Spam Trigger Words to Avoid in Cold Email (200+ Word List)

Let's clear up the biggest myth in cold email first. There is no secret list of words that flips a switch and dumps your message in spam. Type "free" into a one-to-one email from a warmed-up inbox and it lands fine. The "spam trigger words list" you've seen passed around since 2009 is mostly folklore from an era when filters worked like a word-counting blocklist. They don't anymore.

So why does this list exist at all? Two reasons. First, rule-based filters like Apache SpamAssassin still score certain phrases, and if your copy stacks enough of them on top of a weak reputation, the math adds up. Second, and this matters more: these words make you sound like a marketer, not a person. That tone is what actually tanks reply rates. A cold email that reads like a billboard gets ignored even when it lands in the primary inbox.

This is the word-level reference. For the bigger deliverability picture (authentication, warmup, list hygiene, complaint rates), our guide on how to avoid spam filters covers the infrastructure side. Here we stay on the copy.

Key Takeaways

• Single words almost never trigger spam. Gmail's filters block more than 99.9% of spam using reputation and machine learning, not keyword matching.
• Rule-based filters like SpamAssassin do score spammy phrases, but a message gets flagged only when many signals accumulate, not from one word.
• The real cost of spam-trigger language is human: it reads like a mass blast and depresses replies.
• Use the 200+ word list below as a writing-hygiene check. Every entry has a safer alternative that sounds more like a real person.

Do spam trigger words still get you filtered?

Short answer: rarely, and almost never on their own. Here's the honest version.

Gmail's defenses "stop more than 99.9% of spam, phishing and malware from reaching inboxes" and "block nearly 15 billion unwanted emails every day," Google reports. That filtering runs on machine learning, not a vocabulary lookup. Google's spam models use frameworks like TensorFlow to catch image-based spam, hidden content, and sketchy new domains, blocking an extra 100 million messages a day on top of the baseline. A model trained on billions of examples doesn't care that you wrote "discount." It cares whether people who get mail from your domain open it, reply to it, or hit the spam button.

Rule-based filters are the part of the myth that's actually real. Apache SpamAssassin runs hundreds of named tests, each with a point score, and plenty of them target specific phrases: MILLION_USD ("talks about millions of dollars"), ACT_NOW_CAPS, MONEY_BACK, URG_BIZ ("contains urgent matter") and many more are real rules in the engine. But the scoring is cumulative. As the project's own documentation puts it, "a message will only be considered as spam if it matches multiple criteria; matching just a single test will not usually be enough to reach the threshold." One trigger word is a rounding error. A wall of them, sent from a cold domain with no reply history, is a problem.

So the practical rule is simple. If your reputation is healthy and your sends are personal, vocabulary barely registers. If your reputation is shaky, every spammy phrase is one more weight on the wrong side of the scale. Either way, cleaning up the language costs you nothing and helps your replies.

Where filters actually draw the line

The numbers that move your placement aren't word counts. They're behavioral. Google tells senders to "keep spam rates reported in Postmaster Tools below 0.10% and avoid ever reaching a spam rate of 0.30% or higher." Cross that complaint line and no amount of careful wording saves you. Stay under it with engaged recipients and you've got room to spare.

Google's guidelines also flag a few content behaviors directly, and these are worth more than any single word: don't "use HTML and CSS to hide content in your messages," and keep "web links in the message body visible and easy to understand." Hidden text and obscured links are mechanical spam signals. A friendly word like "exclusive" is not.

How to use this spam trigger words list

Treat it as a tone check, not a compliance scan. You don't need to purge every word below from existence. You need to notice when your draft is leaning on the language of late-night infomercials.

A useful test: read your email out loud. If it sounds like something a real colleague would write to one person, you're fine. If it sounds like a coupon, start swapping. Three habits cover most of it:

1. Replace claims with specifics. "Guaranteed amazing results" is a trigger phrase and an empty promise. "Cut onboarding time by 40% for teams like yours" is neither.
2. Drop manufactured urgency. Real deadlines have dates. Fake ones ("act now," "last chance") read as pressure and score on rule-based filters.
3. Sound like a person, not a flyer. One ask, plain link, no symbol pileups. Our guides on cold email CTAs and subject line formulas go deeper on the wording that earns clicks instead of triggering filters.

Now the list. It's grouped by the kind of red flag each phrase raises, and every row has a safer alternative. Over 200 words and phrases across seven categories.

The 200+ spam trigger words list, by category

1. Money, free, and financial bait

The single most overused category. Filters and humans both pattern-match this language to scams and mass discounting.

2. Urgency, pressure, and FOMO

Manufactured scarcity is one of the strongest rule-based signals and the fastest way to sound like a hard sell.

3. Guarantees and overpromising

Big promises with no proof. Filters score them; prospects discount them.

4. Hype and superlatives

Empty intensifiers. They add zero information and a lot of marketing smell.

5. Pushy CTAs and retail tone

Cold email has one ask. Stacked, salesy calls to action read like an ad and bury the actual next step.

6. Scam-flavored and too-good-to-be-true

This vocabulary is borrowed straight from lottery and phishing scams. Using it makes you guilty by association.

