mailskop.

Measurement report ·

Email security measurement, October 2026: DMARC, SPF and DKIM across 26,135 domains

By: Kemal Özleyen, Cybersecurity Engineer · Check Point Workspace Security Expert · Kinetik Bilişim

From the DNS records of 26,135 domains checked on Mailskop between 26 July 2026 and 31 October 2026. Figures about Türkiye come only from the 8,792 .tr domains.

Infographic of the Mailskop October 2026 research. 63.8% of 26,135 domains measured on Mailskop have no DMARC policy that stops forged email; among those measured since September it is unchanged at 70.4%. The figures are given as text in the sections below.

PNG · Turkish version · Data (CSV) · Blog article · CC BY 4.0 — share freely with attribution (Mailskop · Kinetik Bilişim).

The security product and domain age sections were computed from the domain-level export of the same measurement and are not in the tables on this page. Age is taken from the registration date (RDAP, TRABIS WHOIS for .tr) as of 31 October 2026; only the date was read. Rates by product reflect the configuration of the organisations using it, not the product.

Key findings

  • Of 8,792 .tr domains measured on Mailskop, 46.9% have no valid DMARC record, and another 24.2% use p=none — monitoring without protection.
  • Only 28.9% of the same 8,792 .tr domains enforce a DMARC policy: 9.1% reject and 19.8% quarantine.
  • SPF is widespread: 95.0% of 8,721 .tr domains with a measurable SPF lookup publish one. But 1.0% of the 8,591 domains whose SPF records could be counted publish more than one, which invalidates SPF entirely under RFC 7208.
  • A DKIM key was found on 50.6% of 8,192 .tr domains where DKIM could be checked. This is a lower bound: selector names cannot be listed from DNS, so a key we did not find is not a key that does not exist.
  • Transport security is almost absent: of 3,482 .tr domains measured, 29 are DNSSEC-signed, 3 enforces MTA‑STS, and none of 3,378 publish DANE.

Compared with September 2026

The left value is the earlier snapshot, the right value this one. The sample grows with new queries every period, so part of any change in the aggregate rates comes from newly added domains; it should not be read as “improvement” on its own.

SegmentDomainsrejectquarantinenoneNo / invalid record
.tr domains8,673 → 8,7928.9% → 9.1%19.8% → 19.8%24.1% → 24.2%47.2% → 46.9%
Other TLDs12,522 → 17,34313.8% → 20.3%16.2% → 19.6%21.9% → 21.5%48.1% → 38.6%

Looking at the same domains: of 1,339 .tr domains re-measured between the two snapshots, 37 strengthened their DMARC policy and 6 weakened it; the rest stayed the same. Domains not re-measured are not counted: what we did not measure is not counted as unchanged.

Before ↓ / After →rejectquarantinenoneNo / invalid record
reject226020
quarantine942731
none283630
No / invalid record3411280

DMARC policy

The policy in each domain’s most recent scan. reject has forged mail rejected, quarantine sends it to spam, none only asks for reports. An invalid record is the same as no record to a receiver and is counted in that column (107 domains).

SegmentDomainsrejectquarantinenoneNo / invalid record
.tr domains8,7929.1%19.8%24.2%46.9%
Other TLDs17,34320.3%19.6%21.5%38.6%

SPF and DKIM

The most recent scan per domain in which SPF could be measured. More than one SPF record is a “permerror” under RFC 7208: the record looks present but never works. DKIM rates are lower bounds.

SegmentMeasuredHas SPFMultiple SPFDKIM checkedDKIM foundMultiple keys
.tr domains8,72195.0%1.0%8,19250.6%12.8%
Other TLDs17,23590.8%0.9%15,26652.5%21.3%

Transport security and multiple DMARC

These columns were added to the measurement later and are filled only for domains scanned since then. The denominators are therefore smaller and are shown in every cell.

SegmentDNSSECMTA‑STS enforceMTA‑STS testingDANEMultiple DMARC
.tr domains29 / 3,4823 / 3,48211 / 3,4823 / 3,37813 / 3,458
Other TLDs582 / 9,28467 / 9,28491 / 9,28424 / 8,86926 / 9,058

DMARC by mail infrastructure

Whole sample. Infrastructure is inferred from MX, SPF and DKIM traces. Individual hosting companies are deliberately not named: detection is imperfect, and a measurement error should not become a claim about a company. The “Infrastructure not detected” row may include domains that do not use email at all; we did not measure this separately, so compare that row with care.

InfrastructureDomainsrejectquarantinenoneNo / invalid record
Microsoft 3658,10022.6%27.6%22.6%27.2%
Google Workspace3,09612.5%17.6%26.3%43.6%
Yandex 3606531.4%9.5%21.6%67.5%
Self-hosted mail server4,1489.4%20.7%30.2%39.8%
Other hosting and mail services7,6499.5%15.1%20.5%54.8%
Infrastructure not detected2,48938.9%11.3%10.1%39.7%

Sample, method and limits

The sample is not random: it is the set of domains checked on Mailskop. 86.7% of the latest measurements come from bulk checks — organisations queried as lists. The figures therefore describe these lists and the domains checked one by one, not “all domains in Türkiye”.

Each domain is counted once, by its latest measurement; test and invalid domains are excluded. In bulk checks the user’s browser performs the DNS lookups and the server evaluates them (67.7% of measurements).

DMARC is evaluated against RFC 9989. A value that could not be measured was not counted as “absent”; it was removed from the denominator.

This report was frozen on 8 November 2026; its numbers will not change, and corrections to the text are noted with a date. The statistics on the home page are recomputed nightly; since they measure different days, the two can show different numbers.

Data and citation

Download the aggregate counts as CSV. Aggregate counts only; no individual domain.

The data and the tables on this page are licensed under CC BY 4.0: you may use, share and adapt them freely with attribution.

To cite: Kemal Özleyen, Mailskop (Kinetik Bilişim), “Email security measurement, October 2026: DMARC, SPF and DKIM across 26,135 domains”, 8 November 2026. https://mailskop.com/en/research/2026-10

Check your own domain

Last updated: