Nixeny Dijital - Mersin / State of Digital 2026
72 types of business were scanned across the 13 districts of Mersin. The digital address of 19,615 active businesses was classified and 4,167 domains put through a technical audit. One question: what does someone searching for a Mersin business online actually find?
What these numbers measure, and what they do not
This is not a population count but a sample from a defined, repeatable systematic scan. The accurate phrasing: “No independent website could be found for 75.7% of the Mersin businesses included in the study.”
What is measured is whether the business shared a web address on its Google profile. It does not prove the business has no site anywhere on the internet; it shows the business did not put one on its profile.
Every percentage carries its own denominator: 19,615 businesses, 4,167 domains, 3,395 reachable sites, 689 sites with real-user data. Comparisons always state the denominator. The aim is a sense of scale, not a ranking.
The “digital score” used in this report is a derived metric specific to this study that reduces eight measurements to one number; it is not a Google standard. The headline figures are taken from raw measurements.
Glossary · every term explained once
- Google business profile
- The business card that appears in Google Maps and Google Search: name, address, phone, rating and, where given, a web address.
- Independent website
- A site on the business's own domain, under its own control. An Instagram page or a directory listing does not meet this definition.
- Third-party platform
- A page the business uses as its digital address that somebody else owns: a social media profile, a directory record, a classifieds page.
- Lighthouse
- Google's open-source site audit tool. It opens a page on a simulated phone and scores it 0-100 in four categories.
- Performance score
- A 0-100 rendering of how quickly the page becomes usable on a phone. Below 50 is Google's own “poor” threshold.
- Basic mobile test
- Opening the site in a real browser at a 390×844 pixel phone viewport: is there a viewport tag, does content overflow sideways, is anything pushed off-screen.
- Structured data (JSON-LD)
- A machine-readable description embedded in the page: what this place is, where it is, what its hours are. Search engines and AI assistants read it.
- HTTPS
- An encrypted connection between browser and site. Without it the browser writes “not secure” in the address bar.
- HSTS · CSP
- Two security instructions a server sends the browser. Their absence is a hygiene gap; it does not mean “this site is dangerous”.
- CrUX (field data)
- Experience data Google collects from real Chrome users. For a site to appear in it, it needs a measurable level of visitors.
- Core Web Vitals
- Google's three field metrics: when the main content appeared (LCP), how slowly it answered a tap (INP), and how much the page jumped while loading (CLS).
- Median
- Not the mean: the value in the exact middle of a sorted list. Used so that a handful of exceptionally good or bad sites cannot skew the result.
- Denominator (n)
- How many records a percentage was calculated over. In this report it is written next to every percentage.
- CMS
- The content management system or site builder the site is built on.
Three in four businesses have no address to go to
No independent website could be found for 14,842 of the 19,615 businesses. The reason is usually simple: no web link was shared on the business profile at all. A consumer who finds that business on Google has no address to go to.
A small group is visible online, but not at an address of its own. 1,548 businesses built their digital presence on someone else’s platform: an Instagram profile, a directory record, a classifieds page. If the platform closes, changes its rules or suspends the account, that address is no longer the business’s to keep.
Each square represents one percent of the businesses scanned. 100 squares in total.
Akdeniz's rate is three times Mut's
The map below was not drawn: the real latitude and longitude values from 19,615 businesses’ Google profiles were aggregated into 500-metre cells and printed. The shape of Mersin, its coastline and the Taurus settlements emerge from the distribution of the businesses themselves. The red marks are businesses with an independent website.
Each district's own denominator is written on its row: Akdeniz's 37.0% is calculated over 4,106 businesses, Çamlıyayla's 16.9% over 431.
68.8% for hotels, 3.1% for pharmacies
There is no single “Mersin average”. The gap between the highest and lowest type is 65.7 percentage points: 68.8% of hotels have a site of their own, against 3.1% of pharmacies. The 66 types in between are squeezed between those two ends.
The split is not random. Site ownership rises where the customer is outside the province or outside the country: logistics, customs, shipping, software. It falls where the customer is on the street: hairdressers, bakeries, car washes, pharmacies.
Each tile's area varies with the number of businesses in that category and the darkness of its fill with the independent site rate. Two tiles of the same size mean the same number of businesses.
Each tick is one business type, placed on the horizontal axis at that type's independent site rate. Tick height carries no data; it only keeps overlapping values distinguishable. Only the 68 types with at least 50 businesses are drawn; 4 of the 72 fall below that threshold and are left out.
