Email Deliverability Metrics Worth Acting On

The most useful weekly report connects each signal to a decision. Delivery tells the team whether receiving servers accepted messages; inbox placement and downstream actions help reveal whether recipients could see and use them. Complaint rate, hard bounces, and provider-level results deserve attention first because they can expose trouble that an overall delivery number conceals.

A provider split should include the sending domain and stream as well as the mailbox provider. Promotional, lifecycle, and transactional messages can behave differently, so combining them can obscure which traffic needs a fix. Keep the report small enough to review consistently, but preserve the details needed to identify an owner and next step.

Why a Perfect Delivery Rate Still Misses the Inbox

A typical pattern looks like this: the ESP dashboard shows delivery close to 100 percent, while a placement test and provider-level click rates indicate that Gmail is filing a large share of the same campaign into spam. The messages were accepted, but recipients may not see them in the inbox.

The scale of the gap is substantial. Validity's 2025 benchmark found that 83.5 percent of legitimate marketing email reached the inbox worldwide, and Microsoft had the weakest major-provider inbox placement at 75.6 percent. Those figures describe inbox placement, not delivery acceptance, and they show why a green delivery chart can coexist with missed reach and revenue (Validity's 2025 benchmark).

Reconcile evidence by what it measures

ESP delivery logs show whether the receiving server accepted a message.

Provider-level deferrals and blocks identify which mailbox provider is delaying or rejecting traffic.

Seed or inbox-placement tests show where accepted messages appear.

Clicks or replies by provider indicate whether real recipients took action.

Compare Gmail, Microsoft 365, and other major providers separately, then split by domain and sending stream. A sharp decline at one provider is more actionable than a small change in the global average. If the team ignores that split, password-reset complaints and quiet campaign revenue can rise while the delivery chart stays green.

For teams also responsible for outbound programs, this guide to getting outbound emails into inboxes offers a related perspective on inbox reach. The next diagnostic step is to inspect bounce and complaint signals to determine what kind of failure the team is facing.

Hard Bounces, Deferrals, and Complaint Math

A hard bounce is a permanent failure, such as an address that doesn't exist or a recipient rejection that won't clear on retry. A soft bounce means the address is treated as valid, but delivery failed temporarily. A deferral is another temporary failure, often shown as a 4xx response when a provider delays or throttles mail.

Don't treat all failures as one bounce percentage. Keep the receiving server's SMTP status code, enhanced status code, and diagnostic text. An unknown-user response calls for immediate suppression of that address; a provider-specific policy block or deferral calls for investigation of reputation, volume, and authentication rather than blind retries. For a practical process for separating response types, see this guide to managing email bounces.

M3AAWG says hard-bounce rates above 5 percent can indicate problems with list collection or maintenance and recommends active monitoring (M3AAWG's sender complaint-handling guidance). Marketing teams should use the stricter operating line of keeping hard bounces under 2 percent. Break the rate out by provider and list source, or a change in list quality may get mistaken for a reputation block.

Treat complaints as a separate incident signal

Complaints indicate that recipients reported messages as spam, not that an address was invalid. Google updates user-reported spam-rate data daily in Postmaster Tools. Google advises senders to stay below 0.1 percent and avoid reaching 0.3 percent; for bulk senders, a rate above 0.3 percent can make the sender ineligible for delivery mitigations until the rate stays below that threshold for seven consecutive days (Google's spam-rate guidance).

At 0.3 percent, the rate represents 30 complaints per 10,000 delivered messages. Review the date and stream behind a spike rather than smoothing it into a monthly average. Yahoo's bulk-sender expectations also include a complaint rate below 0.3 percent and one-click unsubscribe for promotional mail, so easy unsubscribing is a practical alternative to leaving frustrated recipients with only the spam button.

Common SMTP response	Typical meaning	Classification	First action
`5.1.1` with “user unknown” wording	Recipient address doesn't exist	Hard bounce	Suppress the address immediately
`4.2.2` with “mailbox full” wording	Temporary recipient-side problem	Soft bounce	Retry according to the sending platform's policy
`4.7.1` with “temporarily deferred” wording	Provider delay or policy evaluation	Deferral	Check the provider, volume, and diagnostic text
`5.7.1` with “blocked” or “policy” wording	Rejection tied to provider policy	Permanent rejection	Investigate reputation and authentication before resending

These are example enhanced status codes, not universal response text. Providers use different wording, so copy the exact text from your own logs before writing suppression or retry rules. When list validity is unclear, a tool for checking email addresses before sending can help reduce preventable invalid-address failures.

