Email Authentication in 2026: DMARC, SPF and Mail Gateways at 10,655 Organizations

We looked up the public DNS records of 10,655 companies, government bodies, universities and nonprofits and recorded how each one authenticates its email. 74.6% publish a DMARC policy that tells receivers to quarantine or reject spoofed mail, and 49.1% use the strictest setting, p=reject. The remaining quarter either monitor without enforcing or publish no policy at all.

74.6%enforce DMARC (quarantine or reject)
49.1%publish p=reject
18.6%monitor only (p=none)
6.8%have no valid DMARC policy

The sample is the set of organizations that have a DNSai company profile. It is weighted toward large, recognizable organizations and toward the United States, so read these figures as a picture of established organizations, not of every registered domain. The methodology section describes the sample and its limits.

DMARC policy: three in four enforce

A DMARC record tells receiving mail servers what to do with a message that claims to come from a domain but passes neither SPF nor DKIM in alignment with it. p=none asks only for reports, p=quarantine asks for the message to be treated as suspicious, and p=reject asks for it to be refused.

DMARC policy across all 10,655 organizations

Share of organizations by the policy that applies to their domain

GroupOrganizationsNo policyp=nonep=quarantinep=rejectEnforcing
All organizations10,6556.8%18.6%25.5%49.1%74.6%

Two details sit behind those totals:

  • Monitoring without reports. 364 of the 1,987 organizations at p=none (18.3%) publish no rua address. A p=none policy is meant to be a temporary stage in which aggregate reports reveal which legitimate senders still fail authentication. Without a reporting address the record has no practical effect.
  • Partial enforcement is rare. Only 204 enforcing organizations (2.6%) still limit their policy to a share of failing mail with a pct value below 100. The current DMARC specification, RFC 9989, removed that tag in May 2026, so receivers that follow it apply the policy to all failing mail regardless.

To see where a specific domain stands, the DMARC Lookup shows its record and policy, and the DMARC record generator builds a record for each stage of a rollout.

DMARC enforcement by industry

Enforcement is highest in Banking (87.8%), Insurance (85.9%) and Financial Services (85.8%), the sectors where spoofed email is most directly tied to fraud. It is lowest in Education (61.7%), Sports (61.8%) and Automotive (64.2%).

DMARC policy by industry

Industries with at least 100 organizations in the sample, ordered by the share that enforce (quarantine or reject)

IndustryEnforcingBanking123 organizations87.8%Insurance156 organizations85.9%Financial Services337 organizations85.8%Healthcare244 organizations82.8%Consulting132 organizations81.8%Software & Technology2,700 organizations80.1%Legal Services250 organizations78.8%Nonprofit252 organizations78.2%Retail543 organizations75.7%Logistics & Transportation117 organizations75.2%Manufacturing227 organizations74.4%Aerospace & Defense118 organizations73.7%Apparel & Fashion345 organizations72.8%Consumer Goods619 organizations71.1%Government607 organizations71.0%Entertainment & Media540 organizations70.7%Restaurants172 organizations69.2%Telecommunications126 organizations69.0%Food & Beverage679 organizations66.9%Travel & Tourism200 organizations65.0%Automotive179 organizations64.2%Sports293 organizations61.8%Education436 organizations61.7%
View as a table
IndustryOrganizationsNo policyp=nonep=quarantinep=rejectEnforcing
Banking1234.9%7.3%9.8%78.0%87.8%
Insurance1563.2%10.9%24.4%61.5%85.9%
Financial Services3372.4%11.9%21.1%64.7%85.8%
Healthcare2442.5%14.8%21.7%61.1%82.8%
Consulting1324.5%13.6%17.4%64.4%81.8%
Software & Technology2,7005.7%14.2%34.8%45.3%80.1%
Legal Services2505.2%16.0%20.4%58.4%78.8%
Nonprofit2524.0%17.9%42.5%35.7%78.2%
Retail5434.6%19.7%22.8%52.9%75.7%
Logistics & Transportation1173.4%21.4%24.8%50.4%75.2%
Manufacturing2275.7%19.8%22.9%51.5%74.4%
Aerospace & Defense1184.2%22.0%27.1%46.6%73.7%
Apparel & Fashion3453.2%24.1%22.6%50.1%72.8%
Consumer Goods6196.9%22.0%22.8%48.3%71.1%
Government60711.0%18.0%10.5%60.5%71.0%
Entertainment & Media5408.9%20.4%24.3%46.5%70.7%
Restaurants1728.7%22.1%27.9%41.3%69.2%
Telecommunications1266.3%24.6%17.5%51.6%69.0%
Food & Beverage67913.7%19.4%13.8%53.0%66.9%
Travel & Tourism20011.5%23.5%26.0%39.0%65.0%
Automotive1798.9%26.8%20.1%44.1%64.2%
Sports29311.3%27.0%27.0%34.8%61.8%
Education4361.8%36.5%29.1%32.6%61.7%

Two sectors do not follow the simple ranking:

  • Education publishes DMARC almost universally but enforces least. Only 1.8% of education organizations lack a policy, the lowest share of any industry, yet 36.5% are at p=none. Universities typically have many departments and third-party systems sending mail under one domain, which makes the move from monitoring to enforcement slow.
  • Government is split. 60.5% of government domains are at p=reject, one of the higher rates in the study, while 11.0% have no policy at all. The group mixes U.S. federal agencies, which have been required to publish p=reject since CISA's Binding Operational Directive 18-01, with state, county and city domains that have no equivalent mandate.

