How to avoid spam filters before you send

Avoiding spam filters is not about finding one safe subject line or removing a few words. Mailbox providers judge the sender, authentication, delivery path, audience, message, and sending history together.

The useful goal is to remove preventable risk, send the final message through the real delivery path, and check where it lands before customers depend on it.

Quick answer

Before an important send:

1. Use the same sender domain and email service provider that production will use.
2. Verify SPF, DKIM, and DMARC alignment.
3. Inspect a delivered message's headers and routing path.
4. Confirm the audience expects the email and remove invalid or stale recipients.
5. Check links, HTML, sender details, unsubscribe handling, and message content.
6. Run an inbox placement test across the mailbox providers your customers use.
7. Change one thing at a time, then rerun the same test.

A spam score can help find risks, but it cannot prove that Gmail, Outlook, Yahoo, or a business mailbox will put the message in the inbox.

1. Test the message you will actually send

A clean sample template does not tell you how the finished message will behave. Test the version that includes the real sending domain, tracking links, hosted images, footer, personalization, and unsubscribe controls.

Use a controlled audience or seed list so the test does not reach customers. For transactional email, trigger the same signup, password reset, OTP, billing, or notification flow that the application will use in production.

This matters because the sending path can change the message after it leaves the editor. An email service provider may add tracking redirects, rewrite links, insert a footer, or apply contact-specific content.

2. Verify SPF, DKIM, and DMARC alignment

Authentication should be the first technical check, not the last.

• SPF confirms whether the sending infrastructure is authorized for the domain.
• DKIM confirms that the message was signed and that the signature survived delivery.
• DMARC checks whether the visible sender aligns with the authenticated domains and the domain's published policy.

Passing records are not a guarantee of inbox placement, but failed or misaligned authentication gives mailbox providers a clear reason to distrust the message.

Check the current requirements for every major provider in your audience. Google's email sender guidelines and Yahoo's sender requirements both distinguish baseline requirements from the stricter controls applied to bulk senders.

3. Inspect the delivered message headers

Use the email header analyzer on a message received through the real route. Check:

• Authentication-Results
• From and Return-Path
• the DKIM signing domain
• the Received chain
• the sending IP and provider route
• unexpected forwarding or relay hops

If the visible sender, return path, and signing domain do not tell a consistent story, fix that before rewriting the email body.

4. Protect sender reputation with clean audience practices

Send only to people who asked for the message or whose relationship with your product clearly calls for it. Remove invalid addresses, process bounces, honor unsubscribes, and investigate complaint spikes.

Avoid sudden, unexplained changes in volume. A new domain, new IP, provider migration, or large audience import should be introduced with monitoring and clear sending controls.

There is no universal inbox-rate or spam-folder placement target that is safe for every sender. Compare each stream with its own history and investigate provider-specific changes. Provider-published complaint-rate requirements, such as those from Gmail and Yahoo, are separate controls and should be monitored directly.

5. Check content without chasing spam-word myths

Mailbox providers do not use a public list of forbidden marketing words. A legitimate sale email can reach the inbox, while a plain message can go to spam when its sender or delivery path is untrusted.

Review the message for concrete problems:

• misleading sender names or subject lines
• broken HTML or a missing plain-text alternative
• links that redirect through unfamiliar or low-trust domains
• URL shorteners where the destination is hard to verify
• images that hide essential meaning
• missing unsubscribe or preference controls for marketing email
• a sudden template or tracking-domain change

Run the finished message through the email spam checker and spam score checker. Treat the findings as evidence to investigate, not as a universal pass or fail grade.

6. Test inbox placement through your existing ESP

You do not need to replace HubSpot, Mailchimp, Klaviyo, SendGrid, Amazon SES, or another sending platform to run a placement test. Send the finished message from the existing provider to the controlled seed addresses supplied by the test.

Use the same sender identity and production-like route whenever it is safe to do so. Some ESP test-send modes change the From address or delivery path. When that happens, use the test send for content and rendering checks, then use a controlled campaign or provider-supported seed workflow for placement evidence. Compare the result across Gmail, Google Workspace, Outlook, Microsoft 365, Yahoo, and the other mailbox environments relevant to your audience.

For HubSpot campaigns, the built-in test email is useful for content and rendering, but HubSpot's current documentation says it uses noreply@hubspot.com instead of the campaign's configured From address. Use a controlled send with the connected sending domain when sender reputation and placement are the question. The HubSpot email preview workflow explains the visual-review handoff. Mailchimp and Klaviyo teams can use the Mailchimp device rendering guide and Klaviyo device rendering guide to inspect the finished email visually as well.

Rendering and placement answer different questions. A device preview shows whether the message is readable in iPhone, Android, Gmail, Outlook, and dark mode. An inbox placement test shows whether the provider put it in the inbox, a tab such as Promotions, spam or junk, or nowhere at all.

7. Keep a simple before-and-after record

Record enough evidence to explain the result and reproduce the test:

Evidence	What to record
Sender	From address, return path, signing domain, and sending provider
Message	Template or campaign version and the workflow that triggered it
Authentication	SPF, DKIM, and DMARC results from the delivered message
Placement	Inbox, tab, spam or junk, quarantine, missing, or delayed by provider
Diagnostics	Header, content, link, reputation, and spam-check findings
Next run	One change, the reason for it, and the rerun result

Changing one variable at a time makes the result useful. If you change the sending domain, template, audience, and volume together, you will not know which change affected placement.

Common reasons legitimate email still reaches spam

• authentication passes, but the visible sender is not aligned cleanly
• a provider or routing change introduced a new reputation history
• complaint, bounce, or invalid-recipient rates increased
• marketing and transactional traffic share a sender without clear controls
• tracking links or the template changed abruptly
• the message is accepted by the receiving server but classified as spam afterward

If a live message is already failing, use Why do emails go to spam? for the full investigation sequence.

FAQ

Can a spam score guarantee inbox delivery?

No. A spam score can identify message-level risks, but inbox placement also depends on authentication, reputation, sending behavior, recipient signals, and provider-specific filtering.

Does passing SPF, DKIM, and DMARC keep email out of spam?

It removes important authentication failures, but it does not guarantee placement. Reputation, audience quality, content, routing, and sending history still matter.

Can I test inbox placement with HubSpot or another ESP?

Yes. Send the final test message from the ESP to the controlled seed addresses in the placement test. This lets you inspect the result after the provider has applied its normal message processing.

Should I remove every promotional word?

No. Write honest, expected email and fix concrete content, link, and sender problems. Word substitution alone will not repair broken authentication or a weak sending reputation.

What is the best first test?

Start with the same message and sender path customers will receive. Verify authentication, inspect the headers, run spam diagnostics, and then compare the real placement result across providers.