Metrics That Mislead After Mail Privacy Protection

Open rate is a weak deliverability-health metric. Apple Mail Privacy Protection can preload remote images, including tracking pixels, and Gmail image caching can also weaken the connection between a recorded open and a person reading the message. A high open rate can sit beside poor inbox placement, so a rise in opens doesn't prove that a deliverability fix worked (Stripo's overview of email metrics and privacy).

Use a measurement hierarchy based on the job the metric needs to do:

Reputation: Review provider-level inbox placement, complaint rates, and deferrals. These signals show whether providers are filtering, delaying, or rejecting the mail.

List quality: Track hard bounces and whether recipients have recently engaged. This helps distinguish stale or invalid data from placement trouble.

Business impact: Measure authenticated first-party actions, such as clicks, replies, conversions, product events, or a completed password reset.

There isn't one engagement threshold that fits every stream. A password-reset email can be working well when few recipients click, provided the reset completes. A lifecycle campaign with similarly sparse clicks may indicate a placement or audience problem.

Set a baseline for each stream and provider. Investigate when opens rise but clicks, replies, or conversions fall, or when clicks drop at one provider while other providers hold steady. Mailgun's 2025 State of Email Deliverability report found that only 13 percent of senders used inbox-placement testing to monitor deliverability (Mailgun's 2025 report). Without placement evidence, the open chart often becomes the default health metric. That can lead teams to optimize the wrong thing, continue sending into spam, and notice the revenue gap only after reputation has shifted.

What We Check First When a Metric Moves

A deliverability investigation should follow the evidence, not the most prominent chart. The first three checks are complaint rate, hard bounce rate, and performance split by mailbox provider. Other triggers include a provider-specific rise in deferrals, a new blocklist listing, or an abrupt click decline isolated to one provider.

Check complaint rate first. Gmail and Yahoo enforce complaint expectations directly, so a rate moving toward 0.1 percent warrants investigation. Skipping it can allow recipient feedback to build while the team focuses on a less urgent metric.

Review hard bounces and failure types. A rise after a list change points toward list quality; a temporary error points elsewhere. If the team combines them, it may suppress valid recipients or keep retrying invalid ones.

Split every result by provider, stream, and domain. A Gmail or Microsoft 365 problem can disappear inside the account average. Without this view, the owner may miss the affected traffic until blocks or support tickets rise.

Identify what changed. Compare audience, volume, content, links, and authentication with the last healthy period. If a migration or routing change went unreviewed, a selective failure can persist unnoticed.

Check reputation and blocklist signals. Review Google Postmaster Tools for Gmail-related domain and IP reputation and daily spam-rate data. For Outlook.com sending, Microsoft SNDS provides its own view of volume, complaint, and filter data. A new blocklist listing deserves attention the day it appears, since continuing to send into a block can worsen delivery.

Read the logs and headers

Record the SMTP status, enhanced status code, and the receiving server's diagnostic text. In message headers and authentication reports, examine the DKIM result, SPF result, and DMARC alignment. A pass for an unrelated domain doesn't establish that the visible From domain is authorized.

Google's sender requirements include SPF or DKIM for mail to Gmail personal accounts; bulk senders sending more than 5,000 messages per day to Gmail must also implement SPF, DKIM, and DMARC with an aligned From domain. The requirements also cover valid forward and reverse DNS, TLS, and RFC 5322-compliant formatting (Google's email sender guidelines). After an ESP, return-path, or forwarding change, verify the authentication outcomes separately by provider, domain, and stream. The email authentication overview explains what to review without treating DNS records as a one-time checklist.

For a quick listing check, the free blacklist checker from our sister tool mailX can show whether a domain or IP appears on a DNS blocklist. Tools identify signals; the incident owner still needs to connect the result to a change in sending behavior, audience, or setup. If repeated deferrals are ignored, providers may move from delay to rejection, interrupting both campaigns and transactional messages.

Common Mistakes in Deliverability Reporting

The costliest reporting mistake is averaging complaint rate across a month. Google updates spam-rate information daily, so quiet days can hide a bad day and the provider-specific event that needs attention. A monthly average isn't an incident-control metric.

Other habits leave teams reacting too late:

Combining all bounces: A single percentage without hard, soft, and spam-bounce categories or SMTP text leads to the wrong suppression and retry rules.

