On this page
Authenticate your domain and clean your list. That's the most common email deliverability advice, and it's incomplete. You can pass authentication, send to working addresses, and still land in spam because recipients don't want the message or your sending patterns look suspicious.
Deliverability isn't a setting you flip on. It's a reputation you build through repeated sending decisions. Google and Yahoo evaluate sender identity, infrastructure, and complaint behavior. Your operations team controls the inputs, including which contacts enter a sequence, how quickly volume increases, and whether an unsubscribe stops the next scheduled message.
Start with email authentication and list hygiene, but connect those controls to the rest of your workflow:
- Infrastructure: Authenticate every sending service and inspect actual message headers.
- Data quality: Verify addresses without confusing deliverability signals with permission.
- Recipient behavior: Honor complaints, unsubscribes, and inactivity.
- Monitoring: Measure inbox placement separately from server acceptance.
For outbound teams, the critical decision happens before a contact reaches the sending tool. A verifier result should inform an eligibility rule, not automatically trigger enrollment. The email deliverability best practices below turn those decisions into an operating system you can inspect, debug, and improve.
Table of Contents
- 1. Authenticate Every Sending Domain with SPF, DKIM, and DMARC
- Validate the message, not just the DNS record
- 2. Verify Email Addresses Before Sending
- 3. Build Lists with Explicit Permission and Double Opt-In
- Keep subscription and prospecting workflows distinct
- 4. Warm Up New Domains and IP Addresses Gradually
- Let observed results control expansion
- 5. Keep Spam Complaint Rates Below Provider Thresholds
- Investigate the source before rewriting the subject line
- 6. Support One-Click Unsubscribe with List-Unsubscribe Headers
- Test the entire suppression path
- 7. Handle Hard and Soft Bounces Correctly
- Route failures to the right owner
- 8. Re-Engage or Sunset Unengaged Subscribers
- Define inactivity around the relationship
- 9. Separate Sending Streams and Keep Sending Patterns Consistent
- Separate operations, not accountability
- 10. Monitor Deliverability with Postmaster Tools and Inbox Placement Tests
- Combine provider telemetry with controlled tests
- 10 Email Deliverability Best Practices at a Glance
- Build Deliverability as a System, Not a One-Time Fix
- Start with setup, then enforce the workflow
- Measure whether the fix changed recipient outcomes
1. Authenticate Every Sending Domain with SPF, DKIM, and DMARC
Authentication establishes identity, not permission or inbox placement. SPF authorizes sending infrastructure, DKIM signs messages, and DMARC checks whether an authenticated identity aligns with the visible From domain. DMARC also lets domain owners publish handling policies and receive aggregate reports.
Gmail's sender requirements require SPF, DKIM, and DMARC for bulk senders sending 5,000 or more messages daily to personal Gmail accounts, along with valid forward and reverse DNS records for sending infrastructure. Don't assume your email service provider has configured everything just because its dashboard shows a verified domain.
Validate the message, not just the DNS record
Send a test from every production system, including HubSpot, your newsletter platform, and application mail. Inspect the received headers for SPF, DKIM, and DMARC results, then confirm the authenticated domains align appropriately.
- Inventory senders: Record which service sends from each domain and who owns its configuration.
- Start with monitoring: Publish a DMARC monitoring policy, inspect aggregate reports, and identify legitimate senders before tightening enforcement.
- Check infrastructure: Confirm PTR records and forward DNS with the infrastructure owner.
- Retest changes: Repeat header inspection after migrations, signing changes, or new integrations.
Say marketing passes authentication but CRM sequences fail DMARC. Fix the CRM's signing or return-path configuration rather than weakening the policy for everyone. Use this guide to diagnosing email deliverability issues to separate authentication failures from reputation problems.
This walkthrough provides a visual reference for authentication setup.
2. Verify Email Addresses Before Sending
A contact enrichment job finishes successfully. That doesn't mean every returned address belongs in a sequence.
Verification is a pre-send risk control, not proof that someone wants your email. Check syntax, domain mail configuration, and available mailbox-deliverability signals. Treat unknown and catch-all results separately from positively verified addresses, and don't assume a verifier can identify every spam trap or abandoned inbox.