- Hotel68.8%n 96
- Customs brokerage67.9%n 112
- International freight66.0%n 159
- Logistics company63.5%n 208
- Ship agency59.0%n 61
- Software / IT57.7%n 317
- Pharmacy3.1%n 487
- Bakery3.2%n 410
- Auto electrician4.4%n 250
- Hairdresser / barber4.6%n 836
- Car wash4.6%n 369
- Pilates studio7.8%n 64
Red end: the share of businesses using a third-party platform as their digital address. Navy end: the share with an independent site of their own. The line between them is the distance; a short line means a sector where the two options have moved close together.
One business in every 21 rents its address
In Mersin one business in every 21 has a social media profile as its only digital address; 775 of those 921 businesses gave that address on Instagram. If the account closes, the rules change or the profile becomes unreachable, the business’s only record on the internet goes with it.
Another 430 businesses give a directory, booking or marketplace page as their web address; 165 give a Sahibinden classifieds page directly; for 21 businesses the “website” is a WhatsApp link. The content of those pages, and the customer data that comes through them, are under the platform’s control rather than the business’s.
Circle area varies with business count. The 430 businesses are spread across 51 different platforms; the 10 most crowded are shown here.
A third of the best-known businesses have no address
Review count is the crudest available indicator of how well known a business is. Site ownership rises with reviews, but reaches a majority in no bucket. 34.3% of the 473 businesses with more than 500 Google reviews have no web link on their profile at all. This is not a claim of causation; it is two measurements moving together.
772 domains never opened during the audit
A web address printed on a profile does not mean that address opens. At the time of the audit 772 domains did not respond at all; 561 of them no longer resolve in DNS, meaning the registration has lapsed or been abandoned. On 66 sites the browser threw a certificate error: the visitor meets a full-page security warning.
Nobody's opinion is needed. Type it into the address bar and look.
561 domains do not resolve in DNS, 66 sites have a certificate error, 279 sites still do not use HTTPS. All three can be checked one by one; none of them is open to interpretation.
The step widths are a readable ramp, not a share of the first step. The real rates are written under each step.
Two thirds carry no structured data
Having a site built and having a site that works are not the same thing. The median mobile speed score of the audited sites is 61/100; 19.5% sit below Google’s 50-point “poor” threshold. Of the sites opened at a phone viewport in a real Chromium browser, 14.4% failed the basic mobile test.
The gaps are mostly in the layer nobody sees. 66.1% of the sites carry no structured description data (JSON-LD) of the kind search engines and AI assistants read. 44.5% have no canonical tag and 41.9% no sitemap. 279 sites still do not use HTTPS: the browser writes “not secure” on those pages.
Each row is divided into 25 cells; the filled cells show the share of sites applying that step. One cell is about 135 sites. Rows below 40% are red: below that share a step stops being a default and becomes an exception.
| Gap | Sites | Denominator | Rate |
|---|---|---|---|
| No Content-Security-Policy | 2,957 | 3,395 | 87.1% |
| No HSTS | 2,547 | 3,395 | 75.0% |
| No WhatsApp link | 2,270 | 3,381 | 67.1% |
| No contact form | 2,264 | 3,381 | 67.0% |
| No structured data (JSON-LD) | 2,236 | 3,381 | 66.1% |
| No og:image | 2,172 | 3,381 | 64.2% |
| No single H1 heading | 2,106 | 3,381 | 62.3% |
| No tel: link | 1,678 | 3,381 | 49.6% |
| No canonical | 1,504 | 3,381 | 44.5% |
| No sitemap.xml | 1,392 | 3,325 | 41.9% |
Source: home-page audit of 3,395 reachable sites · “No WhatsApp link” and “no contact form” are not necessarily gaps but may be a channel choice; they are listed as raw observations.
The digital score is a derived composite metric specific to this study; it is not a Google standard. It reduces eight measurements to a single 0-100 number and should be used only to read a site's position relative to the others.
These three measurements are not a lab test but the 75th-percentile experience of real Chrome users. They can only be measured on sites that reach Google's data threshold; each measurement has a different denominator.
How long a real user waits to see the largest content on the page. Google's good threshold is under 2,500 milliseconds; the median Mersin site is at 2,083 ms.
How long the page takes to react after a user taps something. The good threshold is under 200 milliseconds.
Content shifting out of place during load. The good threshold is 0.10 and below.
Two rates are close, the denominators are not
The two rates in this section do not measure the same denominator. TurkStat’s universe is enterprises with at least 10 employees; this study’s universe is 19,615 local businesses registered on Google regardless of size, most of them micro-scale. The global web measurements are taken over millions of sites worldwide, not over Mersin businesses. The figures should therefore be read for a sense of scale, not as a ranking.
