Why emails go to spam, and how to find out which reason applies

There is no single spam score that decides where a message lands. Each mailbox provider combines dozens of signals, and most of them are about you as a sender rather than about one message. The good news: the common causes are few, and you can check most of them yourself.

Start with the evidence. Find a message that went to spam, open its full headers (in Gmail: ⋮ → “Show original”), and paste them into the email header analyzer. The receiving server records its authentication verdict there, and some providers record a spam verdict too. Then go through the causes below in order.

1. Authentication fails or isn’t aligned

Look at the Authentication-Results header added by the receiving provider:

Authentication-Results: mx.google.com;
       dkim=pass header.i=@example.com header.s=s2026;
       spf=pass smtp.mailfrom=bounce@mail.example.com;
       dmarc=pass (p=QUARANTINE sp=QUARANTINE dis=NONE) header.from=example.com

You want pass for all three. Typical failures:

• spf=fail or softfail — the sending service is missing from your SPF record, or the record has more than 10 DNS lookups (SPF checker).
• dkim=none — the message has no DKIM signature: the service doesn’t sign, or DKIM isn’t enabled for your account.
• dmarc=fail although SPF and DKIM passed — they passed for the service’s domain, not yours. Enable custom DKIM in the service; see SPF vs DKIM vs DMARC.

Since 2024 Gmail and Yahoo — and since 2025 Microsoft for Outlook.com — expect SPF, DKIM and DMARC from bulk senders, and treat unauthenticated mail with growing suspicion.

2. People mark your mail as spam

User complaints are the strongest negative signal. Google requires every sender to keep the spam rate reported in Postmaster Tools below 0.3% and recommends staying below 0.1% — one complaint per thousand messages.

Complaints usually mean recipients don’t remember signing up, get more mail than they expected, or can’t find the unsubscribe link. For marketing mail, Gmail and Yahoo require one-click unsubscribe (RFC 8058):

List-Unsubscribe: <https://example.com/unsubscribe?u=abc123>, <mailto:unsubscribe@example.com>
List-Unsubscribe-Post: List-Unsubscribe=One-Click

Making unsubscribing easy reduces complaints; hiding it increases them.

3. Your list contains bad addresses

Sending to addresses that don’t exist produces bounces; sending to abandoned addresses that providers have turned into spam traps damages your reputation quickly. Bought, scraped or very old lists are the usual source.

• Use double opt-in for sign-ups.
• Remove addresses that hard-bounce immediately.
• Stop mailing people who haven’t opened anything in many months.
• Validate the syntax at the point of entry and catch typos such as gmial.com — the email syntax validator does this. (Syntax checks catch typos, not dead mailboxes.)

4. Your sending reputation is new, low or damaged

Providers track reputation per domain and per IP address. A brand-new domain, a new dedicated IP or a sudden jump in volume all look like a spammer’s pattern. Increase volume gradually over a few weeks (“warm-up”) and send to your most engaged recipients first.

Shared IP addresses at email services carry the reputation of other customers too. If deliverability dropped without changes on your side, ask your provider.

Check your standing where the providers publish it: Google Postmaster Tools (spam rate and compliance status for Gmail) and Microsoft SNDS (IP data for Outlook.com). Public blocklists such as Spamhaus show whether an IP or domain is listed and why; the IP blocklist checker queries the lists whose terms allow public lookups and explains each listing.

5. Infrastructure details are missing

These are requirements rather than optimizations:

• Reverse DNS: the sending IP must have a PTR record whose name resolves back to the same IP. Gmail requires this from all senders.
• TLS: Gmail requires encrypted connections from all senders.
• A proper HELO/EHLO name: a fully qualified domain name that resolves in DNS, as RFC 5321 expects.
• Valid message format: one From header, a Date header, a Message-ID. The header analyzer flags missing ones.

Email services take care of these for you; they matter when you send from your own servers.

6. The content looks like spam

Content matters less than it used to, but some patterns still hurt:

• link shorteners, or link text that shows one domain and points to another;
• messages that are a single image with little text;
• attachments recipients didn’t expect, especially archives and macros;
• a From name, subject or body that doesn’t match what people signed up for.

Write like a person, use your own domain in links, and keep HTML simple.

A diagnostic order that works

1. Headers of a spam-foldered message → authentication, alignment and any spam verdict (header analyzer).
2. DNS → SPF, DKIM and DMARC for the From domain.
3. Provider dashboards → Postmaster Tools spam rate and compliance status, SNDS for your IPs.
4. List and sending practices → bounces, complaints, sign-up flow, unsubscribe headers.
5. Content → only after the above are clean.

If the problem affects a single recipient, also check the recipient’s own filters and their organization’s gateway — mail that never arrives at all is a different problem, often in the recipient’s MX setup.