For an implementation walkthrough, use this guide on how to verify an email address. The important engineering decision is what your pipeline does with each result.
- Verified: Continue to consent, relevance, freshness, and suppression checks.
- Invalid: Suppress the address rather than repeatedly retrying it.
- Catch-all or unknown: Hold in a separate risk tier instead of treating it as verified.
- Role-based or disposable: Exclude from ordinary person-based outreach unless the use case specifically justifies inclusion.
A valid address answers a technical question. A sendable contact must also pass your policy checks.
For example, an AI SDR workflow can enrich a lead, verify the address, and write the result to the CRM without enrolling the contact. Enrollment happens only after the workflow checks acquisition source, suppression status, and campaign eligibility.
Store the verification timestamp and result category, not just a Boolean field. Recheck older records before reuse, and compare actual bounce outcomes by provider and acquisition source. Otherwise, you won't know whether poor results came from stale records, weak verification, or incorrect integration logic.
3. Build Lists with Explicit Permission and Double Opt-In
Enrichment doesn't create consent. An address can belong to the right person at the right company and still be inappropriate for a newsletter subscription.
For subscription programs, double opt-in adds a useful confirmation step: the person submits an address, receives a confirmation message, and actively confirms the subscription. That introduces signup friction, but it also helps prevent mistyped addresses and unwanted enrollments from becoming recurring sends.

Keep subscription and prospecting workflows distinct
Say someone downloads a technical integration guide. Send the requested asset, but don't interpret that request as unlimited permission for unrelated newsletters and sales sequences.
Record the signup source, timestamp, disclosure text, and subscription scope. Keep the confirmation message short, explain what the subscriber will receive, and avoid burying confirmation beneath promotional content.
- For newsletters: Enroll confirmed subscribers in the topics and cadence they selected.
- For outbound: Assess applicable rules and your lawful basis separately. An opt-out doesn't turn unsolicited outreach into permission-based marketing.
- For unclear historical records: Investigate provenance before sending a re-permission campaign. Don't email an unexplained database merely to ask whether you can email it.
For a sector-specific discussion of consent and campaign governance, SimpleLab Digital's law firm email guide provides useful context. Keep your own eligibility rules explicit enough that an automated agent can't mistake a newly discovered address for a new subscriber.
4. Warm Up New Domains and IP Addresses Gradually
A migration can preserve your templates and contact database while changing the infrastructure receiving servers evaluate. Launching the entire backlog immediately makes it harder to distinguish an authentication defect from a volume or reputation problem.
Warm up with legitimate, expected messages, not manufactured conversations. Begin with recipients who have recently demonstrated interest through meaningful clicks, replies, purchases, or product activity, depending on the stream.
Let observed results control expansion
Don't copy a universal warm-up calendar into production. A new dedicated IP, an established domain moving providers, and a new outreach subdomain need different operating decisions.
- Establish a baseline: Send a controlled batch and inspect authentication, deferrals, bounces, and complaints.
- Expand cautiously: Increase volume only when receiving-provider results remain stable.
- Pause selectively: Slow the affected destination provider when its responses deteriorate instead of unnecessarily stopping healthy traffic.
- Keep rollback ready: Preserve queue controls so you can reduce volume without losing or duplicating messages.
For example, a SaaS team migrating newsletters can move a recently engaged segment first while keeping critical product notifications on the existing infrastructure. If a receiving provider starts deferring messages, the team can investigate before moving the rest of the audience.
A dedicated IP provides control, but it also leaves you responsible for maintaining its sending history. A reputable shared pool can be more practical for irregular traffic. Neither option compensates for unwanted mail, and automated warm-up features don't remove the need to inspect results.
5. Keep Spam Complaint Rates Below Provider Thresholds
Treat complaints as a targeting and expectation problem, not just a dashboard metric. Recipients complain when messages are unexpected, irrelevant, excessive, or difficult to stop.
Yahoo's sender guidance sets a 0.3% complaint-rate ceiling for bulk senders, calculated against messages delivered to the inbox rather than total attempted sends. That denominator matters: dividing complaints by every attempted message can make a troubled campaign look healthier than it is.
