Website Development Tested on 11 Greek Sites

Website development for a small business needs to work well on mobile, where customers often search for services, products, or contact details. We checked 11 anonymous Greek websites for speed and selected signs of usability by AI agents.

Key findings

• Median mobile PageSpeed was 69/100, based on one lab run per site.
• Median Largest Contentful Paint (LCP) was 9.5 seconds. Ten of the 11 measurements exceeded four seconds.
• Chrome User Experience Report field data was unavailable for all 11 sites, so the times do not describe the average visitor’s experience.
• Agentic Browsing is experimental and is not an overall AI-readiness score.

What did the 11 anonymous site checks show?

The clearest finding in this sample was the slow display of the main content on mobile. The measurements help identify areas for further investigation, but they do not describe all Greek SMEs.

The check was carried out on November 13, 2026, on 11 public websites from five business groups: accommodation, travel services, handmade goods, food producers, and local services. Each group included two or three sites. The sample was neither random nor representative.

Each site had one lab test with PageSpeed Insights and Lighthouse, using a simulated Moto G Power phone and a slow 4G connection. Company names and web addresses are not shown in the table.

The dash means that no Agentic Browsing result appeared in the report for that site. PageSpeed Insights combines lab tests with real-user data when those data are available. [LINK: https://developers.google.com/speed/docs/insights/v5/about]

How should you interpret the speed scores?

The median score was 69 out of 100, with results ranging from 53 to 90. Only one site reached 90. The score alone, however, does not explain what caused a delay or what needs to change.

For a new website build, a lower score can be a useful starting point for diagnosis. It does not, on its own, prove that visitors abandon the page or that the business loses sales. Those conclusions require traffic, conversion, and repeat-measurement data.

Largest Contentful Paint (LCP) measures when the largest main element in the viewport appears. The median was 9.5 seconds, with results ranging from 3.6 to 17.6 seconds. Ten of the 11 lab measurements exceeded four seconds. Google considers an LCP of 2.5 seconds or less a good field result at the 75th percentile. [LINK: https://web.dev/articles/lcp]

There is an important difference between a lab test and real-world use. CrUX data, which summarizes real Chrome visits, was unavailable for all 11 sites. The table therefore does not provide an official Core Web Vitals assessment.

What can affect website speed the most?

Website speed can be affected by large image files, scripts that delay rendering, heavy code, and slow server response. A measurement points to where to look first; a technical review is needed to identify the cause on each site.

CLS measures how much page elements shift unexpectedly as the page loads. Ten sites had a result of 0.1 or less, while one recorded 0.207. A CLS of 0.1 or less is the recommended threshold for a good experience. [LINK: https://web.dev/articles/cls]

A business can start with these practical checks:

• Check whether the main image is too large or loads before the essential text.
• Reduce third-party scripts that are not needed on the initial screen, such as widgets or tracking tools.
• Make sure images and embedded elements have set dimensions so the page does not shift while loading.
• Repeat the measurement and compare the same URLs on the same device and network settings.

Wechevo describes responsive design and technical optimization as part of its website development service and offers speed optimization as a separate service. The distinction matters: a design can look attractive, but measurements are also needed to evaluate its mobile performance.

What does AI readiness mean for a business website?

AI readiness is not one official score. It helps to distinguish eligibility for AI-powered search results from an AI agent’s ability to interact with a website’s elements.

The Agentic Browsing check appeared in 10 of the 11 reports. Results ranged from 0/2 to 2/2, with a different total number of checks for each site. Since the denominators differ, the fractions are not directly comparable and should not be averaged.

The category is experimental. It checks technical signals related to interaction by agents; it does not measure whether an AI tool will mention or recommend a business. [LINK: https://developer.chrome.com/docs/lighthouse/agentic-browsing/scoring]

For Google Search AI features, the official guidance is different. A page needs to be indexed and eligible to appear with a snippet in Search. Google does not require a special schema or separate AI file. [LINK: https://developers.google.com/search/docs/appearance/ai-features]

The practical priorities remain clear: content that answers customer questions, a well-structured page, indexability, clear links, and a good mobile experience. SEO optimization can help organize these priorities, but it does not guarantee appearance in AI responses.

What can a Greek SME conclude from this study?

The study shows why mobile speed and content accessibility are useful parts of website development. It does not show that all websites in Greece perform the same way or that one sector is consistently faster.

Broader data provide a different perspective. According to Eurostat, in 2025, 7.3% of small and 15.2% of medium-sized enterprises in Greece used AI technologies, compared with 17% and 30.4% in the EU. These figures measure business use of AI, not website performance. [LINK: https://ec.europa.eu/eurostat/documents/7870049/23260410/KS-01-26-009-EN-N.pdf/37d063cb-28cf-3b4e-91f3-c3784c970842?t=]

A business can use this process to assess its own site:

1. Choose key pages such as the homepage, services, products, or contact form.
2. Test them on mobile with PageSpeed Insights and record the score and individual findings.
3. Find the cause of delays or layout shifts before changing the design or hosting.
4. Repeat the test after making changes, using the same method.
5. Compare business results such as enquiries, calls, bookings, or purchases alongside the technical measurements.

Good website development cannot be judged by one number or tool. It should make the business’s offer clear, help customers find the next step, and let them use the site without unnecessary delays.

Lessons from the available data

The data in this study do not include Wechevo project results that would support broader claims about what “usually” happens. They do point to practical mistakes worth checking: relying only on the PageSpeed score, treating a lab LCP as an official Core Web Vitals result, or treating an experimental AI score as a prediction of visibility in AI answers.

Keep each measurement with its date, URL, device, and network settings. That way, a later website review can compare like with like and show whether changes improved the experience.

What should you do next?

Start with the page that matters most to your business and measure it on mobile. If the results show slow rendering or unexpected shifts, ask for a review that connects the technical findings with your customers’ needs.

To discuss your website and next steps, contact Wechevo.