Data: September 2026
Web Measurement Report Card: Türkiye
Türkiye's top 10,000 websites measure but rarely manage it: 41.1% use a tag manager and 14.2% a consent platform, the lowest of ten countries.
How Türkiye's 10,000 most visited websites measure, manage consent and act on data, compared with nine countries.
Personal research, built only from anonymous, aggregate, public data. No site is named. Sources are listed under Method.
Editions: September 2026
- Mean measurement maturity, out of 7
- 1.61 the lowest of the ten
- Sites with a tag manager
- 41.1% the lowest of the ten
- Sites with a consent platform
- 14.2% the lowest of the ten
- Google-tag sites with server-side traces
- 5.4% the lowest of the ten
- Sites with good responsiveness on phones
- 69.9% the lowest with Spain
Summary
Türkiye's top websites measure, but they rarely manage, govern or act on what they measure: 68.0% run web analytics, 41.1% a tag manager, 14.2% a consent platform and 19.3% of those with Google tags use Consent Mode.
Across ten countries, Türkiye scores lowest on measurement maturity, and the gap is widest in the layers that come after a tag is installed: management, consent, server-side ownership and activation. The gap sits in Turkish-built sites: global sites in Türkiye score like global sites everywhere, and the same tools cost Turkish sites more speed.
- Technology explains about a third. What a site runs explains at most 31.2% of the differences in phone responsiveness between top sites. Most of Türkiye's gap is not about what its sites run.
- Local platforms raise the floor. 31.9% of Turkish stores run on Turkish SaaS platforms, and 74.7% of those carry a tag manager.
- Customers are reached on WhatsApp. 14.0% of Turkish sites offer WhatsApp chat, the highest of the ten.
- Phones matter most here. 61.9% of human traffic from Türkiye is mobile, the highest of the ten, and phones are where Turkish sites are weakest.
- The gap opened recently. In 2021 Türkiye sat near the other nine; the distance to their median has since doubled.
The opportunity is less in adding tools than in running the ones already installed: managed, consented, owned and fast.
Measurement stack
68.0% of Türkiye's top websites run web analytics, yet only 41.1% use a tag manager, the lowest of the ten.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| Web analytics | 68.0% | 57.5% | 66.9% | 70.3% | 68.6% | 68.3% | 72.4% | 72.2% | 70.6% | 67.9% |
| Tag manager | 41.1% | 46.2% | 51.2% | 56.0% | 52.4% | 52.0% | 54.1% | 56.1% | 48.2% | 52.5% |
| No tag manager (of sites with analytics) | 43.2% | 37.9% | 33.8% | 28.9% | 31.1% | 31.7% | 32.1% | 28.3% | 37.8% | 27.9% |
| Behaviour analytics | 11.8% | 11.1% | 15.1% | 18.7% | 12.4% | 15.0% | 18.8% | 16.5% | 12.8% | 13.8% |
| Yandex Metrica | 8.1% | 3.1% | 1.6% | 1.3% | 1.0% | 1.0% | 2.7% | 1.9% | 1.7% | 1.3% |
| Privacy-first analytics | 1.8% | 9.0% | 14.2% | 4.2% | 8.3% | 4.8% | 7.9% | 3.1% | 2.8% | 2.8% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Hard-coded measurement. 43.2% of Turkish sites with web analytics have no tag manager, against 27.9–37.9% elsewhere. Their tags change only when the site's code does.
- The top tier is not immune. Among Türkiye's top 1,000 sites, 51.5% use a tag manager; in the rest of the top 10,000, 39.9%.
- A local second tool. Yandex Metrica runs on 8.1% of Turkish sites, the highest of the ten; elsewhere 1.0–3.1%.
- Little privacy-first analytics. Matomo, Plausible, Fathom and Piwik PRO together reach 1.8% of Turkish sites, the lowest of the ten.
- Dead code. 3.7% of Turkish sites with Google Analytics still load Universal Analytics, shut down in 2024, the highest with Poland.
Web analytics here is an explicit list of analytics products. Ad pixels and server monitoring tools, which Wappalyzer also files under analytics, are counted in other sections.
The crawl never accepts anything, so it cannot see tags that wait for consent. These shares are minimums. Other countries may measure more than they appear to here (see Privacy and consent).
Advertising and media
Turkish advertisers concentrate on two platforms and Turkish publishers sell the simplest way: 43.7% of ad-supported Turkish sites run AdSense alone, the highest of the ten.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| Meta Pixel | 20.6% | 11.8% | 14.8% | 21.9% | 17.6% | 18.7% | 19.9% | 21.4% | 29.8% | 22.6% |
| Google Ads conversion tag | 20.3% | 10.7% | 13.1% | 17.6% | 13.1% | 16.2% | 18.0% | 17.3% | 20.4% | 23.2% |
| TikTok Pixel | 7.8% | 3.9% | 5.0% | 7.7% | 5.0% | 6.1% | 5.9% | 7.0% | 13.7% | 9.4% |
| Microsoft Advertising | 4.3% | 18.9% | 20.0% | 22.3% | 21.6% | 16.2% | 25.9% | 18.5% | 12.8% | 17.7% |
| Criteo | 6.5% | 5.4% | 6.0% | 6.9% | 7.1% | 7.0% | 6.2% | 6.8% | 4.9% | 9.3% |
| X (Twitter) Ads | 3.0% | 1.4% | 1.7% | 4.9% | 1.6% | 2.6% | 1.8% | 1.8% | 1.2% | 5.3% |
| 1.9% | 1.9% | 2.5% | 4.0% | 2.2% | 2.3% | 4.6% | 1.8% | 1.6% | 5.4% | |
| 1.5% | 2.3% | 2.9% | 3.6% | 1.8% | 2.1% | 3.9% | 2.3% | 1.6% | 6.4% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| AdSense only, no ad server or auction (of ad-supported sites) | 43.7% | 16.2% | 28.5% | 14.1% | 27.0% | 21.3% | 25.4% | 31.0% | 37.5% | 9.4% |
| Header bidding (Prebid) (of ad-supported sites) | 23.7% | 43.6% | 30.7% | 56.5% | 35.9% | 37.9% | 40.1% | 38.4% | 29.8% | 53.5% |
| An identity solution (of ad-supported sites) | 26.6% | 35.1% | 41.3% | 51.3% | 31.0% | 37.5% | 35.0% | 33.3% | 28.3% | 61.9% |
| Amazon publisher services (of ad-supported sites) | 17.2% | 22.6% | 19.6% | 41.3% | 21.9% | 28.8% | 24.5% | 11.0% | 15.9% | 48.5% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Two-platform bets. Of Turkish sites with any ad pixel, 40.2% run both Meta and Google Ads, the highest with the US.