For an internal warning level, Validity's 2025 benchmark recommends that best-in-class senders keep spam complaints consistently below 0.1%. Use that as an investigation trigger, not as a quota you can safely consume.

Investigate the source before rewriting the subject line
Suppose complaints rise after an agency adds a new lead source. Break results down by that source, receiving domain, campaign, and record age. Pause the questionable segment while checking its acquisition process and message relevance.
- Ingest feedback: Enable complaint feedback mechanisms where available.
- Suppress promptly: Prevent reported complainants from entering another promotional sequence.
- Check expectations: Compare what the signup or outreach context promised with what you sent.
- Respect measurement differences: Label provider-specific metrics rather than blending incompatible denominators.
Low reported complaints don't prove universal recipient satisfaction. Not every provider supplies equivalent feedback, so combine complaint monitoring with opt-outs, replies, and delivery errors.
6. Support One-Click Unsubscribe with List-Unsubscribe Headers
An unsubscribe link that works only after a login isn't a sensible exit path. Neither is an opt-out that updates the newsletter platform while leaving sales automation untouched.
Implement List-Unsubscribe headers and the associated List-Unsubscribe-Post mechanism for eligible subscription messages. Keep a visible unsubscribe link in the message body too, because recipients shouldn't have to depend on their email client's interface.

Test the entire suppression path
Send a production-template message to a controlled inbox and unsubscribe. Then inspect the underlying suppression record and attempt to enqueue another promotional message through a different tool.
- Handle requests directly: Don't require passwords, surveys, or confirmation screens.
- Make processing idempotent: Repeated requests should preserve the same suppressed state.
- Check queued messages: Cancel pending promotional sends, not just future enrollments.
- Define scope clearly: Separate topic preferences from global marketing opt-outs.
Operational test: An unsubscribe succeeds only when the next unwanted message cannot leave your system.
For example, a contact opts out through a newsletter, then an enrichment sync updates their company title. That update must not reset suppression or trigger a sales sequence.
Keep essential transactional communications governed separately, and don't disguise promotional content as transactional mail to bypass an opt-out. The practical trade-off is straightforward: an easy unsubscribe reduces list size, but retaining unwilling recipients creates avoidable complaint risk.
7. Handle Hard and Soft Bounces Correctly
A nonexistent mailbox and a temporary provider deferral need different responses. Suppress permanent recipient failures immediately. For temporary failures, let your sending provider apply bounded retries and backoff, then inspect recurring patterns.
Don't classify every rejection as a bad address. Authentication failures, policy blocks, and rate limits can affect valid recipients, so preserve the SMTP response and enhanced status code alongside your normalized bounce category.
Route failures to the right owner
Consider a sequence that receives mailbox-not-found responses from one imported list while another campaign receives policy rejections across an entire provider. The first points toward contact quality; the second needs infrastructure or reputation investigation.
- Permanent recipient failure: Suppress the address and attach the original diagnostic.
- Temporary mailbox issue: Retry under a defined policy, then stop repeated unsuccessful attempts.
- Provider-wide deferral: Reduce concurrency or sending rate and inspect the provider response.
- Authentication or policy rejection: Escalate to the sending-infrastructure owner instead of deleting valid contacts.
Your suppression store should be authoritative across CRM workflows, enrichment jobs, and sequence tools. An import must never reactivate a permanently failed address just because another system labels it new.
Make webhook handlers resilient to duplicate and out-of-order events. Otherwise, an old delivery event can overwrite a later hard bounce. Track failures by acquisition source and verification age so you can stop feeding a poor-quality segment before it affects the next campaign.
8. Re-Engage or Sunset Unengaged Subscribers
An address can accept mail and still be a poor sending choice. It may belong to an abandoned inbox, a reassigned role account, or someone who has lost interest.
Mailgun's list-hygiene guidance emphasizes the need to manage stale-but-valid addresses, not just obvious failures. Technical verification can't replace an address-level lifecycle that includes source, permission, activity, and suppression history.
Define inactivity around the relationship
A weekly product newsletter and a seasonal purchasing update shouldn't share the same inactivity rule. Choose an observation window that reflects expected contact frequency and meaningful recipient actions.
