Your emails are hitting spam folders, and you have no idea why. You're following the basics - you have a domain, you're sending from a real email address, your copy looks legitimate. But every day, a chunk of your sends disappear into the void, and your reply rates suffer for it.

The problem isn't usually one big thing. It's the accumulation of small signals that tell Gmail, Outlook, and other providers: "This sender is probably spam." Understanding how inbox filtering actually works - and more importantly, how to avoid triggering it - is the difference between campaigns that generate replies and campaigns that generate nothing.

This isn't theoretical. It's the actual criteria email providers use to decide where your message lands.

How Modern Email Filters Actually Work

Email providers use a combination of automated systems and machine learning to filter messages. Here's the hierarchy of what they check:

Authentication checks come first. Your email won't pass the initial gates if SPF, DKIM, and DMARC aren't set up correctly. These are non-negotiable. If you skip this, you're starting at a 40% deliverability disadvantage before anything else matters. We've covered cold email infrastructure setup in detail elsewhere, but understand this: authentication is the baseline.

Content analysis happens next. Filters scan your email for spam signals - certain words, formatting patterns, link density, images. But here's what most people miss: modern filters aren't looking for obviously bad copy. They're looking for patterns that legitimate business emails don't follow.

Sender reputation sits underneath everything. If your IP or domain has a history of spam complaints, bounces, or poor engagement, the content of your email almost doesn't matter. Sender reputation is built over time through consistent, clean sending practices.

User behavior seals the deal. Gmail and Outlook watch what recipients do with your emails. If people mark them as spam instead of deleting them, that's a strong signal. If they move them to a folder or ignore them consistently, that also matters. If they reply and engage, your reputation goes up.

The Email Content Filters That Actually Catch People

Your copy is being scanned automatically. Here are the specific things that trigger filters:

Spam trigger words. Yes, they still exist. Words like "free," "limited time," "act now," "winner," "click here" are red flags. But the bigger issue is overuse of capitalization, excessive punctuation (!!!), and urgent language patterns. One exclamation mark is fine. Three looks like spam.

Link density. If more than 15-20% of your email is links, filters get suspicious. A cold email should have 1-2 links maximum, usually just the CTA at the end. Multiple links scattered throughout looks like you're trying to get clicks for the sake of clicks.

URL formatting. Shortened links (bit.ly, tinyurl) are filtered more aggressively than full URLs. Use your actual domain or a branded short link tied to your domain. Also, never use link redirects in your signature - just your actual website URL.

Image-to-text ratio. If your email is mostly images, filters treat it like spam. Cold emails should be plain text or text-heavy HTML. No banner images, no embedded graphics that carry your message.

HTML vs. plain text. This is less obvious. Overly complex HTML with lots of styling, hidden text, or formatting tricks triggers filters. Simple, clean HTML - or just plain text - performs better. Most cold emails should be plain text or very minimal HTML.

Here's what a clean cold email looks like technically:

Hi [Name], Quick thought - saw you manage the ops side at [Company]. Most ops teams we work with spend 6-8 hours a month on manual reporting. We built a tool that cuts that to 15 minutes. Works with your existing stack. Worth a conversation? [Your Name] [Title] [Phone] [Website]

That's it. No HTML formatting, no images, no links until the CTA. Plain text, short paragraphs, one ask. Filters like this. Recipients like this.

SPF, DKIM, and DMARC - The Three Things That Matter

You need all three. Not one. Not two. All three.

SPF (Sender Policy Framework): This tells email providers which servers are allowed to send emails from your domain. You add an SPF record to your DNS. It should look something like: v=spf1 include:sendgrid.net ~all (depending on your email service).

If you're sending from multiple services, you need multiple includes. Gmail checks this first. If it fails, you're getting blocked immediately.

DKIM (DomainKeys Identified Mail): This cryptographically signs your emails so providers can verify they actually came from you. Your email service generates the DKIM record, you add it to DNS. Without it, your emails look spoofed, even if they're legitimate.

DMARC (Domain-based Message Authentication, Reporting and Conformance): This tells providers what to do if an email fails SPF or DKIM checks. Set it to p=quarantine at minimum - this sends failed emails to spam instead of rejecting them outright (which helps you debug issues). The record looks like: v=DMARC1; p=quarantine; rua=mailto:[email protected]

Check these are actually working: Go to mxtoolbox.com, search for your domain, check DKIM and SPF records. If either shows red, you have a problem. Fix it before sending at scale.

Volume and Warmup Matter More Than Copy

Here's what kills campaigns: sending 500 emails from a brand new domain on day one.

Email providers have warmup algorithms built in. A new sending IP or domain is treated like an unknown actor. If you dump a huge volume immediately, filters assume you're a spam operation. Legitimate senders ramp up slowly.

The right approach: start with 20-30 emails per day for the first week. Move to 50-75 per day in week two. By week three or four, you can hit 100-150 per day depending on your domain age and reputation history.

This isn't arbitrary. Gmail and Outlook literally observe your sending patterns. Sudden volume spikes from new senders trigger automated holds and scrutiny. Gradual, consistent volume looks legitimate.

Bounce Rates and List Quality Are Filter Signals

When you send to bad email addresses, providers notice. High bounce rates - anything above 2% - tell filters you're using poor quality lists or scraping addresses. This hurts your domain reputation going forward.

Before you send any cold email campaign, clean your list. Remove obvious patterns: generic addresses like info@ or support@, free email domains if you're B2B, anything that bounced on a previous send. List cleaning takes 30 minutes and saves your sender reputation for months.

Benchmark: aim for a bounce rate below 1.5%. If you're hitting 3%+, your list is dirty. Stop sending and clean it.

Unsubscribe Rate and Spam Complaint Rate Are Real Metrics

This one is simple but people ignore it. If more than 0.5% of recipients mark your email as spam, providers start filtering your future sends. If more than 5% unsubscribe from a campaign, that's a signal your targeting is off or your copy is misleading.

If you're seeing spam complaint rates above 0.5%, the problem is usually targeting (you're emailing the wrong people) or subject lines (you're overpromising and disappointing).

Here's the email formula that minimizes both:

Subject: [Name] - [Company] ops question Hi [Name], We work with ops leaders at companies like [Similar Company] on streamlining reporting workflows. Looking at your site and the size of your team, thought there might be a fit. Worth chatting about? [Your Name]

This is direct, specific, and doesn't oversell. Low spam complaint rate because there's no clickbait in the subject line. Low unsubscribe rate because you've narrowed to people who actually care.

The Gap Between Knowing This and Running It Well

You now understand the mechanics. Authentication setup, content signals, list quality, warmup pacing, reputation monitoring - it's all actionable. But the gap between "I understand inbox filtering" and "my cold email campaigns consistently hit inboxes and generate replies" is real.

Most teams don't have someone dedicated to monitoring sender reputation data, adjusting warmup schedules based on bounce rates, or testing subject line patterns to keep complaint rates low. And they definitely don't want to build and manage email infrastructure from scratch. That's where having a cold email operation run by people who do this at scale makes sense - they handle the infrastructure, the list quality checks, the warmup pacing, and the ongoing reputation management so your campaigns actually land.

Related Guides