- Microsoft is nearly absent. Its tag reaches 4.3% of Turkish sites, against 12.8–25.9% elsewhere.
- TikTok follows Meta. 37.8% of Turkish sites with the Meta Pixel also run TikTok's.
- Pixel stacking is moderate. 26.1% of Turkish pixel sites run three or more ad platforms; in the US, 45.2%.
Ad pixels are detected technologies, except Google Ads, which is read from requests to its conversion endpoints.
Pixels that wait for consent are not seen either, so these shares are minimums. Do not read small gaps between countries as firm.
Privacy and consent
14.2% of Türkiye's top 10,000 websites run a consent platform, the lowest of the ten, against 27.4–40.3% in the other nine countries.
Share of sites on first load, nothing accepted.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| Consent platform | 14.2% | 35.5% | 36.1% | 40.3% | 40.1% | 33.9% | 35.6% | 29.1% | 27.4% | 29.4% |
| Consent Mode (of sites with Google tags) | 19.3% | 38.9% | 40.8% | 44.8% | 45.3% | 43.6% | 49.1% | 49.4% | 50.4% | 27.2% |
| No consent platform (of sites with an ad pixel) | 74.2% | 34.7% | 36.6% | 38.3% | 33.5% | 43.7% | 36.3% | 49.4% | 58.6% | 48.5% |
| Tracking starts anyway (of sites with a consent platform) | 66.6% | 45.5% | 61.2% | 66.4% | 63.8% | 65.1% | 73.8% | 68.7% | 70.4% | 79.1% |
| Google Analytics cookie | 56.1% | 32.5% | 37.3% | 47.6% | 44.1% | 44.3% | 49.9% | 44.8% | 46.2% | 55.6% |
| Any tracking cookie | 63.7% | 42.0% | 49.1% | 57.9% | 53.1% | 54.5% | 59.7% | 58.1% | 61.9% | 63.6% |
| Tag loaded, nothing sent | 3.9% | 17.4% | 13.9% | 10.3% | 9.5% | 9.0% | 8.7% | 7.2% | 6.3% | 6.6% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
| Consent platform | Sites in Türkiye | Share in Türkiye | Share in the other nine |
|---|---|---|---|
| Google Funding Choices (not counted) | 662 | 10.6% | 6.4% |
| OneTrust | 248 | 4.0% | 9.0% |
| Efilli | 200 | 3.2% | 0.0% |
| Cookie Seal | 50 | 0.8% | 0.0% |
| Cookiebot | 46 | 0.7% | 5.8% |
| consentmanager | 41 | 0.7% | 1.7% |
| Osano | 40 | 0.6% | 0.7% |
| Usercentrics | 30 | 0.5% | 1.8% |
| IdeaSoft cookie bar | 29 | 0.5% | 0.0% |
| InMobi Choice | 27 | 0.4% | 1.5% |
| Cookie Script | 27 | 0.4% | 1.0% |
| CookieYes | 25 | 0.4% | 1.8% |
- Tracking starts anyway. On 66.6% of Turkish sites with a consent platform, tracking starts before anything is accepted, against 45.5–79.1% elsewhere. On Turkish sites with no platform it is 64.0%, so the two barely differ. Germany is lowest at 45.5%.
- A local player. OneTrust is the most used platform in Türkiye, on 248 of its 889 sites with one. Efilli follows on 200, and reaches 8 sites in the other nine countries together.
- Google's tool is counted apart. 9.4% of Turkish sites load only Google Funding Choices. It is a message tool for ad publishers and may show this visitor no banner, so it does not count as a consent platform here.
- Tag management predicts consent. Among Turkish sites with a tag manager, 27.2% run a consent platform; without one, 5.0%.
- Tracking before consent. 63.7% of Turkish sites set a tracking cookie on first load, the highest with the US.
- Consent state at load. Of Turkish sites sending Google tag requests, 10.6% load with ad storage denied, inside the 5.7–29.9% range, and 9.3% with it granted. The rest send no consent state.
- IAB frameworks. TCF v2 appears on 9.2% of Turkish sites (others 4.0–17.2%), GPP on 3.0%.
What a first load shows, not a legal assessment. HTTP Archive crawls from the US, so sites that show a banner only to EU visitors look less compliant than they are for those visitors. Tracking cookies are counted by known names, so the true shares are likely higher.
That trace is not proof: a tag can also be empty or use a tool not tracked here. A consent platform is a tool that collects consent and passes it to the tags, read from detected technologies and from the hosts of known tools. Hand-built banners are not counted in any country, and how many exist is not measured. The crawl is one visit from the US: a site may hold its tags for visitors elsewhere, so this does not say what any site does for a visitor in Türkiye.
Server-side and first-party
Server-side tagging is where measurement is moving, and 5.4% of Turkish top sites with Google tags show traces of it, the lowest of the ten, against 21.7% in the Netherlands.
The data
| the Netherlands | 21.7% |
|---|---|
| Germany | 16.5% |
| France | 14.0% |
| Italy | 11.5% |
| the UK | 10.6% |
| Spain | 9.9% |
| the US | 8.7% |
| Poland | 7.3% |
| Romania | 7.1% |
| Türkiye | 5.4% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| Endpoint on a default cloud domain (of sites with a first-party endpoint) | 11.7% | 2.2% | 2.6% | 4.3% | 4.9% | 6.5% | 2.6% | 5.7% | 7.3% | 6.7% |
| GA4 also sent straight to Google (of sites sending GA4 to their own endpoint) | 40.7% | 23.0% | 23.5% | 35.0% | 29.9% | 34.3% | 24.7% | 28.5% | 29.0% | 39.5% |
| Runs a consent platform (of server-side sites) | 50.4% | 64.5% | 69.0% | 67.4% | 73.3% | 69.4% | 65.5% | 63.6% | 53.1% | 60.4% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Often on a borrowed domain. 11.7% of Turkish sites with a first-party endpoint run it on a default cloud domain such as appspot.com or run.app, the highest with Romania. The data goes through a server, but the cookie advantage of a first-party domain is lost.