- Prioritize deliberate activity: Use relevant clicks, genuine replies, and product actions rather than relying on opens alone.
- Reduce frequency first: Offer a lighter cadence when that fits the subscription.
- Request confirmation carefully: Send a clear re-engagement message only where the existing relationship supports it.
- Sunset nonresponders: Stop routine promotional sending when the recipient doesn't re-engage.
For example, a developer-tools publisher might offer inactive subscribers a choice between release announcements and technical tutorials. Subscribers who make no selection and show no meaningful activity move out of regular campaigns.
Keep the suppression history even when you remove a contact from an active audience.
Deleting the record entirely can let the next CRM import recreate it without context. Preserve enough information to prevent accidental re-enrollment, subject to your retention obligations, and don't let a fresh verification result erase a prior opt-out or inactivity decision.
9. Separate Sending Streams and Keep Sending Patterns Consistent
Password resets shouldn't share the same operational controls as prospecting sequences. They serve different purposes, tolerate different delays, and produce different recipient responses.
Separate transactional, subscription, and outbound traffic using appropriate subdomains, provider configurations, queues, and reporting. This improves diagnosis and limits some shared infrastructure risks, but it doesn't create perfect reputation isolation or authorize unwanted sending.
Separate operations, not accountability
Say a product team launches a webinar campaign while sales imports a prospect list and engineering sends onboarding notifications. Without shared scheduling and suppression rules, each team can independently overload the same audience or infrastructure.
- Give streams owners: Assign responsibility for authentication, volume, suppression, and incident response.
- Smooth the queue: Distribute campaigns instead of releasing the entire backlog simultaneously.
- Keep identity recognizable: Use consistent From names and domains within each stream.
- Coordinate integrations: Make CRM automations and agent-driven workflows obey the same throttling and eligibility controls.
When reviewing creative, see recent marketing campaign work for presentation context, but keep campaign design separate from the infrastructure decision. A polished email still needs the correct sending stream and recipient eligibility.
Dedicated IPs offer more control at the cost of additional reputation management. Subdomains provide organizational separation without guaranteeing that receiving providers will ignore relationships between streams. Don't rotate domains to escape complaints; fix the acquisition, relevance, or frequency problem that caused them.
10. Monitor Deliverability with Postmaster Tools and Inbox Placement Tests
Accepted mail isn't necessarily inboxed mail. A delivery dashboard can look healthy while recipients find the messages in spam.
DMARC Report's summary of Validity's benchmark reports 83.5% global inbox placement in 2024 for legitimate marketing email. That's context, not a performance target for your specific B2B stream.

Combine provider telemetry with controlled tests
Use Google Postmaster Tools for available Gmail reputation, authentication, complaint, and delivery-error information. For relevant Microsoft sending infrastructure, investigate SNDS and JMRP access rather than assuming every domain can be registered identically.
Follow a repeatable process to test email deliverability, particularly after infrastructure or template changes.
- Review provider signals: Compare reputation and errors by destination provider.
- Run seed tests: Inspect placement, headers, rendering, and links in controlled inboxes.
- Log changes: Record DNS edits, sending-volume changes, template releases, and new acquisition sources.
- Connect outcomes: Join delivery events to campaign, source, verification result, and sending stream.
For example, if Outlook placement deteriorates after a migration while Gmail stays stable, inspect Microsoft-specific responses and infrastructure changes before rewriting every template.
Seed tests reveal useful symptoms, but they don't represent every recipient's personalized inbox. Open rates are also imperfect proxies. Use multiple signals, review dashboards weekly and after major changes, and assign someone who can pause the affected queue.