SPF: nearly universal, evenly split between hard and soft fail

94.7% of organizations publish an SPF record. The mechanism at the end of the record states what receivers should do with mail from servers the record does not list.

How SPF records end

Share of all 10,655 organizations

-all (hard fail)47.0%~all (soft fail)43.6%redirect= to another record3.0%?all (neutral)0.9%No all mechanism or +all0.2%No SPF record5.3%

Read from each domain's own SPF record. Records that hand off to another record with redirect= are counted separately because the final qualifier lives in the target record.

The near-even split between -all and ~all matters less than it once did. Under DMARC, a message fails when neither SPF nor DKIM passes in alignment, whichever qualifier the SPF record ends with, so an organization at p=reject with ~all is better protected than one with -all and no DMARC policy. Taken together, 47.7% of organizations publish both SPF and p=reject, and 3.3% publish neither SPF nor DMARC. Check any record with the SPF Analyzer or build one with the SPF record generator.

Where inbound mail is delivered first

MX records name the servers that accept mail for a domain, which shows whether mail goes straight to a mailbox provider or passes through a separate security gateway first. 92.6% of organizations publish MX records that accept mail. 7.0% publish none, and 0.4% publish a null MX, the explicit statement that a domain accepts no email.

First destination for inbound mail

Share of the 9,871 organizations that accept mail, by the operator of their MX hosts

Microsoft 3653,344 organizations33.9%Google Workspace2,330 organizations23.6%Proofpoint1,741 organizations17.6%Mimecast791 organizations8.0%Cisco Secure Email191 organizations1.9%Barracuda91 organizations0.9%Self-hosted or unrecognized757 organizations7.7%Other named provider626 organizations6.3%

An organization that routes mail through a gateway to Microsoft 365 or Google Workspace is counted under the gateway, because that is what its MX records name. "Self-hosted or unrecognized" covers MX hosts inside the organization's own domain and hosts that match no known provider.

Microsoft 365 and Google Workspace receive mail directly for more than half of the sample. Among organizations that place a dedicated gateway in front of their mailboxes, Proofpoint and Mimecast account for most of the rest.

DMARC policy by mail gateway

DMARC is published by the domain owner, not the mail provider, so differences between these groups reflect the kind of organization that chooses each setup. Proofpoint customers have the highest p=reject rate at 66.4%. Organizations running their own mail servers or using smaller providers are the most likely to have no policy.

DMARC policy by first mail destination

Organizations that accept mail, grouped by the operator of their MX hosts

MX operatorEnforcingMicrosoft 3653,344 organizations76.7%Google Workspace2,330 organizations78.9%Proofpoint1,741 organizations82.8%Mimecast791 organizations77.6%Cisco Secure Email191 organizations74.9%Barracuda91 organizations65.9%Self-hosted or unrecognized757 organizations65.0%Other named provider626 organizations63.3%
View as a table
MX operatorOrganizationsNo policyp=nonep=quarantinep=rejectEnforcing
Microsoft 3653,3442.6%20.7%31.0%45.6%76.7%
Google Workspace2,3303.0%18.1%34.0%44.9%78.9%
Proofpoint1,7411.4%15.8%16.4%66.4%82.8%
Mimecast7913.7%18.7%25.8%51.8%77.6%
Cisco Secure Email1911.6%23.6%26.7%48.2%74.9%
Barracuda911.1%33.0%30.8%35.2%65.9%
Self-hosted or unrecognized75712.7%22.3%18.2%46.8%65.0%
Other named provider62614.1%22.7%24.1%39.1%63.3%

To identify the gateway for a single domain, use the MX Lookup. For a list of accounts, the email gateway lookup returns the gateway and DMARC policy for each one.

Who receives DMARC reports

92.4% of organizations with a DMARC policy publish an address for aggregate reports. The domain of that address shows who processes them: 59.2% send reports to one of the commercial or government reporting services below, and the rest mostly collect them in their own mailboxes.

DMARC report destinations

Share of the 9,935 organizations with a DMARC policy whose rua address belongs to each service

Proofpoint Email Fraud Defense1,561 organizations15.7%Valimail1,146 organizations11.5%dmarcian678 organizations6.8%Mimecast DMARC Analyzer613 organizations6.2%Cloudflare DMARC Management459 organizations4.6%Red Sift OnDMARC330 organizations3.3%EasyDMARC279 organizations2.8%Postmark DMARC222 organizations2.2%CISA (U.S. federal reporting)199 organizations2.0%Agari (Fortra)195 organizations2.0%MxToolbox182 organizations1.8%DMARC Advisor143 organizations1.4%

Twelve most common services. An organization that reports to two services is counted under both.