- Often half migrated. 40.7% of Turkish sites that send GA4 hits to their own endpoint still send them straight to Google as well, the highest with the US. A second, direct stream can be deliberate, but it often marks a migration left unfinished, with two sets of numbers to reconcile.
- Server-side, often without consent tooling. 50.4% of Turkish server-side sites run a consent platform, against 53.1–73.3% elsewhere. Moving tags to a server does not remove the need for consent.
- The server-managed client ID is rare. The FPID cookie, which server-side GTM sets itself so the visitor ID survives browser limits, appears on 1.8% of Turkish Google-tag sites, the lowest of the ten.
- The top tier leads. Among Türkiye's top 1,000 sites with Google tags, 9.2% show server-side traces; in the rest of the top 10,000, 4.9%.
A site counts as server-side when a Google tag loads from, or GA4 hits go to, a host that is not Google's, or when the FPID or FPLC cookie is set. Conversion APIs that run server to server, such as Meta CAPI, leave no trace in the browser, so the true share is higher.
Marketing technology stack
12.3% of Türkiye's top sites run a customer data platform, an audience tool, A/B testing or personalisation, the lowest with Romania: the country collects data but rarely acts on it on the site.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| Marketing automation | 8.8% | 16.4% | 21.7% | 26.9% | 18.7% | 20.5% | 23.1% | 13.3% | 14.7% | 25.7% |
| Customer data platform | 3.9% | 5.4% | 6.3% | 9.7% | 5.3% | 6.0% | 6.9% | 5.4% | 2.6% | 14.4% |
| Audience segmentation (DMP) | 4.7% | 7.5% | 7.4% | 12.1% | 7.1% | 8.9% | 6.7% | 4.8% | 5.4% | 18.2% |
| A/B testing | 4.5% | 9.1% | 10.9% | 13.0% | 8.0% | 9.2% | 10.9% | 4.8% | 6.3% | 13.2% |
| Personalisation | 2.0% | 5.6% | 5.5% | 11.4% | 5.6% | 8.1% | 5.3% | 2.4% | 3.5% | 14.0% |
| Reviews widget | 1.2% | 8.4% | 9.9% | 13.3% | 6.5% | 6.4% | 7.9% | 4.2% | 4.9% | 5.2% |
| Live chat, any | 18.5% | 11.3% | 14.1% | 17.0% | 16.0% | 16.7% | 15.6% | 9.0% | 18.6% | 15.6% |
| WhatsApp chat | 14.0% | 2.5% | 1.4% | 1.5% | 5.7% | 6.8% | 5.5% | 1.2% | 11.3% | 0.5% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Automation is the widest gap. 8.8% of Turkish sites run a marketing automation tool, the lowest of the ten. Self-serve tools such as Klaviyo, HubSpot and Mailchimp reach 4.3% (UK 17.8%).
- The conversation moved to WhatsApp. 14.0% of Turkish sites offer WhatsApp chat, the highest of the ten, and 75.3% of Turkish sites with any live chat use it (UK 8.6%, US 3.2%).
- A local platform leads locally. Insider, founded in Türkiye, runs on 1.9% of Turkish sites and at most 0.3% elsewhere.
- Advanced tools sit in the top tier. 21.7% of Türkiye's top 1,000 sites run them; in the rest of the top 10,000, 11.2%.
- Acting on data, without consent tooling. 60.9% of Turkish sites with advanced tools run no consent platform, the highest of the ten.
- Few public reviews. Review widgets appear on 1.2% of Turkish sites, the lowest of the ten.
Categories are Wappalyzer's, detected on the home page. Tools that run only on the server, inside an app or behind a login are not seen; a WhatsApp button may be a simple link rather than a managed chat.
Performance
On phones, 57.5% of Turkish top sites pass Core Web Vitals, inside the 56.6–72.6% range. Good responsiveness (INP) stands at 69.9%, the lowest with Spain, and good layout stability (CLS) at 79.4%, the lowest of the ten.
Share of top sites with good field performance at p75.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| Core Web Vitals pass, phone | 57.5% | 71.1% | 63.1% | 66.8% | 57.1% | 56.6% | 72.6% | 62.0% | 62.9% | 58.3% |
| Core Web Vitals pass, desktop | 71.1% | 79.3% | 75.0% | 71.9% | 71.8% | 72.1% | 77.7% | 74.5% | 74.3% | 64.3% |
| Good LCP (loading), phone | 79.7% | 88.0% | 84.7% | 85.7% | 78.2% | 79.2% | 88.9% | 85.1% | 83.0% | 76.6% |
| Good INP (responsiveness), phone | 69.9% | 84.2% | 76.5% | 81.0% | 74.1% | 71.1% | 86.4% | 76.0% | 79.8% | 78.1% |
| Good CLS (layout stability), phone | 79.4% | 87.2% | 86.7% | 86.2% | 85.2% | 86.3% | 87.2% | 85.8% | 85.7% | 82.3% |
Chrome UX Report on phones and HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
Core Web Vitals pass rate on phones, by marketing tools on the home page.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| 0 to 2 tools | 66.9% | 78.0% | 70.6% | 75.3% | 66.8% | 66.2% | 76.4% | 67.5% | 70.1% | 64.9% |
| 3 to 5 tools | 60.2% | 69.3% | 61.0% | 68.7% | 58.6% | 54.4% | 74.9% | 64.3% | 62.5% | 58.8% |
| 6 or more tools | 45.8% | 62.5% | 55.8% | 59.2% | 43.7% | 46.5% | 69.4% | 57.3% | 51.2% | 56.0% |
Chrome UX Report on phones and HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Phones lag desktop. 71.1% of Turkish sites pass on desktop, 13.6 points more than on phones.
- Each marketing tool band costs more here. Turkish sites with six or more marketing tools pass 21.1 points less often than those with two or fewer; in the Netherlands, 7.0 points.
- Turkish sites do not carry more tools. The median Turkish home page runs 2 marketing tools and calls 9 third-party hosts.
- The top tier is no faster. Among Türkiye's top 1,000 sites, 56.1% pass on phones; in the rest of the top 10,000, 57.6%.
Traffic context
The phone results weigh more in Türkiye than anywhere else in the set. In Cloudflare's traffic, 61.9% of likely-human page requests from Türkiye come from mobile devices, the highest of the ten; the others sit at 46.2–54.0%.
