on 25 October 2026. Aggregated data and the methodology are available from me on request, without shop names or domains.