Home · Deliverability

How to check if your emails land in spam (step-by-step)

Suspecting mail is going to spam is common. Guessing without a workflow wastes time. This is a practical sequence: authenticate → clean the list and content → seed-test → fix one variable → re-test. It soft-sells proper placement testing only where it belongs — after the basics.

No fear-mongering, no “never land in spam” promises, and no invented success rates.

1. Confirm it is really a spam problem

Before tools:

• Check bounces and suppressions in your ESP (hard bounces ≠ spam folder).
• Ask a few engaged subscribers on major providers (Gmail, Outlook/Microsoft 365, Yahoo where relevant) to search “All Mail” / junk.
• Compare open/click shifts carefully — privacy changes make opens noisy; treat them as weak signals only.

If mail never arrives anywhere, think infrastructure or blocklisting first, not subject-line tweaks.

2. Authentication first (non-negotiable)

Mailbox providers weight domain trust heavily. Before paying for placement tests:

1. SPF — sending hosts authorised for the domain.
2. DKIM — messages signed; signatures validating.
3. DMARC — policy published; start monitoring (p=none) if you are new, then tighten with care.
4. Alignment — From domain matches what SPF/DKIM authenticate (as your ESP docs describe).

Check the domain itself with the email authentication checker, or start from the record you are fixing: SPF, DMARC, DKIM. A pass there does not promise the inbox.

Skip this step and every spam test will mostly tell you “you look untrustworthy.”

3. List and sending hygiene

• Stop mailing chronic non-responders per a written policy.
• Never buy or scrape lists.
• Honour unsubscribes and complaints quickly.
• Warm new domains/IPs gradually if you are migrating.

A clean test to a poisoned list still looks like spam behaviour downstream.

4. Content and technical sanity (cheap checks)

Use free or built-in checks for:

• Broken HTML, missing plain-text part, huge images-only bodies
• Misleading From names / subject lines
• Working unsubscribe and physical address where required for your jurisdiction
• Link domains that match your brand (not random shorteners on every URL)

These catch silly mistakes. They do not prove multi-provider inbox placement.

5. Seed / inbox placement test

When auth and hygiene are in decent shape, run a seed-list placement test: send the real creative to a panel of addresses across providers and see inbox vs promotions vs spam vs missing.

That is the job class GlockApps and similar tools sit in. Use them when:

• You have a big send coming
• Metrics slipped after a template or ESP change
• You need evidence for a client or boss

How to run a sensible test:

1. Use the same From, subject, and body you will send live.
2. Send from the production domain / stream you care about.
3. Note time, ESP, and any A/B variant.
4. Read inbox / promotions / spam / missing — not just a single “score.”
5. Change one variable, then re-test.

6. Fix, then re-test

Common levers (pick one at a time):

• Auth / DNS mistakes
• List segment (engaged vs rusty)
• From name and subject honesty
• Template weight (tracking scripts, image ratio)
• Link domains and redirect chains
• Sending volume spikes on a new domain

Document what you changed. Without a re-test, you are still guessing.