The data
| Türkiye | 61.9% |
|---|---|
| Romania | 54.0% |
| Italy | 53.8% |
| the UK | 52.4% |
| Spain | 51.5% |
| France | 48.9% |
| Poland | 46.8% |
| the US | 46.8% |
| Germany | 46.3% |
| the Netherlands | 46.2% |
Cloudflare Radar, likely-human traffic by visitor country · September 2026
An origin passes when its p75 LCP is at most 2.5 s, INP at most 200 ms and CLS at most 0.1, as measured by real Chrome users. Sites with more tools also tend to be larger and more complex, so the tool bands show an association, not a measured cost per tag. Radar measures traffic by the visitor's country and only through Cloudflare's network.
E-commerce and app signals
31.9% of Turkish stores among the top sites run on a Turkish SaaS platform, and those stores carry a tag manager more often than any other Turkish group.
Share of stores among the top 10,000 sites per country.
Turkish SaaSShopifyOpen sourceEnterprise suiteCustom or not detected
The data
| Turkish SaaS | Shopify | Open source | Enterprise suite | Custom or not detected | |
|---|---|---|---|---|---|
| TR | 31.9% | 7.9% | 4.6% | 2.0% | 53.7% |
| DE | 0.0% | 17.0% | 13.3% | 4.9% | 64.9% |
| FR | 0.0% | 13.6% | 23.8% | 6.6% | 56.0% |
| GB | 0.0% | 23.1% | 13.4% | 7.7% | 55.8% |
| IT | 0.0% | 15.2% | 19.6% | 6.3% | 58.9% |
| ES | 0.0% | 17.3% | 26.2% | 6.4% | 50.1% |
| NL | 0.0% | 17.3% | 23.3% | 4.5% | 54.9% |
| PL | 0.0% | 8.8% | 30.4% | 2.6% | 58.2% |
| RO | 0.0% | 14.3% | 30.7% | 1.3% | 53.7% |
| US | 0.0% | 16.4% | 4.5% | 10.7% | 68.5% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Four local platforms. Ticimax runs 14.0% of Turkish stores, T-Soft 10.2%, IdeaSoft 5.3% and ikas 2.5%. Shopify holds 7.9%, the lowest with Poland, and open-source carts such as WooCommerce and Magento 4.6%.
- The platform sets the measurement. 74.7% of stores on Turkish SaaS run a tag manager, against 41.1% of Turkish top sites overall and 34.0% of Turkish Shopify stores. A platform that ships tag management to every merchant lifts the floor for all of them.
- Stores are pixel-heavy. 50.6% of Turkish stores run the Meta Pixel, inside the 25.5–55.6% range.
- App links concentrate at the top. 24.3% of Turkish top sites declare Android or iOS app links, inside the 21.9–32.7% range. Among the top 1,000 the share is 42.3%; in the rest of the top 10,000, 22.2%.
- Sites with an app measure more. 50.6% of Turkish sites with app links run a tag manager; without, 38.5%.
Platform by platform
"Turkish SaaS" is not one thing. 71.8% of IdeaSoft stores and 70.1% of T-Soft stores pass Core Web Vitals on phones, around Shopify's 69.2%, while 36.6% of Ticimax stores and 33.3% of ikas stores do.
Stores passing Core Web Vitals on phones. Lines are 95% intervals; n is stores with field data.
Dashed line: all Turkish stores: 58.5%
The data
| 95% interval | Stores | ||
|---|---|---|---|
| IdeaSoft | 71.8% | 60.5% – 81.0% | 71 |
| T-Soft | 70.1% | 61.9% – 77.2% | 134 |
| Shopify | 69.2% | 59.8% – 77.3% | 104 |
| Other or custom | 60.6% | 57.1% – 64.1% | 752 |
| WooCommerce | 55.6% | 39.6% – 70.5% | 36 |
| Ticimax | 36.6% | 30.0% – 43.8% | 183 |
| ikas | 33.3% | 19.2% – 51.2% | 30 |
Chrome UX Report on phones and HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- The gap holds for comparable stores. Against other Turkish stores of the same rank tier and marketing tool count, T-Soft stores pass 15.0 points more often and IdeaSoft stores 8.8 points more; Ticimax stores 19.4 points less and ikas stores 29.6 points less.
- Responsiveness separates them. Good INP: IdeaSoft 85.9%, T-Soft 79.9%, Ticimax 50.3%, ikas 33.3%.
- Ticimax stores carry more. The median Ticimax store runs 7 marketing tools and loads from 23 hosts; the median T-Soft store, 5 and 19.
- ikas is a small sample. 30 stores have field data, so the direction is clear and the exact level is not.
| Platform | Stores | Tag manager | Consent platform | Consent Mode | Meta Pixel |
|---|---|---|---|---|---|
| Ticimax | 184 | 86.4% | 15.2% | 32.1% | 92.9% |
| T-Soft | 134 | 74.6% | 14.9% | 23.9% | 84.3% |
| Shopify | 104 | 33.7% | 8.7% | 80.8% | 80.8% |
| IdeaSoft | 71 | 71.8% | 42.3% | 2.8% | 56.3% |
| WooCommerce | 36 | 52.8% | 16.7% | 13.9% | 30.6% |
| ikas | 30 | 30.0% | 13.3% | 6.7% | 93.3% |
| Other or custom | 757 | 61.6% | 23.6% | 23.5% | 29.9% |
A store is a home page with a detected cart or e-commerce platform. Turkish SaaS platforms are not in Wappalyzer's list, so they are read from the asset hosts each page loads. A platform's numbers mix the platform with its merchants' themes, apps and content, so they describe the stores on each platform, not the platform alone. App links count when a site's assetlinks.json or apple-app-site-association file declares deep links.
