Published · Updated
In short: in our Estonian online stores speed and SEO study we tested the home pages of 47 well-known Estonian online stores on 14 November 2026. The median mobile Lighthouse performance score was 61/100, and only 2 stores (4%) reached 90. The median Largest Contentful Paint (LCP) was 8.5 seconds against Google’s 2.5-second “good” threshold, which only 3 stores met. 40% of home pages had no schema.org markup at all, and 8 stores (17%) had an llms.txt file. Estonian stores have the most to gain in mobile speed, structured data and content that AI search can read.
This is our own Estonian online stores speed and SEO study. We checked the home pages of 47 Estonian online stores from two angles: how fast the page loads on a phone, and how well Google and AI assistants such as ChatGPT, Perplexity and Google AI Overviews can read it. We report aggregate results only; we do not rate individual stores but show where the whole market has the biggest opportunities.
How was the study done?
- Sample. 47 well-known Estonian online stores in electronics, groceries, pharmacy, DIY, sports, books, toys, furniture and fashion (full list at the end).
- Speed. Google Lighthouse 12.2 in mobile mode (simulated slower network and CPU), one run per home page, 14.11.2026, from Tallinn.
- SEO and GEO. The home page HTML without running JavaScript – which is how many AI crawlers see a page. We checked the title, meta description, H1, hreflang, schema.org JSON-LD, Open Graph, robots.txt, sitemap and llms.txt.
A single Lighthouse run varies by a few points, and lab results are not the same as real-user data (CrUX). That is why we look at medians and distributions rather than individual numbers.
How fast are Estonian online stores on mobile?
| Metric (mobile, Lighthouse) | Result across 47 stores |
|---|---|
| Median performance score | 61 / 100 (mean 59) |
| Score 90–100 (“good”) | 2 stores (4%) |
| Score 50–89 (“needs improvement”) | 30 stores (64%) |
| Score below 50 (“poor”) | 15 stores (32%) |
| Median LCP | 8.5 s |
| LCP up to 2.5 s | 3 stores (6%) |
| LCP over 4 s | 40 stores (85%) |
| Median First Contentful Paint (FCP) | 3.0 s |
| Visual stability CLS up to 0.1 | 34 stores (72%) |
| Median home page weight | 2.9 MB |
| Median server response (TTFB) | 0.24 s |
The key finding: servers respond fast (median TTFB under a quarter of a second), but the page itself is heavy on mobile. The slowness comes mainly from large images, banner sliders at the top of the home page, several tracking and marketing scripts and large JavaScript bundles. The best results in the sample were Prisma (99), Bauhof (90) and Euronics (86) – proof that a fast store is entirely possible in Estonia, even with a large catalogue.
Can Google and AI assistants understand Estonian online stores?
| Check (home page HTML without JavaScript) | Stores | Share |
|---|---|---|
| No schema.org JSON-LD at all | 19 | 40% |
| Organization markup present | 21 | 45% |
| BreadcrumbList on the home page | 7 | 15% |
| No H1 heading in the HTML | 22 | 47% |
| No meta description | 10 | 21% |
| Meta description over 160 characters | 10 | 21% |
| Canonical tag present | 32 | 68% |
| Open Graph (share preview) present | 36 | 77% |
| hreflang tags present | 21 | 45% |
| Russian version linked | 15 | 32% |
| English version linked | 14 | 30% |
| Sitemap listed in robots.txt | 44 | 94% |
| llms.txt file present | 8 | 17% |
| robots.txt blocks at least one AI crawler | 3 | 6% |
Three things stand out. First, almost half of the home pages have no H1 in the HTML before JavaScript runs. People will not notice, but many AI crawlers (for example GPTBot and ClaudeBot) do not run JavaScript and see an incomplete page. Second, 40% of stores do not say in machine-readable form who they are – there is no Organization, WebSite or other schema.org markup. Third, almost nobody blocks AI crawlers (only 3 stores), so the door to AI search is mostly open, but the pages are not prepared for it.
How multilingual are Estonian online stores?
People in Estonia search in Estonian, Russian and English, yet only a third of the stores linked a Russian version and fewer than half used hreflang tags. Without hreflang, Google may not show Russian-speaking searchers the right language version. We cover this Estonian specific in more detail in SEO in Estonia in 2026.
What does this mean for an online store owner?
- Mobile speed is the biggest opportunity. When 85% of stores have an LCP over 4 seconds, a fast home page is a clear advantage. Start with images (WebP/AVIF, correct sizes), the hero slider and third-party scripts.
- Add schema.org markup. Organization and WebSite on the home page, Product and Offer on product pages, BreadcrumbList in categories. This tells Google and AI assistants what you sell and to whom.
- Check what a crawler sees without JavaScript. The title, H1, main text and product descriptions must be in the HTML even if scripts do not run.
- Do language versions properly. hreflang tags, separate URLs and translated metadata in Estonian, Russian and English.
- Add llms.txt. It is not an official standard, but it is a simple way to give AI assistants an overview of your categories, delivery and payment. Only 17% of Estonian stores have one today.
How can IT Solutions Group help?
Since 2016 we have built WordPress and WooCommerce online stores and worked on speed optimisation, technical SEO and AI search visibility. We can run the same checks used in this study on your store and give you a prioritised fix plan. See our speed optimisation and AI search visibility services, or run a first check with our free website audit.
Further reading: Website speed and SEO and SEO vs AEO vs GEO.
Frequently asked questions
How fast are Estonian online stores on mobile?
In our 14.11.2026 test, the median mobile Lighthouse score of 47 Estonian store home pages was 61/100. Only 2 stores reached 90, and the median LCP was 8.5 seconds.
What is the most common SEO issue of Estonian online stores?
Missing structured data on the home page: 40% of the sample used no schema.org JSON-LD at all. A missing H1 in the HTML and a missing or overlong meta description are also common.
Do Estonian online stores block ChatGPT’s crawler?
Mostly not. Only 3 of 47 stores blocked at least one AI crawler in robots.txt, and one store blocked GPTBot.
How many Estonian online stores have an llms.txt file?
8 stores in the sample, or 17%. The file is not an official standard, but it gives AI assistants a quick overview of the store.
How accurate are the results?
A single Lighthouse lab run varies by a few points and does not replace real-user data. That is why we report medians and distributions, which reliably show the overall picture of the market.
Sample: 47 online stores
apollo.ee, apotheka.ee, arvutitark.ee, barbora.ee, bauhof.ee, benu.ee, bikestore.ee, decora.ee, demoog.ee, denimdream.com, ehitus24.ee, ehituseabc.ee, espak.ee, euroapteek.ee, euronics.ee, gamestar.ee, hansaplant.ee, huvikeskus.ee, jukukeskus.ee, kalev.eu, kaubamaja.ee, kaup24.ee, kidsland.ee, klick.ee, kodumasinad.ee, kriso.ee, lillepood.ee, lumi.ee, matkasport.ee, motonet.ee, nutiseade.ee, on24.ee, onoff.ee, oomipood.ee, petcity.ee, photopoint.ee, pood.telia.ee, prismamarket.ee, puumarket.ee, rahvaraamat.ee, rimi.ee, selver.ee, soulz.ee, sportland.ee, sudameapteek.ee, tradehouse.ee, veloplus.ee.