This study's universe is 19,615 local businesses registered on Google regardless of size; TurkStat's universe is enterprises with at least 10 employees. The difference in denominator alone explains most of the distance between them.
Here the two rates are very close. But the Mersin denominator is the 689 sites for which field data exists — that is, the minority that already has measurable traffic. 72.6% of the 3,378 audited sites never reach that threshold.
The global measurement is at page level, this study at home-page level. The two rates do not count the same thing; the direction is the same, the level is not comparable.
Limitations
- The source is a single system of record.
- The whole digital presence classification comes from Google business profiles. A business that has a site but has not updated its profile does not appear as one in this study.
- The scan is a sample.
- 1,131 queries were run across 87 search terms × 13 districts, with 20 results per page and at most 5 pages per query. Businesses falling outside that design were not covered.
- Sector assignment is single-choice.
- 3,399 businesses appeared in more than one sector query; each was assigned to a single study sector. The sector rates depend on that assignment rule.
- The measurement belongs to a date.
- All technical audits were run around 2026·08·25. The speed, certificate and content of the sites are as they stood that day.
- Only home pages were audited.
- Structured data, a contact form or a booking step that might exist on inner pages was not measured; the rates should therefore be taken as lower bounds.
- The 772 unreachable domains could not be measured.
- The denominator for the quality rates is the 3,395 reachable sites. Addresses that do not work are excluded in a direction that improves the quality average.
- District assignment comes from the query.
- A business's district is taken from the district query that returned it; the map labels are the centre of gravity of business coordinates, not an official district boundary.
- Security headers are a hygiene indicator.
- The absence of HSTS or CSP does not show that a site has been attacked or is unsafe.
The method, briefly
- 01DiscoveryGoogle Places API (New) text search, 1,131 queries on the pattern “{term} {district} Mersin”; 2,594 API requests. Results were de-duplicated by Place ID.
- 02Geographic verificationProvince check from the address components. 151 records falling outside Mersin were removed from the main dataset, leaving 19,615 active local businesses.
- 03ClassificationThe web address field on the profile was split into own domain, social media profile, directory or marketplace page, classifieds page and other third party. A reason and a confidence value were recorded for every decision.
- 04Site auditHTTP and TLS audit of 4,167 de-duplicated domains, a 390×844 mobile viewport test in a real Chromium browser, a Google PageSpeed Insights mobile audit and Chrome UX Report field data.
- 05Missing dataAny metric that could not be measured is removed from its own denominator; a blank is never counted as zero. That is why you see different denominators within one section.
- 06Verification23 automated consistency checks and a replay sample from the raw API data over 20 businesses and 20 sites. All checks passed, and the replay produced no field mismatch.
Sources
- 01Google Places API (New) business profilesThe primary data source; name, address, coordinates, phone, rating, review count, business status and the web address field. Fieldwork date 2026·08·25.
- 02HTTP and TLS audit of 4,167 domainsDirect requests to 4,167 de-duplicated domains; response code, redirect chain, certificate validity and security headers.
- 03Mobile viewport test in a real Chromium browserViewport tag, horizontal overflow, off-screen elements and the amount of data downloaded, at a 390×844 pixel viewport.
- 04Google PageSpeed InsightsMobile Lighthouse audit: Performance, Accessibility, Best Practices, SEO and the lab timings.
- 05Google Chrome UX ReportReal-user field data: the 75th-percentile values for LCP, INP and CLS.
- 06TurkStat, ICT Usage Survey in Enterprises 2025For the outside reference in Figure 16 only. Website ownership among enterprises with at least 10 employees is 56.5% (2024: 51.8%). Published September 2025.veriportali.tuik.gov.tr
- 07HTTP Archive Web Almanac 2025For the outside reference in Figure 16 only. On July 2025 CrUX data, 48% of sites pass Core Web Vitals on mobile; the median mobile home page weighs 2,362 KB; 17.2 million sites.almanac.httparchive.org
- 08HTTP Archive Web Almanac 2024, structured data chapterFor global comparison: JSON-LD is on 41% of pages (34% in 2022).almanac.httparchive.org
Both outside sources are primary publications and are given with their own denominators. The TurkStat data is Türkiye-wide and the Almanac data is the global web; neither is a measurement specific to Mersin. That difference is not a gap in this report but a limit written down plainly.
Colophon
Conducted and published by Nixeny Dijital, Mersin. Data collection, classification, technical audit and report design were carried out by Nixeny.
Method notes, raw datasets and verification logs are shared on request. For corrections and questions: info@nixeny.com