Technical SEO and AI readiness
16.4% of Turkish top sites block AI training crawlers, inside the 16.3–19.6% range, but only 31.7% of those also block AI search agents, against 40.5–53.4% in Western Europe and the US.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| robots.txt present | 84.3% | 91.6% | 89.4% | 88.8% | 87.0% | 87.6% | 92.9% | 89.3% | 90.4% | 83.4% |
| Sitemap declared in robots.txt | 50.5% | 58.9% | 54.0% | 57.8% | 54.9% | 54.9% | 62.1% | 52.1% | 60.1% | 50.2% |
| Blocks AI training crawlers | 16.4% | 18.8% | 16.3% | 19.4% | 16.3% | 16.6% | 18.3% | 18.2% | 19.6% | 18.8% |
| Blocks Google-Extended | 12.9% | 11.3% | 10.2% | 10.9% | 9.8% | 9.6% | 11.2% | 9.5% | 9.3% | 12.0% |
| Blocks AI search and user agents | 5.2% | 9.3% | 7.1% | 9.8% | 6.6% | 7.8% | 7.6% | 5.1% | 4.5% | 10.0% |
| Also blocks AI search (of sites blocking AI training) | 31.7% | 49.5% | 43.7% | 50.5% | 40.5% | 47.0% | 41.5% | 28.2% | 22.8% | 53.4% |
| Blocks AI training (of ad-supported sites) | 23.0% | 28.8% | 30.5% | 41.5% | 29.3% | 30.7% | 31.8% | 28.8% | 23.5% | 38.2% |
| Blocks AI training (of sites not on Cloudflare) | 9.7% | 17.2% | 12.2% | 13.7% | 12.4% | 12.9% | 16.0% | 13.7% | 17.4% | 11.5% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Open to AI answers. Agents such as OAI-SearchBot, ChatGPT-User and PerplexityBot fetch a page when someone asks an assistant about it. 5.2% of Turkish sites block them (US 10.0%), so Turkish content stays quotable in AI answers even where it is closed to training.
- Google-Extended. 12.9% of Turkish sites block it, the highest with the US. It stops Gemini training without touching Google Search.
- Cloudflare does much of the blocking. Turkish sites on Cloudflare block AI training at 20.1%; elsewhere, 9.7%. A setting at the CDN, more than a site decision, seems to drive the Turkish rate.
- Publishers block less. 23.0% of ad-supported Turkish sites block AI training, the lowest with Romania (UK 41.5%).
- Basic hygiene lags. 84.3% of Turkish sites serve a robots.txt, inside the 83.4–92.9% range, and 50.5% declare a sitemap in it, inside the 50.2–62.1% range.
HTTP Archive records how many rules a robots.txt sets for each crawler, not their paths, so a block here is any Disallow rule for that crawler, whole site or part. Training crawlers: GPTBot, ClaudeBot, anthropic-ai, CCBot, Google-Extended, Applebot-Extended, Meta-ExternalAgent, Bytespider. Search and user agents: OAI-SearchBot, ChatGPT-User, PerplexityBot, Perplexity-User, Claude-SearchBot, Claude-User.
Infrastructure
The Turkish web runs through one vendor: 63.7% of Turkish top sites sit behind Cloudflare, and 92.9% of those with an edge CDN use it, the highest of the ten.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| Cloudflare | 63.7% | 48.8% | 54.2% | 64.4% | 52.6% | 53.7% | 59.1% | 52.3% | 55.1% | 60.2% |
| CloudFront, Akamai or Fastly | 7.0% | 12.5% | 13.0% | 19.7% | 14.0% | 16.1% | 13.2% | 8.2% | 7.5% | 22.7% |
| HTTP/3 | 43.5% | 31.6% | 34.1% | 35.4% | 33.0% | 32.5% | 37.3% | 35.1% | 47.7% | 28.8% |
| HSTS (HTTPS enforced) | 50.5% | 58.2% | 56.5% | 63.4% | 49.8% | 52.6% | 60.5% | 46.3% | 46.1% | 62.3% |
| AWS, Google Cloud or Azure | 15.9% | 26.1% | 25.0% | 44.2% | 28.2% | 31.3% | 29.9% | 17.5% | 24.5% | 46.6% |
| ASP.NET or IIS | 18.2% | 3.0% | 2.6% | 8.8% | 6.5% | 5.0% | 5.5% | 3.0% | 3.3% | 10.7% |
| React, Vue, Next.js or Angular | 24.7% | 30.9% | 28.1% | 32.4% | 24.8% | 27.5% | 32.4% | 25.2% | 21.2% | 35.5% |
| jQuery | 61.0% | 51.8% | 55.9% | 52.7% | 59.9% | 57.6% | 50.5% | 57.6% | 63.7% | 51.2% |
| Code from public library CDNs | 22.9% | 13.6% | 19.6% | 22.0% | 20.8% | 19.5% | 17.9% | 17.3% | 18.9% | 19.5% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Modern protocol by default. 43.5% of Turkish sites serve HTTP/3, inside the 28.8–47.7% range. 88.6% of them are on Cloudflare: HTTP/3 arrives with the CDN, not as a separate choice.
- Little global CDN or cloud. CloudFront, Akamai and Fastly together serve 7.0% of Turkish sites, the lowest with Romania. AWS, Google Cloud and Azure host 15.9%, the lowest of the ten.
- A Microsoft-heavy stack. ASP.NET or IIS runs 18.2% of Turkish sites, the highest of the ten; elsewhere 2.6–10.7%.
- Older front ends. jQuery appears on 61.0% of Turkish sites and React, Vue, Next.js or Angular on 24.7%. 22.9% load code from public library CDNs such as cdnjs and jsDelivr at runtime, the highest with the UK, which adds third-party code outside the site's control.
- HTTPS enforcement. 50.5% of Turkish sites send HSTS, inside the 46.1–63.4% range; the top 1,000 reach 54.7%.
Outdated and vulnerable versions
Where a site exposes the version of its software, the version can be checked against known vulnerabilities and end-of-life dates. 29.0% of Turkish top sites run at least one flagged version, inside the 21.2–33.6% range. The exception is Next.js.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| Next.js with the 2025 middleware bypass (of sites with a readable version) | 20.0% | 9.9% | 10.3% | 9.0% | 12.6% | 10.4% | 7.7% | 15.9% | 14.3% | 7.2% |
| jQuery UI before 1.13 (XSS) (of sites with a readable version) | 55.3% | 41.9% | 56.7% | 41.3% | 46.4% | 55.8% | 44.8% | 53.9% | 51.6% | 46.4% |
| jQuery before 3.5 (XSS) (of sites with a readable version) | 38.7% | 35.8% | 39.3% | 34.2% | 36.9% | 41.7% | 37.5% | 43.4% | 44.8% | 35.1% |
| PHP past end of life (before 8.2) (of PHP sites showing their version) | 56.5% | 53.6% | 62.5% | 57.9% | 54.6% | 61.0% | 49.5% | 60.6% | 57.6% | 58.3% |
| IIS on Windows Server 2012 R2 or older (of sites with a readable version) | 15.4% | 6.0% | 10.1% | 6.0% | 17.8% | 17.4% | 5.2% | 15.0% | 19.4% | 4.8% |
| At least one of the above (of all top sites) | 29.0% | 21.2% | 28.3% | 21.3% | 27.3% | 30.0% | 22.2% | 30.3% | 33.6% | 21.7% |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Next.js left unpatched. 20.0% of Turkish Next.js sites run a version affected by the middleware authorisation bypass disclosed in March 2025 (CVE-2025-29927), the highest with Poland.
