This article is for hospital marketing and IT teams whose emails, such as appointment confirmations, patient newsletters, lab-report notices and bills, are landing in spam or not arriving. It is also for anyone told by a vendor that the domain needs "SPF, DKIM and DMARC" without being told what that means in a hospital with half a dozen systems sending as the same name.
Hospitals are harder to authenticate than most businesses. The hospital information system sends billing mail. A diagnostics platform sends reports. The CRM or email platform sends newsletters. The website form plugin sends enquiries. The corporate mailbox sends everything else. Each of them may claim to be the hospital's domain, and each needs authorising separately. Miss one and its mail fails authentication, which now has consequences that did not exist a few years ago.
We cover the three protocols, the current sender requirements from Gmail and Yahoo, a method for inventorying your sources, and the headers needed for unsubscribe. A sister article walks through the phased move to a strict DMARC policy.
Why this matters more since 2024
Gmail and Yahoo introduced requirements for bulk senders in early 2024. Google's email sender guidelines say that senders who send more than 5,000 messages a day to Gmail accounts must meet additional requirements, on top of baseline rules that apply to every sender. A hospital with a patient newsletter and daily appointment mail can cross that line without noticing, since the count is per sending domain across all systems.
The baseline and bulk requirements overlap in these areas:
- Authentication. All senders need SPF or DKIM. Bulk senders need SPF, DKIM and DMARC. Google states that the DMARC policy can be set to none.
- Alignment. For direct mail, the domain in the visible From address must align with the SPF domain or the DKIM domain.
- DKIM key length. Google requires a key of 1024 bits or longer for personal Gmail recipients and recommends 2048 bits if your provider supports it.
- TLS and DNS. Mail must travel over TLS, and sending IPs need valid forward and reverse DNS.
- Unsubscribe. Marketing and subscribed messages must support one-click unsubscribe and carry a clearly visible unsubscribe link in the body.
- Complaint rate. Google asks senders to keep their spam rate in Postmaster Tools below 0.10 percent and to avoid reaching 0.30 percent or higher.
Yahoo's sender best practices set out a parallel list. Bulk senders implement SPF and DKIM, publish a DMARC policy of at least none, support one-click unsubscribe, honour unsubscribes within two days and keep complaints below 0.3 percent. Yahoo's page does not define a numeric volume threshold for bulk senders, so a hospital should assume the rules apply once it sends meaningful volume.
Large hospital chains will often be bulk senders by volume. Smaller clinics may not be, but the same setup protects their mail from being spoofed, so it is worth doing regardless.
The three protocols in plain terms
SPF (Sender Policy Framework) is a DNS record that lists which servers may send mail for your domain's envelope address. Receivers check whether the connecting server is on the list.
DKIM (DomainKeys Identified Mail) adds a cryptographic signature to each message. The sending system signs with a private key, and the receiver verifies using a public key published in DNS at a name of the form selector._domainkey.yourdomain.
DMARC (Domain-based Message Authentication, Reporting and Conformance) ties them together. It requires that the domain the reader sees in the From field aligns with either the SPF or DKIM domain that passed, and it tells receivers what to do with failures: nothing, quarantine or reject. It also requests reports. The tags are defined in RFC 7489, including the policy tag p, the subdomain policy sp, the report address rua and the alignment modes adkim and aspf.
The detail that trips up hospitals is alignment. SPF can pass for the vendor's bounce domain, and DKIM can pass for the vendor's own domain, while neither matches the From address your patient sees. That mail passes the individual checks and still fails DMARC. The fix is to have the vendor sign with your domain (a custom DKIM selector) and, where possible, to use a custom return-path domain.
Inventory every system that sends as your domain
Begin with a source inventory, not DNS edits. Collect:
- Every system that sends mail with an address on your domain in the From field: HIS, LIS and radiology systems, patient portal, appointment engine, CRM, email marketing platform, payment gateway, careers site, website forms, helpdesk, and the mailboxes in Google Workspace or Microsoft 365.
- For each, the owner (internal or vendor), the sending domain or subdomain, the approximate monthly volume and the purpose.
- Whether the system supports a custom DKIM selector and a custom return-path.
Then check the evidence. The best source is DMARC aggregate reports, which you start receiving once you publish a record with an rua address. Even at p=none, they list every IP and source that sent mail claiming to be your domain, with pass or fail results. Expect surprises: an old survey tool, a shared mailbox used by a department, and a vendor that a previous team set up without telling IT.
Use subdomains to separate mail streams
A clean architecture is to separate streams by subdomain:
| Stream | Example subdomain | Why |
|---|---|---|
| Corporate mailboxes | Root domain | Staff email, highest trust |
| Transactional alerts | alerts. | Confirmations, OTPs, reports |
| Marketing and newsletters | news. | Complaint-prone, isolated reputation |
| Vendor-managed systems | Per vendor | Easier to revoke |
Each subdomain has its own SPF and DKIM. DMARC at the root can apply to subdomains through sp, or each subdomain can publish its own record. The point is that a complaint spike on the newsletter should not damage the delivery of appointment confirmations. Google's guidance on DMARC setup notes that subdomains inherit the parent policy unless sp is set, and that strict alignment can reject mail from subdomains you do not control.




