Why your cold emails land in spam (and the 15-minute fix)

The typical SDR runs this loop. Write sequence. Import list into Apollo, Lemlist, Instantly or Smartlead. Send. Check stats the following week. 0.8% reply rate. Blame the copy. Rewrite. Send again. 0.7% reply rate. Blame the ICP. Rewrite the ICP. Send again. 0.9%. Blame the tool.

The problem is almost never the copy or the ICP. The problem is that 70%+ of those sends are landing in Junk. Three fixes, in the order below, typically move the inbox rate from 20% to 75%+ within a week. Total time: fifteen minutes if your DNS admin is responsive, thirty if not.

Fix 1: authentication (the 10-minute SPF/DKIM/DMARC check)

Cold emails from a properly authenticated sender land in the inbox at 70%+ rates on Gmail. Cold emails from a sender with broken or missing authentication land at 15–25%. This is the single largest lever on the list.

The 10-minute routine:

1. Send a test message from your cold-outreach sending address to check-auth@verifier.port25.com. Read the reply. It tells you SPF result, DKIM result, DMARC alignment, and any specific failure.
2. If SPF fails: check the TXT record on your sending domain apex. It should include every ESP and transactional service. Watch the 10-lookup limit — include: chains can silently break SPF past ten lookups.
3. If DKIM fails: enable DKIM in your ESP admin and publish the provided selector record. Google Workspace, Microsoft 365, SendGrid, Postmark and Resend all offer one-click DKIM with a copy-paste DNS record.
4. If DMARC fails alignment: publish v=DMARC1; p=none; rua=mailto:reports@yourdomain.com at _dmarc.yourdomain.com, read the reports for two weeks, then move to p=quarantine.

80% of cold-outreach teams have one of these three broken right now. Fixing them takes ten minutes and typically moves inbox rate 20–40 percentage points on its own.

Fix 2: stop sending cold from your brand domain

This is the second-biggest fix, and it takes five minutes plus DNS propagation. A fresh sending domain — something like send.yourbrand.com, yourbrand-mail.com, or go.yourbrand.com — carries its own reputation separate from your main brand. A reputation hit on cold outreach doesn't poison transactional mail, billing notifications or your newsletter.

Beyond isolation, receiving filters also treat subdomain senders differently. Gmail weights patterns like "real person from a subdomain that looks like a dedicated sending surface" more favourably than "random SDR using the company apex".

Setup: register a new sending domain (not the apex), publish SPF / DKIM / DMARC on it, run the 3–4 week warm-up (see the related warm-up guide), then move cold outreach to it. Your brand apex continues to handle transactional and marketing mail untouched.

Fix 3: remove the three generic-template patterns

Content filters at Gmail, Outlook and Mail.ru all have weights against patterns that appear in millions of cold outreach sequences. If your template contains any of the following verbatim, you are paying a content tax on every send:

• "I hope this email finds you well" — or any variant. Every spam-detection corpus trained on public cold-outreach datasets learned this phrase as a near-perfect spam signal.
• "Quick question" in subject lines. Used in roughly one in four cold-outreach templates. Has measurable negative weight on Gmail and Outlook.
• "5 minutes of your time", "just 15 minutes", "brief call" — the "small ask" cliché. These phrases appear across SDR playbooks and carry weight on both content filters and human reply rates.

Removing these three patterns typically moves Gmail inbox rate 5–10 percentage points and, separately, tends to improve reply rate among humans who get tired of seeing the same opener every day. Replace with direct, specific openers that reference something real about the recipient.

Why these three matter more than everything else

Most deliverability advice for cold outreach is long and expensive. The three fixes above cover 70–80% of the gap between "cold emails land in spam" and "cold emails land in inbox". The remaining 20–30% — list hygiene, tracking pixels, link counts, warm-up curves — matters at the margin. But none of the margin fixes matter if the foundational three are broken.

The 15-minute routine, step by step

1. Minute 0–5: send test to check-auth@verifier.port25.com. Record the three results.
2. Minute 5–10: fix any failing authentication record. Republish TXT records as needed.
3. Minute 10–15: open your cold-email template. Delete "hope this finds you well", rewrite the subject if it contains "quick question", rewrite the ask if it asks for "5 minutes of your time".
4. Optional, +24 hours: register a dedicated sending domain, set up DNS, start the warm-up.

Before and after numbers

A typical before/after for teams running this fix on a live sequence:

• Inbox rate: 18–25% before, 70–80% after.
• Reply rate: 0.6–1.0% before, 2.5–4.5% after.
• Bounce rate: usually unchanged (list hygiene is separate).
• Unsubscribe rate: unchanged or slightly lower.

The reply-rate lift is partly mechanical (more mail reaches the inbox, so more mail can be replied to) and partly content (the removed phrases are tiresome to read).

Secondary fixes worth doing

• Remove the tracking pixel on first contact. Pixels correlate with ~12% lower Gmail inboxing on cold outreach. Tracking opens is useless for MPP-protected recipients anyway.
• Keep link count low. One link in a cold email. Two maximum. More than two measurably hurts inbox rate.
• Fill the plain-text alternative. Multipart messages with empty plain-text parts look like spammer output.

When 15 minutes isn't enough

If after doing the three fixes your inbox rate is still below 50%, the likely remaining causes are: (1) a shared sending IP on a low-tier ESP with bad neighbours, (2) a brand-new sending domain that hasn't been warmed, or (3) a tracking-domain CNAME that points at a listed host. All three are solvable, but they take days, not minutes.

Frequently asked questions