7. Symbols, caps, and formatting tics

Not words exactly, but the same family of signals. These add mechanical spam scores and look amateurish.

That's 210-plus distinct words and phrases. You won't memorize them, and you don't need to. Internalize the seven categories and your ear will catch the rest.

Rewriting a spammy cold email: before and after

The list is easier to feel than to recite. Here's a draft that hits a dozen triggers, then the same message rewritten.

Before (triggers everywhere):

1Subject: ACT NOW - Limited Time FREE Offer!!!
2
3Dear Friend,
4
5Congratulations! You've been selected for an AMAZING, risk-free
6opportunity to DOUBLE your revenue with our world-class, #1 platform.
7100% guaranteed results or your money back, no catch!
8
9This is a once-in-a-lifetime deal and it won't last. Click here now
10before it's gone. Don't miss out!!!

Count the problems: faked urgency, all caps, triple exclamation marks, "free," "congratulations," "you've been selected," "amazing," "risk-free," "double your revenue," "world-class," "#1," "100% guaranteed," "money back," "no catch," "once-in-a-lifetime," "click here," "don't miss out." It reads like a scam because it borrows scam language.

After (same offer, human tone):

1Subject: cutting onboarding time at [Company]
2
3Hi [First name],
4
5I noticed [Company] is hiring three onboarding specialists this
6quarter. Teams in that spot usually spend 6-8 hours per new hire on
7manual setup.
8
9We helped [Similar company] cut that to about two hours. Happy to
10walk you through how in 15 minutes if it's relevant.
11
12Worth a quick look? If not, no worries and I'll leave it there.
13
14[Your name]

Same pitch, zero trigger words, one clear ask, a real specific, and an easy out. It will land in the inbox, and more importantly, it reads like a person who did their homework. That's what earns the reply. If you want the engagement side of this, our guide on reply rate optimization digs into the structure that turns a clean send into a conversation.

Words aren't your deliverability problem (usually)

Worth saying plainly so you spend effort where it pays. If your cold email is landing in spam, the cause is almost never your vocabulary. It's one of these, roughly in order of how often it's the culprit:

• A weak or unwarmed sender reputation. A brand-new domain with no history gets filtered no matter how clean the copy. Warm it first; see the email warmup guide.
• A dirty list driving complaints and bounces. Spam-trap hits and high complaint rates cross the lines providers actually enforce.
• Missing or broken authentication. No SPF, DKIM, or DMARC alignment, and you're treated as suspicious before a filter reads a single word.
• Volume that spikes too fast. Going from 20 to 500 sends a day reads as a compromised account.

All four outrank wording by a wide margin. For the full triage, run the 10-step deliverability audit, and if you want to understand how Gmail and Outlook actually classify cold mail, beating AI spam filters covers the detection side. The word list is the last 5%, not the first 95%.

That said, the last 5% is free to fix, and the copy cleanup pays off twice: marginally better placement, and meaningfully better replies. Clean language and a good signature signal a real sender. Spammy language signals a blast. Recipients can tell, and increasingly, so can the filters trained on how recipients react.

Common questions

Does using the word "free" send my email to spam?

No. "Free" by itself is harmless in a personal email from a reputable domain. The myth comes from old keyword-based filters. Modern systems weigh reputation and engagement far more than any single word. "Free" only matters if it stacks with many other spammy signals on a weak sender.

How many spam trigger words is too many?

There's no fixed number, because no provider publishes a public word-count threshold. Rule-based filters like SpamAssassin accumulate scores, so a handful of mild words on a healthy sender is fine, while a wall of urgency and hype on a cold domain adds up. The better test is tone: if a draft sounds like an ad, cut until it sounds like a person.

Do spam words matter more in the subject line?

Somewhat, because the subject line is short, visible, and the first thing both filters and humans read. Caps, exclamation marks, and "free" in a subject are more conspicuous than the same words buried in the body. For subject lines that get opened instead of filtered, see our subject line formulas.

Will spinning words with spintax help me avoid filters?

Rotating phrasing across a campaign can reduce the fingerprint of identical mass sends, but it won't rescue a bad reputation or spammy core copy. Use it to vary tone naturally, not to disguise junk. Our spintax guide shows how to do it without producing awkward, robotic sentences.

Is there an official list of spam trigger words from Google or Microsoft?

No. Neither Google nor Microsoft publishes a keyword blocklist, because their filters don't work that way. Google's published guidance focuses on authentication, complaint rates below 0.30%, and not hiding content, not on banned words. Any "official spam words" list you find online is a community-built reference, like this one.

The bottom line

A spam trigger words list is a writing-hygiene tool, not a filter-evasion trick. Modern providers block almost all spam on reputation and behavior, so no single word will sink a well-sent email, and no amount of word-swapping will save a badly sent one. Use the 210-plus entries above to catch the moments your copy slips into infomercial mode, swap in the plainer alternative, and you'll get the small deliverability edge plus the much bigger win: an email that sounds like a person worth replying to.

Sources

• Google, Gmail introduces new requirements to fight spam
• Google Workspace, Ridding Gmail of 100 million more spam messages with TensorFlow
• Google, Email sender guidelines
• Apache SpamAssassin, Tests Performed: v3.3.x
• Wikipedia, Apache SpamAssassin