10 Email Deliverability Best Practices at a Glance
| Practice | Primary Goal | Key Metric or Standard | Difficulty | Time to See Impact |
|---|---|---|---|---|
| Authenticate Every Sending Domain with SPF, DKIM, and DMARC | Verify sender identity and protect domain reputation | SPF, DKIM, and DMARC alignment; SPF under 10 DNS lookups | Moderate | Days to weeks |
| Verify Email Addresses Before Sending | Remove invalid or risky addresses before campaigns | Hard bounce rate and verification risk results | Easy to moderate | Before the next send |
| Build Lists with Explicit Permission and Double Opt-In | Grow an engaged, consent-based list | Confirmed opt-ins and documented consent | Moderate | Over several campaigns |
| Warm Up New Domains and IP Addresses Gradually | Build a positive reputation for new sending infrastructure | Gradual volume increases; monitor bounces, complaints, and placement | Moderate | Four to eight weeks |
| Keep Spam Complaint Rates Below Provider Thresholds | Protect inbox placement by reducing spam reports | Stay below 0.3%; aim for near 0.1% | Moderate | Weeks, depending on send volume |
| Support One-Click Unsubscribe with List-Unsubscribe Headers | Make opting out easy and reduce spam complaints | List-Unsubscribe and RFC 8058 one-click support | Easy to moderate | Immediate after implementation |
| Handle Hard and Soft Bounces Correctly | Suppress failed addresses and protect sender reputation | Suppress hard bounces; retry soft bounces a limited number of times | Moderate | Immediate to a few days |
| Re-Engage or Sunset Unengaged Subscribers | Focus sending on active, interested recipients | Inactivity window, typically 90 to 180 days | Moderate | Several weeks |
| Separate Sending Streams and Keep Sending Patterns Consistent | Isolate reputation risk and avoid sudden sending spikes | Separate streams and maintain a steady cadence | Moderate to difficult | Weeks to months |
| Monitor Deliverability with Postmaster Tools and Inbox Placement Tests | Detect reputation and inbox placement issues early | Provider reputation dashboards, spam rates, and seed tests | Moderate | Days to weeks |
Build Deliverability as a System, Not a One-Time Fix
No single configuration saves an email program from unwanted sending. Authentication establishes identity, verification reduces recipient-address risk, suppression honors recipient decisions, and monitoring shows where your assumptions failed.
The system works only when those controls share state. A verifier can correctly reject an address while an old CRM sequence keeps mailing it. An unsubscribe endpoint can return success while a queued follow-up still leaves the sending platform.
Start with setup, then enforce the workflow
Complete the foundational work first: authenticate each sending service, validate alignment, inspect DNS, and implement working unsubscribe headers. Test real messages from production integrations rather than relying exclusively on configuration screens.
Then build repeatable operating habits:
- Before enrollment: Check verification status, freshness, source, relevance, and permission or applicable policy.
- Before dispatch: Recheck suppression and stream eligibility, including messages already queued.
- After sending: Process bounces, complaints, opt-outs, and replies into the same contact lifecycle.
- During review: Compare provider-level results with acquisition sources and recent system changes.
For outbound pipelines, the most useful boundary is between data availability and sending authorization. A successful enrichment lookup means you have information to evaluate. It shouldn't mean the contact immediately enters a campaign.
Say an agency adds a new contact provider to its waterfall. Before enabling automatic enrollment, store the source and verification result, test the mapping into the CRM, and confirm that suppressed contacts remain suppressed after updates. Release a controlled segment and inspect actual outcomes before expanding use.
Measure whether the fix changed recipient outcomes
Choose a weak area you can own this week. If authentication is inconsistent, inventory senders and inspect headers. If complaints cluster around a lead source, stop that source and examine how the contacts were acquired.
If hard bounces recur, trace the path from lookup to verification to enrollment. Check whether your code treats an HTTP success response as a usable contact, ignores unknown verification results, or allows stale records to bypass the pre-send gate.
Don't judge the repair only by whether the dashboard turns green. Look for fewer unwanted sends, correctly applied suppressions, healthier provider responses, and improved placement in controlled tests. Keep a change log so the next operator can connect outcomes to decisions.
Deliverability improves when your team makes fewer bad sending decisions and catches failures before scaling them. Pick the weakest handoff, assign an owner, implement its control, and measure the resulting placement and delivery signals.
Our Email Finder and Email Verifier APIs help you find and verify work email addresses before they enter sending workflows. Visit RichAPI to connect enrichment and verification to your existing GTM pipeline through REST or MCP, while keeping permission and suppression checks in your own sending logic.