- Old Windows servers. Türkiye has the most IIS sites in the set (416 with a readable version), and 15.4% of them run IIS older than version 10, which means Windows Server 2012 R2 or older, out of support since October 2023.
- jQuery is catching up. 38.7% of Turkish jQuery sites still load a version before 3.5, with known XSS flaws, down from 76.0% in 2021.
- PHP past end of life is the norm. Of PHP sites that show their version, 56.5% in Türkiye run a version before 8.2, inside the 49.5–62.5% range. 20.9% of Turkish PHP sites put the version in their response headers, which is a disclosure in itself.
Only sites that expose a version are judged, so each share is over that group, and a version read from a page can lag what the server runs. A flagged version is a risk, not proof that a site can be attacked: some vulnerabilities need a specific feature to be in use, and some vendors back-port fixes. The report keeps aggregates only; it names no site and keeps no list of sites. HTTP/3 and HSTS are read from home page response headers.
Measurement maturity index
On a seven-signal maturity index, Türkiye averages 1.61 of 7, the lowest of the ten, against 2.03–2.60 elsewhere; 4.5% of its top sites reach the advanced level.
One score per site brings the sections together. A site earns a point for each of seven signals on its home page: web analytics and a tag manager (collect); a consent platform and Google Consent Mode (govern); a server-side trace (own); a CDP, DMP, A/B testing or personalisation tool, and marketing automation (activate). Levels: none (0), basic (1–2), managed (3–4), advanced (5–7).
None (0)Basic (1–2)Managed (3–4)Advanced (5–7)
The data
| None (0) | Basic (1–2) | Managed (3–4) | Advanced (5–7) | Mean | |
|---|---|---|---|---|---|
| GB | 17.7% | 32.6% | 29.9% | 19.8% | 2.60 |
| NL | 17.8% | 34.3% | 28.0% | 19.9% | 2.55 |
| IT | 20.3% | 34.9% | 30.6% | 14.2% | 2.35 |
| ES | 21.4% | 34.1% | 30.1% | 14.4% | 2.32 |
| US | 22.4% | 34.0% | 27.9% | 15.6% | 2.31 |
| FR | 21.0% | 36.9% | 26.7% | 15.4% | 2.28 |
| PL | 19.8% | 36.6% | 33.9% | 9.8% | 2.23 |
| RO | 20.1% | 38.7% | 32.4% | 8.8% | 2.13 |
| DE | 23.8% | 39.7% | 25.8% | 10.8% | 2.03 |
| TR | 26.2% | 50.3% | 19.0% | 4.5% | 1.61 |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Stuck at basic. 50.3% of Turkish sites score 1 or 2, the highest of the ten: analytics installed, little around it.
- The top tier pulls ahead. Türkiye's top 1,000 sites average 1.99, the rest 1.57: the widest gap of the ten.
- Maturity costs speed in Türkiye. Turkish sites at the advanced level pass Core Web Vitals on phones at 37.1%, against 64.8% at none or basic, a drop of 27.7 points. The Netherlands loses 3.3 points, which shows a full stack does not have to be slow.
Türkiye against the nine other countries; the line is their range.
The data
| TR | DE | FR | GB | IT | ES | NL | PL | RO | US | |
|---|---|---|---|---|---|---|---|---|---|---|
| None or basic | 64.8% | 75.7% | 68.6% | 73.5% | 65.4% | 63.9% | 75.0% | 66.1% | 68.1% | 63.6% |
| Managed | 48.1% | 68.6% | 59.2% | 65.2% | 52.9% | 51.9% | 72.7% | 63.7% | 58.7% | 56.9% |
| Advanced | 37.1% | 61.3% | 55.2% | 56.6% | 41.8% | 42.0% | 71.7% | 51.6% | 49.7% | 53.3% |
Chrome UX Report on phones and HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
The index counts what a crawler sees on first load. Tags held back until consent and server-to-server data are not seen, so sites that do these well can score lower than they are. The signals are equally weighted; the index ranks breadth of the stack, not how well each part is used.
Cross readings and robustness
Türkiye's gap is in its own sites: global sites in Türkiye score 1.81, in line with global sites elsewhere, while sites that appear only on Türkiye's list score 1.56, against 2.08–2.94 for local sites in the other countries.
Before reading the sections as a verdict on Türkiye, three checks: whether the gap comes from global sites that appear in every list, whether it comes from the mix of site types, and where it shows inside each type. Türkiye's list is the most local of the ten: 77.3% of its sites are in no other list. Elsewhere local sites score higher than global ones; in Türkiye they score lower.
Local: in this country's top 10,000 only. Global: in two or more lists.
Local sitesGlobal sites
The data
| Local sites | Global sites | |
|---|---|---|
| NL | 2.94 | 1.92 |
| GB | 2.86 | 2.15 |
| IT | 2.44 | 2.11 |
| FR | 2.43 | 1.93 |
| US | 2.40 | 2.14 |
| ES | 2.37 | 2.22 |
| PL | 2.36 | 1.88 |
| RO | 2.25 | 1.80 |
| DE | 2.08 | 1.94 |
| TR | 1.56 | 1.81 |
HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- Consent follows the same split. 11.9% of Türkiye's local sites run a consent platform, against 29.3–47.2% of local sites elsewhere; its global sites reach 21.8%, against 22.2–31.2%.
- Pixels without consent. 78.3% of Turkish local sites with an ad pixel have no consent platform, against 30.4–60.4% elsewhere.
Not a matter of site mix
Türkiye's top list holds more public sites: government and education domains are 12.6% of it, against 2.0% across the other nine. Those sites score low everywhere. Reweighting every country to the same mix of site types lifts Türkiye's mean from 1.61 to 1.75, still the lowest. Within each commercial type it is also the lowest: stores 2.41 (others 2.78–3.81), publishers 2.50 (2.96–3.67), other sites 1.28 (1.46–1.88).