Treating authentication as a record-presence check: In the Mailwarm network study, October 2026, Mailwarm, MailAdept's parent company, assessed 12,936 custom sending domains, with legitimate senders and spammers mixed. SPF was missing or unusable for 25 percent, while 45 percent had DMARC at p=none. These figures describe that population; they don't establish a cause of spam placement. A record's presence alone doesn't show whether mail passes and aligns.

Reporting only the account average: A Microsoft 365 deferral can disappear inside a blended number, leaving the affected team without an owner or a timely fix.

Waiting for a reputation dashboard to move: Complaints and provider-level clicks can signal a problem before a reputation view changes. Recovery also takes consistent sending over several weeks, so delayed action can extend lost reach.

The same reporting discipline applies to content checks: a useful spam-words checker can flag terms for review, but it can't replace provider-level evidence or explain a delivery failure by itself. If these habits remain in place, the team is likely to act after a block instead of at the first meaningful provider-level change.

Frequently Asked Questions

Do transactional emails use the same thresholds?

Complaint and bounce lines still matter because providers evaluate the sender, but transactional messages need their own engagement baseline. A password-reset email can be healthy even with few clicks if the reset completes. Measure transactional and promotional streams separately so low click activity on a utility message doesn't trigger the same interpretation as a weak lifecycle campaign.

Do Postmaster Tools and SNDS cover the same data?

No. Google Postmaster Tools reports Gmail-related domain and IP reputation, daily spam-rate information, and authentication signals. Microsoft SNDS provides volume, complaint, and filtering data for Outlook.com. Each view describes a different provider environment, so a healthy Google view doesn't rule out a Microsoft-specific deferral or filtering problem.

How quickly should a blocklist listing be investigated?

Investigate it the day it appears, not at the monthly review. Identify the listing and its scope, then check for changes in volume, list source, and authentication. Pause sending into an affected block until the cause is understood; continuing without a diagnosis can deepen the delivery problem.

When to Bring in an Expert

Outside help is useful when the same provider incident repeats, an ESP, domain, or platform migration is approaching, or a mailbox provider keeps blocking or deferring mail without a clear owner. It's also worth escalating when complaints keep moving toward 0.1 percent despite list and content changes, or when no one can say which stream and metric they own.

A one-time review captures the setup at one moment; ongoing monitoring has to account for changes in volume, audience, and authentication. MailAdept provides a dedicated deliverability expert who joins the team for an audit, fixes, and monitoring with 12+ daily checks per domain and a weekly expert review. Calls with experts are included in every plan, and the service is built by Mailwarm Inc., backed by Y Combinator (S20).

Report complaints, hard bounces, and provider splits before delivery rate, and treat opens as an unreliable health measure. Investigate a complaint rate nearing 0.1 percent, hard bounces at or above 2 percent, a one-provider deferral rise, or a new blocklist listing. Judge recovery through complaint rates and provider-level clicks over several weeks, not a reputation chart alone.

MailAdept pairs ongoing deliverability monitoring by experts with diagnosis and remediation across sending domains and streams. For a second look at complaint trends, bounces, provider splits, and authentication, visit MailAdept and request a free deliverability audit.

Metric	Healthy range	Investigate when	What it actually shows
Delivery rate	Stable against your own baseline at each provider	A sudden drop, or one provider far below the rest	Server acceptance, including messages that land in spam
Inbox placement	Primary inbox at each major provider you send to	Placement tests or provider-level clicks diverge from delivery	Whether people can see the message
Hard bounce rate	Under 2 percent	At or above 2 percent, or a sharp rise after a list change	Invalid addresses and list quality
Soft bounces and deferrals	Low, short-lived, and spread across providers	A sustained rise from one mailbox provider	Temporary failure, throttling, or policy delay
Spam complaint rate	Under 0.3 percent, with a working target under 0.1 percent	Any move toward 0.1 percent, or a daily spike	Recipient feedback that Gmail and Yahoo enforce
Blocklist status	Not listed on any DNS blocklist	Any new listing	A reputation block that can stop acceptance
Authentication pass rate	SPF, DKIM, and DMARC pass and align	Failures after a migration, new ESP, or routing change	Whether the visible From domain is authorized
Open rate	No health target	Opens rise while clicks, replies, or conversions fall	A weak proxy, inflated by privacy protections

Email Deliverability Metrics Worth Acting On