Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Email Marketing › Email Fundamentals

Email Deliverability

How email deliverability works: SPF, DKIM and DMARC, sender reputation, bulk sender requirements, list hygiene, warming a domain, and monitoring.
Guideintermediate12 min readUpdated

Delivery vs deliverability

A platform's "delivered" figure means the receiving server accepted the message. It says nothing about where the message went afterwards: inbox, promotions tab, spam folder, or silently discarded. Deliverability is that second question, and it is decided by the mailbox provider using signals the sender only partly controls. Delivery problems are visible as bounces and are usually technical; deliverability problems are invisible in the platform's own report and show up as falling engagement, customers saying that mail is "not arriving", and movements in the provider's postmaster tools.

The signals divide into three groups. Identity: can the provider verify that the message was sent by the domain it claims? Reputation: what have recipients done with mail from this domain and these addresses before, and how consistent is the sending pattern? Content and behaviour: does this message look like mail people want, and does the sender honour unsubscribe requests? Authentication fixes identity; list quality, consent and segmentation drive reputation; the campaigns themselves drive the third. Only the first is purely technical, which is why a team that "fixed deliverability" by adding DNS records and changed nothing else usually sees the problem return within a few months.

Authentication: SPF, DKIM, DMARC

Three standards let a receiving server verify that a message is genuinely from the domain it names. They are complementary, and bulk senders are expected to use all three.

StandardWhat it doesWhere it livesWhat it checksCommon failure
SPFLists the servers allowed to send mail for a domainA TXT record in the domain's DNSThe domain in the envelope sender (the return path) against the connecting server's addressThird-party platforms not included; record exceeding the lookup limit; envelope domain differing from the visible From
DKIMSigns each message with a private key so the receiver can verify it was not altered and that the signing domain takes responsibilityA public key in DNS under a selector; a signature header on each messageThe signature against the published keyPlatform not configured to sign with the sender's own domain; key rotated without updating DNS
DMARCPublishes a policy telling receivers what to do when a message fails authentication, and asks them to report backA TXT record at the _dmarc subdomain of the From domainThat SPF or DKIM passes and that the passing domain aligns with the visible From domainPolicy set to reject before all legitimate senders are aligned; reports never read

As dmarc.org describes it, DMARC builds on SPF and DKIM by adding alignment and policy. Alignment means the domain that passed SPF or DKIM must match, exactly or at the organisational level, the domain the recipient sees in the From header; without that, a message could pass SPF for one domain while displaying another. The policy, published in DNS, is one of p=none (monitor only), p=quarantine (treat failures as suspicious) or p=reject (refuse failures), and the record can request aggregate reports, which summarise who is sending as the domain and whether they pass, and failure reports with more detail. The path dmarc.org recommends is to start at p=none, read the aggregate reports until every legitimate source is authenticated and aligned, then move to quarantine and finally to reject.

Two practical points. A dedicated subdomain for marketing mail keeps its reputation separate from the domain used for personal and transactional correspondence, and makes alignment simpler. And every system that sends as the domain, including the CRM, the support desk, the billing platform and the marketing platform, must be in SPF and signing with DKIM, or DMARC enforcement will eventually block it. When a new domain, subdomain or platform is about to send its first campaign, the email sending domain setup checklist puts authentication, unsubscribe headers, warm-up and suppression into one list to tick.

Sender requirements from mailbox providers

Google publishes email sender guidelines that document what it expects before mail is accepted into Gmail inboxes, and other large providers publish their own. Google's guidelines distinguish all senders from bulk senders, the latter defined by a documented volume of messages per day to its users. For all senders, the guidelines document: authentication with SPF or DKIM for the sending domain; valid forward and reverse DNS records for sending IP addresses; a TLS connection for transmitting mail; a spam rate, as reported in Postmaster Tools, kept below the documented threshold; messages formatted according to the internet message format standard; no impersonation of Gmail in From headers; and ARC headers on mail that is forwarded.

For bulk senders the guidelines document additional requirements: authentication with both SPF and DKIM; a DMARC policy for the sending domain, with an enforcement level that may be none; alignment of the From header domain with either the SPF domain or the DKIM domain; one-click unsubscribe for marketing and subscribed messages, implemented with the List-Unsubscribe and List-Unsubscribe-Post headers that RFC 8058 specifies, alongside a clearly visible unsubscribe link in the message body; and processing of unsubscribe requests within the documented window.

The requirements are documented as thresholds and mechanisms, not as a guarantee of placement. Meeting them removes documented reasons for rejection; it does not earn the inbox. Where the guidelines give specific numbers, this site points to the source rather than repeating them, because they change. For the subscriber-facing side of unsubscribe, including why hidden or slow unsubscribes turn into complaints, see the list building and consent guide. M3AAWG, the industry working group on messaging abuse, publishes best common practice documents for senders that cover the same ground from the operator's side: bounce handling, complaint feedback loops, list hygiene and sending infrastructure.