- Publishers are low too. 18.6% of Turkish ad-supported sites run a consent platform, the lowest of the ten.
- Stores run pixels without consent tooling. 21.2% of Turkish stores run a consent platform, the lowest of the ten, and 77.9% of those with an ad pixel have none.
- Platforms lift measurement, not consent. Of stores on Turkish SaaS, 74.7% run a tag manager but 19.2% run a consent platform.
What explains website speed
Technology explains about a third of the speed differences between top websites; pixels are the clearest measurable cost, and most of Türkiye's gap is not about what its sites run.
How much can technology explain?
What a site runs explains 31.2% of the differences in phone responsiveness (INP) between top sites and 35.5% of the differences in loading (LCP), at best.
This covers 77,314 site and country pairs with field data on phones. Most of the difference between sites sits outside what a crawl can see: the devices and networks of their visitors, their content, and how their code is written.
| Share of differences explained | Country and rank only | Adding technology | Flexible model | Türkiye alone |
|---|---|---|---|---|
| Responsiveness (INP) | 2.7% | 22.8% | 31.2% | 25.0% |
| Loading (LCP) | 2.3% | 26.0% | 35.5% | 24.3% |
| Core Web Vitals pass | 1.3% | 11.5% | 19.0% | 15.5% |
Share of what technology adds to explaining INP. Shapley split: each group's average gain over every order of adding the groups.
The data
| Front-end technology | 28.4% |
|---|---|
| Marketing and measurement tools | 28.2% |
| Page weight | 21.9% |
| CDN and delivery | 9.2% |
| Platform and server | 7.7% |
| Ad technology | 3.2% |
| Site type | 1.5% |
Chrome UX Report on phones and HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
- No single cause. For responsiveness, front-end technology carries 28% of the explained part, marketing and measurement tools 28% and page weight 22%.
- Loading is about weight. For LCP, page weight leads at 29% and marketing tools fall to 18%.
- The pass rate is harder to move. Technology explains 19.0% of who passes Core Web Vitals, and front-end technology is 40% of that.
Is Türkiye's gap about what sites run?
Mostly not: Turkish sites pass Core Web Vitals on phones 13.2 points less often than Dutch ones, and only 2.6 points of that comes from what they run.
| Türkiye against | Pass rate gap, points | From what sites run | Not explained by it |
|---|---|---|---|
| the Netherlands | −13.2 | −2.6 | −10.5 |
| Germany | −11.9 | −3.3 | −8.5 |
| the UK | −7.2 | +2.1 | −9.4 |
| France | −3.5 | +0.1 | −3.5 |
| Poland | −3.3 | −1.2 | −2.1 |
| Romania | −2.8 | −0.7 | −2.0 |
| the US | +0.5 | +4.7 | −4.2 |
| Italy | +2.2 | +2.9 | −0.7 |
| Spain | +3.2 | +4.1 | −0.9 |
- Same stack, different result. Against Germany the gap is 11.9 points, of which 3.3 points comes from the stack. Against the UK the stack even favours Türkiye (+2.1), yet Türkiye is 7.2 points behind.
- What the rest is, the data cannot say. The unexplained part is not automatically poor implementation. It also holds the phones and networks of each country's visitors, which a crawl does not see. Field data is measured on those phones, so a country with slower devices looks slower with the same site.
Oaxaca-Blinder decomposition with pooled coefficients, on the Core Web Vitals pass. A negative gap means Türkiye is behind. A positive figure under "from what sites run" means the stack favours Türkiye; it does so against 5 of the nine.
What happens when a site adds a tool
After adding the Meta Pixel, INP rose 7.2% more than at similar sites, and 14.7% more in Türkiye; after the TikTok Pixel, 11.1% and 16.7%.
A snapshot shows which sites choose a tool, not what the tool does. So this follows sites across September 2024, 2025 and 2026, and compares those that added a tool with similar sites that never had it, in the same country and rank tier. A second test looks at the year before they added it: if adders were already moving that way, the change is not the tool's.
Percent change in p75 INP against similar sites that never had the tool, 2024 to 2026. Higher is slower.
TürkiyeAll ten countries
The data
| Türkiye | All ten countries | |
|---|---|---|
| Insider | +23.3 | +16.2 |
| TikTok Pixel | +16.7 | +11.1 |
| Meta Pixel | +14.7 | +7.2 |
| Google Tag Manager | +4.4 | +1.9 |
Chrome UX Report on phones and HTTP Archive mobile crawl, top 10,000 sites per country · September 2026
| Tool | Sites that added it | INP change | In Türkiye | Year before | Year after | Test |
|---|---|---|---|---|---|---|
| Insider | 113 | +16.2% | +23.3% | – | – | Too few sites |
| TikTok Pixel | 3035 | +11.1% | +16.7% | +4.4% | +15.8% | Holds |
| Meta Pixel | 4219 | +7.2% | +14.7% | +2.2% | +12.1% | Holds |
| Yandex Metrica | 327 | +4.8% | +9.2% | +0.1% | +8.5% | Holds |
| Hotjar | 1388 | +3.6% | – | +0.8% | +9.1% | Holds |
| Google Analytics | 5410 | +3.1% | – | +0.1% | +8.0% | Holds |
| Google Tag Manager | 5420 | +1.9% | +4.4% | +1.7% | +6.8% | Holds |
| Cloudflare | 9361 | +1.7% | – | +0.3% | +3.2% | Holds |
| HTTP/3 | 7709 | +0.8% | – | +0.9% | +2.9% | Holds |
| Microsoft Clarity | 5216 | +5.1% | +10.4% | +2.1% | +2.9% | Does not hold |
| WhatsApp chat | 1070 | +3.5% | – | +3.9% | +4.7% | Does not hold |
| Criteo | 2831 | +0.0% | +5.3% | −3.7% | +5.3% | Does not hold |
| OneTrust | 1317 | −7.7% | – | −7.5% | −5.7% | Does not hold |
- Pixels are the clearest cost. Meta's and TikTok's pixels hold up: little change the year before, a clear rise after. TikTok is also the one tool that clearly lowers the pass rate, by 2.4 points.
- A tag manager is cheap, not free. Adding Google Tag Manager raised INP by 1.9%, well below the pixels.