Services verified in DNS

Many services ask a customer to prove ownership of a domain by publishing a verification token in a TXT record, usually when setting up single sign-on or an organization-wide account. Those tokens are public, so they show which services an organization has connected to its domain.

AI services with a domain verification record

Share of all 10,655 organizations

Anthropic (Claude)3,383 organizations31.8%OpenAI (ChatGPT)2,168 organizations20.3%Cursor1,034 organizations9.7%Perplexity169 organizations1.6%ElevenLabs113 organizations1.1%Mistral AI24 organizations0.2%

Selected business software referenced in DNS

Share of all 10,655 organizations with a verification token or other record that names the service

Atlassian4,466 organizations41.9%DocuSign2,851 organizations26.8%Salesforce2,203 organizations20.7%Zoom1,576 organizations14.8%Miro1,436 organizations13.5%HubSpot1,283 organizations12.0%Smartsheet1,244 organizations11.7%Zendesk1,152 organizations10.8%Canva877 organizations8.2%Slack802 organizations7.5%Figma764 organizations7.2%KnowBe4721 organizations6.8%1Password540 organizations5.1%Airtable531 organizations5.0%Notion530 organizations5.0%

These figures are a lower bound on adoption, not a measure of it. A record shows that someone at the organization verified the domain with the service at some point. It says nothing about how many people use the service, and an organization can use a service without ever publishing a record for it. Any company's profile lists the services found in its DNS, for example stripe.com.

Methodology

  • Sample. The 10,750 organizations with a DNSai company profile, one domain each. They were selected for being established and recognizable, not at random: the list is weighted toward the United States, toward larger organizations, and toward software and technology companies, which make up about a quarter of it. 95 domains were left out because they no longer resolve or did not return a complete lookup, leaving 10,655.
  • Collection. Each domain was queried between November 15 and November 21, 2026 through public recursive resolvers (Google Public DNS, with Cloudflare as the fallback) for its MX records, its TXT records and the TXT record at _dmarc. Each organization is counted once, using its most recent complete lookup in that window.
  • DMARC policy. The policy is the p tag of the record at _dmarc.<domain>. Where the profiled domain is a subdomain with no record of its own, the organizational domain's record applies, using its sp tag when present. Records with no valid policy tag are counted as having no policy.
  • SPF. The qualifier on the all mechanism of the domain's own SPF record. Included records are not expanded.
  • Mail gateway. MX hostnames are matched against known provider hostname patterns, and product variants of one vendor are combined. When a domain's MX hosts belong to more than one provider, the first one identified is used.
  • Services. TXT, SPF and related records are matched against known verification-token and hostname patterns.
  • Not measured. DKIM is absent from this study. DKIM keys are published under selector names that cannot be listed from DNS, so their absence cannot be established from outside.

All percentages are rounded to one decimal place. The aggregate figures behind every chart on this page are available as a JSON file. It contains counts and percentages only, with no per-domain data.

Citing this study

The figures may be quoted with attribution to DNSai and a link to this page.

DNSai Research. "Email Authentication in 2026: DMARC, SPF and Mail Gateways at 10,655 Organizations." November 21, 2026. https://dnsai.com/research/email-authentication-2026/

Frequently asked questions

What share of organizations enforce DMARC in 2026?

In this sample of 10,655 organizations, 74.6% publish an enforcing DMARC policy: 49.1% use p=reject and 25.5% use p=quarantine. A further 18.6% publish p=none, which only monitors, and 6.8% have no valid DMARC policy. The sample is weighted toward large and well-known organizations, so these figures are higher than they would be for all registered domains.

Which industries have the highest DMARC enforcement?

Among industries with at least 100 organizations in the sample, enforcement is highest in Banking (87.8%), Insurance (85.9%) and Financial Services (85.8%) and lowest in Education (61.7%), Sports (61.8%) and Automotive (64.2%).

Does a p=none DMARC record protect a domain from spoofing?

No. A p=none policy asks receivers to deliver mail that fails DMARC as usual and only send reports about it. It is the intended first step of a rollout, because the reports show which legitimate senders still need SPF or DKIM alignment, but spoofed mail is not blocked until the policy is raised to quarantine or reject.

How was this data collected?

Each organization's domain was queried through public recursive DNS resolvers between November 15 and November 21, 2026 for its MX records, its SPF record and the DMARC record at _dmarc. Only complete lookups are counted. All of the records involved are public by design, since mail servers have to read them to deliver and authenticate email.

Can I cite these figures?

Yes. The figures may be quoted with attribution to DNSai and a link to this page. The aggregate data behind every chart is available as a JSON file linked in the methodology section.

Data snapshot: 2026-11-15 to 2026-11-21. Page last updated: .