Reputation and engagement

Reputation is the provider's memory of a domain and its sending addresses. It is built from what recipients do: complaints weigh heaviest, then whether mail is deleted unread, moved, replied to, opened or clicked, and whether the sender keeps mailing addresses that bounce or do not exist. Google's Postmaster Tools show domain and IP reputation as categories rather than a numeric score, which is a hint about how to think of it: a state that moves slowly in response to consistent behaviour, not a dial.

The sender's levers are indirect. Mailing only people who asked keeps complaints low. Removing hard bounces at once, and avoiding sources likely to contain abandoned addresses, means the provider never sees a sender that fails to maintain its list. Which sign-up method each form uses matters as well, and the comparison of single and double opt-in shows where a confirmation step keeps typos and bot sign-ups from ever reaching a send, and the double opt-in test is a design for settling it on one form, with first-send bounces and complaints as guardrails. Suppressing people who have not engaged within a chosen window, which the segmentation guide covers in detail, raises the engagement of what is sent and stops the long tail of inactive addresses from diluting it. Consistent cadence matters too; a domain that sends steadily is easier to trust than one that is silent for a month and then mails everyone.

Engagement cuts both ways. Low engagement is a cause of poor placement and a consequence of it, which makes the spiral hard to read from the inside. The honest measures are the ones the provider also sees: complaint rate, bounce rate and the proportion of the list that is active. Click-through (CTR) is useful because a click is unambiguous; opens are a weaker signal now that many clients pre-fetch images, as the email metrics and testing guide explains. The conversion rate (CVR) of email traffic is a business measure rather than a deliverability one, but it is often the first number to fall when placement does.

Warming and volume changes

A new domain or IP address has no reputation, and providers treat an unknown sender that appears with high volume as a risk. Warming is the practice of starting with a small volume to the most engaged subscribers and stepping it up gradually while watching the signals. Google's sender guidelines advise against sudden spikes in volume and recommend a gradual ramp. M3AAWG's sender practices cover the mechanics for operators. In practice, warming schedules are the sender's own judgement: begin with recent, active subscribers, grow the daily volume in steps rather than jumps, hold or step back if complaints or bounces rise, and keep the content representative of what will be sent later so that the reputation earned applies to it.

Volume changes on an established domain deserve the same care. A seasonal campaign to the whole list after months of segmented sends, a reactivation push to long-inactive addresses, or a migration to a new sending platform all look, from the provider's side, like a change in behaviour. Stage them: split large sends over days, start with engaged segments, and never combine a platform migration with a list import and a new cadence in the same week. Separate transactional mail from marketing mail, on different subdomains where possible, so that a marketing problem never delays an order confirmation; the promotional vs transactional comparison explains why that separation matters beyond deliverability.

When the change is a move to another email platform, the tutorial on migrating to a new email platform loads suppressions before contacts and stages the first sends from the most engaged subscribers.

Monitoring and postmaster tools

Deliverability cannot be managed from the sending platform alone. Google's Postmaster Tools, which its sender guidelines point to, show a verified domain's spam rate, domain and IP reputation, authentication results, encryption and delivery errors as the provider sees them. Verifying the domain there is the first monitoring step for any sender with meaningful Gmail volume. Other providers offer sender programmes and feedback loops; check each provider's own documentation for what it exposes.

DMARC aggregate reports are the second source. They arrive as XML files, which most teams route into a report processor rather than read raw, and they answer a question nothing else does: who is sending as the domain, and does it pass? Unknown sources in the reports mean either an unauthorised sender or a legitimate system nobody configured. Read them weekly during rollout and monthly afterwards.

The third source is the sending platform's own data read over time rather than per send: bounce rates by type and by list source, complaint rates, unsubscribe rates and engagement by mailbox provider. A drop in engagement from one provider while others hold steady is a placement problem at that provider, not a content problem. Seed-list placement tests and a few monitored addresses at each major provider add a direct view, with the caveat that seed addresses do not behave like real subscribers and should be read as a smoke test rather than a measurement. Put the handful of measures that matter into the monthly report as deliverability indicators (KPI), so that the trend is visible before placement collapses.