- Two snapshot results did not survive. OneTrust looks costly in a snapshot and even seems to help here, but its adders were already improving the year before. Microsoft Clarity's adders were already slowing. Neither change can be credited to the tool.
INP change is the percentage change in p75 INP against sites that never had the tool, 2024 to 2026, within country and rank tier; a dash in the Türkiye column means no clear result. A tool holds when the step after adding it is clear and at least twice the step before. Year before and year after use sites seen in all three Septembers that added the tool in the last year, so they differ from the main column. This is closer to cause than a snapshot, but it is not an experiment: a site adding a pixel may be changing other things too, and detection can miss a tool in one crawl. Why a tool costs more in Türkiye the data cannot say: the way it is installed and the devices and networks of Turkish visitors both fit. An earlier version of this report used a snapshot comparison and showed OneTrust and Microsoft Clarity among the costliest tools; this test did not support that.
2021 to 2026
The gap is recent: on a five-signal version of the index Türkiye was 0.24 below the median of the other nine in 2021 and is 0.55 below it in 2026.
Türkiye, the nine others.
Maturity score, 5 signals
20211.68 → 1.382026
Sites with a tag manager
202161.7% → 41.1%2026
Sites with marketing automation
20219.3% → 8.8%2026
Core Web Vitals pass on phones
202226.8% → 57.5%2026
The data
| 2021 | 2022 | 2023 | 2024 | 2025 | 2026 | |
|---|---|---|---|---|---|---|
| Maturity score, 5 signals, Türkiye | 1.68 | 1.72 | 1.49 | 1.48 | 1.48 | 1.38 |
| Sites with a tag manager, Türkiye | 61.7% | 63.2% | 41.7% | 43.7% | 44.0% | 41.1% |
| Sites with marketing automation, Türkiye | 9.3% | 8.7% | 9.5% | 12.4% | 9.6% | 8.8% |
| Core Web Vitals pass on phones, Türkiye | 26.8% | 34.5% | 47.3% | 51.1% | 57.5% |
HTTP Archive and Chrome UX Report, each September, top 10,000 sites per country · September 2026
How many years behind
On speed, Türkiye today is where the median of the other nine was 1.4 to 2.2 years ago. On consent platforms and marketing automation the distance is longer than the data: that median had already passed today's Turkish level in 2021. On TikTok Türkiye is not behind.
| Measure | The others were here | Türkiye, 2026 | Others' median, 2026 |
|---|---|---|---|
| Core Web Vitals pass on phones | 1.4 years ago | 57.5% | 62.9% |
| Good loading (LCP) | 2.2 years ago | 79.7% | 84.7% |
| Good responsiveness (INP) | 2.2 years ago | 69.9% | 78.1% |
| Good layout stability (CLS) | 1.8 years ago | 79.4% | 86.2% |
| Consent platform, detected tools | before 2021 | 7.8% | 29.3% |
| Marketing automation | before 2021 | 8.8% | 20.5% |
| TikTok Pixel | not behind | 7.8% | 6.1% |
- Automation never took off. In 2021, 9.3% of Turkish sites ran marketing automation, near the 10.4% median of the others. It is 8.8% today; the others' median is 20.5%.
- Tag management fell behind. Türkiye was 2.6 points below the median in 2021 and 11.3 points below in 2026. Every country drops in 2023, when detection changed; Türkiye dropped furthest and has not recovered.
- Consent tooling barely moved. Consent platforms went from 6.0% to 7.8% of Turkish sites, with a peak of 9.5% in 2023. The series counts detected tools only, so it misses local ones and sits below today's figure. The gap to the median is 21.5 points.
- Speed improved, but less. Turkish sites passing Core Web Vitals on phones went from 26.8% in 2022 to 57.5% in 2026, a gain of 30.7 points; Germany, the Netherlands and the UK each gained 36.0 points or more.
- TikTok grew fastest. Its pixel went from 1.6% to 7.8% of Turkish sites.
History months use detected technologies only, so the five-signal score leaves out Consent Mode and server-side traces and reads lower than the seven-signal index. CrUX ranks start in 2021 and INP in 2022. Wappalyzer's detection rules and HTTP Archive's crawl list change over time, so moves shared by every country, such as the tag manager drop in 2023, are more likely detection than behaviour; this section reads the comparison between countries. Years behind interpolates between Septembers, so read them as approximate; the median of nine countries moves as a group, and no single country is the yardstick.
Method
The report reads the 10,000 most visited websites in each of ten countries (Türkiye, Germany, France, the UK, Italy, Spain, the Netherlands, Poland, Romania, the US) and measures what a browser sees on first load with nothing accepted. Rank comes from the Chrome UX Report; page data from HTTP Archive's mobile crawl of each home page.
A country counts as lowest or highest only when the gap to the next country is larger than sampling noise, tested at the 95% level; otherwise the text names the country it ties with. Every number on this page is computed from the data, including these comparisons.
HTTP Archive covers a smaller share of Türkiye's top 10,000 than of the other countries', because sites popular only in Türkiye fall outside its global crawl list more often. It crawls from the US and sees home pages only. Server-to-server data, apps and pages behind a login are not seen. The crawl accepts nothing, so tags that wait for consent are not seen either, and analytics and ad pixel shares are minimums.
Speed models read field data from real Chrome users on phones at the 75th percentile, against what HTTP Archive detects on each home page plus the requests, scripts and hosts it loads. The share explained is R² of log INP, log LCP and the pass: a linear model with country and rank tier always in it, split between groups by Shapley values, then gradient boosting tested on a quarter of sites held out. Differences mean differences between sites in this set, not the cause of any one site's speed.
Site-level rows are used only during the analysis. They are not published or stored in the project; this page and its data file hold aggregates.
Sources
- HTTP Archive: mobile crawl of each home page: technologies, cookies, requests
- Chrome UX Report: the top sites per country and field performance from real Chrome users
- Cloudflare Radar: mobile share of likely-human traffic by visitor country
License: CC BY 4.0. The analysis and aggregates may be reused with credit to arif.media. This does not cover the Cloudflare Radar figures: they come from the Radar API under CC BY-NC 4.0, so they may be reused with credit to Cloudflare Radar, for non-commercial purposes only.
Download the data (CSV) · Read as markdown
How to cite
Hangişi, A. O. (2026). Web Measurement Report Card: Türkiye. arif.media. Data: September 2026. https://arif.media/reports/web-measurement-report-card-turkiye/2026-09/