Why do cold emails go to spam?

Why are my emails going to spam? The general diagnosis

The question is bigger than cold email, and the diagnosis order is the same whether the mail that is going to spam is outreach, a newsletter or plain one-to-one correspondence. Receivers weigh three layers, in this order. Authentication: if SPF, DKIM or DMARC is missing or broken, everything you send starts guilty - check any domain's records in seconds with our deliverability checker. Reputation: the sending domain's history of bounces, complaints and engagement; new domains, sudden volume jumps and shared infrastructure all read as risk, which is why warmup exists and why bounce discipline matters. Content and behaviour: trigger-heavy copy, misleading subjects, image-only bodies and no working opt-out - screen a draft against the 1,130-phrase trigger list or run it through the spam checker.

Two provider notes that answer most "suddenly going to spam" cases: Gmail and Outlook both tightened bulk-sender rules from 2024 onward - authentication became mandatory at volume and complaint thresholds became hard gates - so mail that survived for years on missing DMARC stopped surviving. And a sudden change is almost always a change on your side: a new sending tool, a volume spike, a purchased list, a broken DNS edit. Diagnose in that order before blaming filters; the eight cold-email-specific causes above are the same physics applied to outreach.

The eight causes, in the order receivers weigh them

1. Missing SPF, DKIM or DMARC

Since February 2024, Google and Yahoo require all three for bulk senders, and weigh them for everyone else. A domain missing any one of them is treated as unauthenticated mail. The fix is three DNS records: an SPF TXT record beginning v=spf1 and ending in ~all or -all; a DKIM key published at selector._domainkey.yourdomain by your provider; and a DMARC TXT record at _dmarc.yourdomain, starting at v=DMARC1; p=none; rua=mailto:you@yourdomain so you receive failure reports, then tightening to quarantine and reject once legitimate mail passes alignment. You can confirm all three in about ten seconds with our deliverability checker.

2. An SPF record that exceeds the 10-lookup limit

This is the most common silent failure we find on client domains. SPF permits a maximum of 10 DNS lookups per evaluation (RFC 7208). Every include:, a, mx, ptr, exists and redirect costs one, and an include also counts everything it includes in turn. Past ten, the record returns a permanent error and receivers treat your mail as unauthenticated even though SPF "exists". Nothing in your sending tool will warn you. Count your real lookups, then flatten or remove unused includes; our checker reports the live count.

3. An unwarmed sending domain

A brand-new domain with no sending history has no reputation for a receiver to trust, and volume from it looks like a burst. Warmup takes two to two and a half weeks per mailbox before it carries campaign volume, ramping gradually rather than jumping to a daily target. Attaching a still-warming mailbox to a live campaign is one of the fastest ways to burn a new domain, because the first impression a receiver forms is the one that sticks.

4. A blocklisted sending domain

Blocklists apply to the domain in your envelope, not just the links in your message, and Gmail rejections frequently name the sending domain rather than the recipient's problem. A pattern worth knowing: when an agency spins up many lookalike domains on one obvious prefix (brand1.com, brand2.com, brand3.com), listing one tends to bring the family with it. Check the actual sending domain, not only your website domain, and never build an entire sending fleet on a single recognisable prefix.

5. Unverified lists and bounce rate

Bounce rate is a direct reputation input. Keep it under roughly 2 percent; sustained rates above 5 percent invite throttling and blocks. The trap is trusting a data vendor's own "valid" or "verified" flag: those flags are frequently stale, and a list can pass a vendor check while its mailboxes no longer exist. Re-verify every address live, immediately before sending, and suppress anything that is not a clean pass.

6. Recipient-gateway policy blocks, which look identical to bad lists

Corporate security gateways such as Mimecast, Proofpoint and Barracuda reject mail on their own policy grounds before a human ever sees it. In reporting, those rejections land in the same bucket as invalid addresses, so teams "fix" their list quality when the actual cause was the recipient's gateway. Read the rejection strings: a gateway block names the gateway, and no amount of list cleaning changes it. This distinction decides whether you re-verify a list or change the sending domain and approach.

7. Volume ramped faster than reputation

Provider trust is earned per mailbox, not per campaign. Roughly 10 to 12 sends per mailbox per day on Google Workspace, and around 1 to 2 on Microsoft 365, is a defensible steady state for cold outreach; scale by adding warmed mailboxes and additional provider pools rather than by raising the per-mailbox number. Bulk senders above about 5,000 messages a day to Google or Yahoo also have to hold spam complaints under 0.3 percent and offer one-click unsubscribe.

8. Trigger-word density and formatting

Genuinely the least important item on this list, and the only one most articles cover. A single flagged word in an otherwise normal message is usually harmless; density is what pattern-matches to promotional mail. Our spam checker scans copy against 1,130 trigger phrases grouped into urgency, overpromise, money, salesy and list-mail language, and weights hard classic spam terms far above everyday words for exactly that reason. The same page flags the structural signals that matter more than vocabulary: link count, image references, ALL CAPS and stacked exclamation marks.

Deliverability is a system, not a setting

Every cause above interacts. Perfect authentication on a domain you ramped too quickly still lands in spam; a beautifully warmed domain sending to an unverified list bounces its way into the same place. The order that works in practice is: authenticate first, warm second, verify the list third, ramp fourth, and treat copy last. If you want the current state of your own domain before changing anything, the deliverability checker reads your live DNS and grades each record with the specific fix.

Common questions

How long does it take to recover a domain that is landing in spam?

If the cause is a missing or broken DNS record, the fix is immediate once the record propagates, typically within an hour. If the cause is reputation, expect weeks: you reduce volume, fix the underlying signal, and rebuild slowly. If the domain is blocklisted, request delisting only after fixing what caused it, or move to a clean domain and warm it properly.

Does using a separate domain for cold email actually help?

Yes, and it is the standard practice for a reason: it isolates your primary domain's reputation from outbound experimentation. Your company mail keeps working even if a sending domain gets burned. Use a lookalike domain you own, authenticate it fully, warm it, and keep the volume per mailbox low.

Do open-tracking pixels hurt deliverability?

They add a signal you do not need. Tracking pixels are increasingly blocked or pre-fetched, which both distorts your open metrics and adds a third-party domain to your message. For one-to-one cold outreach, plain text with at most one link filters cleanest.

Can I check all of this without signing up for anything?

Yes. Our deliverability checker reads MX, SPF with its lookup count, DMARC, DKIM selectors, MTA-STS, TLS-RPT and BIMI directly in your browser, and the spam checker and cold email grader analyse copy the same way. Nothing is uploaded or stored, and there is no signup. All of them are listed on our free tools page.