Diagnosing placement problems

SymptomFirst checkLikely causeFix
Mail rejected or bouncing with an authentication errorSPF, DKIM and DMARC records; alignment with the From domainRecord missing, platform not signing, From domain not alignedCorrect DNS; configure signing on the sender's own domain; align or change the From
Sudden move to spam at one providerThat provider's postmaster view: spam rate and reputationComplaint spike from a recent send or importPause sends to the affected list source; suppress complainers; re-permission or remove the source
Gradual decline in engagement across providersShare of the list that is active; cadence; whole-list sendsInactive addresses diluting engagementIntroduce engagement tiers and suppression; reduce whole-list sends
New domain lands in spam from the first sendSending history; warming planNo reputation and high initial volumeRestart warming from engaged subscribers; step volume gradually
Promotions tab rather than primary inboxContent and sending patternProvider classification of marketing mailAccept it as placement, not failure; focus on relevance and engagement
Unknown sources in DMARC reportsEach source's IP and DKIM domainUnconfigured legitimate system, or spoofingAdd legitimate systems to SPF and DKIM; keep the policy moving towards enforcement
Engagement fine, conversions downLanding pages and tracking rather than deliverabilityNot a placement problemHand to the campaign and analytics review
Transactional mail delayed or filteredWhether it shares infrastructure with marketingMarketing reputation affecting transactionalSeparate streams by subdomain and, where possible, infrastructure
Illustrative example — not a real result

Warming a marketing subdomain for a furniture retailer

A furniture retailer moves marketing mail from its main domain to a new subdomain. The list has 85,000 addresses, of which 31,000 have clicked in the last 90 days. SPF, DKIM and a p=none DMARC record are published and verified in Postmaster Tools before the first send. Day 1 sends 2,000 messages to the most recent clickers; days 2 to 5 double the volume each day while complaints stay under 0.05% and hard bounces under 0.3%. On day 6 the complaint rate reaches 0.12% on a batch that included addresses from a store-event sign-up sheet, so the team holds volume for two days, moves that source to a re-permission segment, and resumes. By day 18 the subdomain is sending 40,000 messages a day with a spam rate below 0.1%, domain reputation shown as high, and DMARC reports showing every message aligned. The policy moves to p=quarantine in week 5 and p=reject in week 9.

When a symptom is already in front of you, the email troubleshooting section starts from it; mail that is accepted but lands in junk is diagnosed cause by cause in emails going to spam.

Because bounced messages reached nobody, measuring email click rate on delivered messages keeps list hygiene problems from masquerading as weak content.

To check deliverability alongside consent, flows and reporting in one pass, score the whole programme with the email programme audit scorecard.

When aggregate reports list sources that fail, diagnosing DMARC failures source by source decides which to fix, which to tolerate as forwarding and which to leave for the policy to refuse.

Whether suppressing inactive subscribers costs more revenue than it saves in complaints can be measured with the inactive subscriber sunset test, which splits the inactive segment between a re-engage-then-suppress policy and the normal cadence.

Common mistakes

  • Reading "delivered" as "in the inbox". Delivery is acceptance by a server; placement is decided afterwards.
  • Authenticating only the marketing platform. Every system that sends as the domain must be in SPF and signing with DKIM, or DMARC enforcement will block it.
  • Jumping straight to reject. Enforcement before the aggregate reports are clean blocks legitimate mail from systems nobody remembered.
  • Treating sender requirements as the goal. They are a floor; the inbox is earned by relevance and consent.
  • Mailing the whole list after months of segmented sends. The volume spike and the inactive addresses arrive together.
  • Never opening postmaster tools. The provider's view is the only one that shows placement; the sending platform cannot.
  • Fixing DNS and changing nothing else. Identity is one of three signal groups; reputation and behaviour decide the rest.
  • Sharing infrastructure between marketing and transactional mail. A promotional problem then delays receipts and password resets.

The fundamentals hub places deliverability between list building and segmentation; most placement problems are solved in one of those two neighbours.

Frequently asked questions

Does a small sender need DMARC?

Google's guidelines document DMARC as a bulk-sender requirement and SPF or DKIM as the requirement for everyone, but a monitoring-only DMARC policy costs nothing and gives any sender aggregate reports showing who is sending as its domain. For a small sender that is the cheapest way to discover a forgotten system or an impersonation attempt. Move towards enforcement once the reports are clean.

Is the promotions tab a deliverability problem?

No. The promotions tab is inbox placement; the provider has classified the mail as marketing, which it is. Subscribers who want the mail find it there, and the answer to low engagement is relevance rather than tricks to avoid the tab, which tend to harm reputation when they work.

How long does warming a new domain take?

There is no documented duration, and providers do not publish one. The sender steps volume up while watching complaints, bounces and reputation, holds when the signals worsen and continues when they recover, so the time depends on the list's quality and the target volume. A clean, engaged list warms in weeks; a neglected one may never warm until its sources are fixed.

Terms used on this page

Where this page is used

Also related

Sources

  1. Email sender guidelines — Google Workspace Admin Help
  2. DMARC overview — DMARC.org
  3. Messaging, Malware and Mobile Anti-Abuse Working Group — M3AAWG
  4. RFC 8058: Signaling One-Click Functionality for List Email Headers — IETF

Related