Spam Trigger Words: How Filters Score Them

What are spam trigger words?

Spam trigger words are words and phrases that turn up far more often in unwanted mail than in normal mail, so filters treat them as a weak signal. Typical examples are money promises ("cash bonus", "double your income"), urgency ("act now", "limited time"), guarantees ("risk-free", "no obligation") and prize language ("you've been selected", "winner"). The name makes them sound like a blocklist, but they aren't one. No mailbox provider publishes a list of banned words. The lists you'll find online come from email vendors who compiled them from what they've seen in practice, and they overlap heavily because they copy from each other. Treat a trigger word as one point against a message, not a verdict. A cold email to one named prospect that says "free" once in a normal sentence reads differently from a bulk blast that repeats "FREE!!!" in the subject, the preheader and three buttons. Filters that score content pick up exactly that difference in density and formatting.

How do spam filters actually use trigger words?

There are two kinds of filter, and they handle words differently. Rule-based filters like Apache SpamAssassin run a message through many tests. Each test that matches adds or subtracts points. Some match phrases in the body or subject, some check formatting, and some check headers and authentication. The Apache SpamAssassin documentation sets the default threshold, required_score, at 5.0. Above it, the message is tagged as spam. A trigger phrase usually adds a fraction of that, so a message only crosses the line when several weak signals pile up. Large mailbox providers such as Gmail and Microsoft run their own machine-learned systems. These don't publish rules, and they lean mainly on sender signals: whether SPF, DKIM and DMARC pass and align with your From domain, how that domain's past mail was received, how often recipients mark it as spam and how many addresses bounce. That's why scrubbing every listed word won't help much if your domain fails authentication or your list is full of dead addresses. The full list of causes is in our guide on why emails go to spam and how to fix them.

What doesn't work: swapping letters for numbers or symbols ("fr3e", "$ave") to slip past a word match. Filters treat that kind of obfuscation as a spam signal in its own right, and it costs you more than the plain word would.

Top spam trigger words by category

Most of the published lists group into a few categories. What follows is a list of patterns, not a ranking by measured impact. No provider publishes figures that would support a ranking.

Category	Typical phrases	Why filters count it	Plainer alternative in B2B outreach
Money and earnings	cash bonus, double your income, make money, earn extra	Common in scam and get-rich mail	Name the business outcome in the prospect's terms, e.g. the process you'd shorten
Urgency	act now, urgent, limited time, expires today, don't miss out	Pressure tactics are typical of bulk promotions	Give a concrete reason to reply, or a date tied to something real
Guarantees	guaranteed, risk-free, no obligation, satisfaction guaranteed	Unverifiable promises show up in fraudulent offers	Describe what the product does and leave the promise out
Free and discounts	free, free trial, percent off, lowest price, best price	Heavily used in mass retail mail	Mention a trial once, in a sentence, not in caps or a button
Prizes and selection	winner, you've been selected, congratulations, claim your	Classic lottery and phishing wording	Say why you picked this recipient, specifically
Account and security bait	verify your account, password, unusual activity, confirm now	Phishing imitates these notices	Leave these out of sales mail entirely

Email spam trigger words in the body: what else gets scored

In the body, the words themselves count for less than how the message is built around them. These patterns add points in content-scoring filters, whatever words you use:

• Whole sentences in capitals, or stacked exclamation marks. Shouting formatting is scored on its own.
• Image-only emails or a large image with very little text. The filter can't read the image, and spam often hides its text that way.
• Link text that names one domain while the link points to another. That's how phishing works, and filters compare the two.
• Public link shorteners. They hide the destination, and spammers use them to rotate domains.
• HTML with no plain-text part, or broken HTML from pasting out of a word processor.
• Hidden text, such as white text on a white background, used to pad a message with neutral words.

In a one-to-one cold email, the safe default is short plain sentences, one link at most, a real signature and a working way to opt out.

Subject line spam trigger words

The subject line gets checked by content rules and read by a person who decides in a second whether to open, ignore or report the message. In SpamAssassin, some rules look only at the subject header, so an all-caps subject or one full of punctuation scores even if the body is clean. The words that do the most harm in subject lines are the ones that misstate the message. A fake "Re:" or "Fwd:" on a first email, "Your invoice", or "Following up on our call" when there was no call. These may get past a word filter, but recipients who feel tricked hit the spam button, and providers weigh complaints far more than vocabulary. In the US, the CAN-SPAM Act explicitly prohibits deceptive subject lines. Keep subjects short, lower-case where natural and specific to the recipient. Leave out money signs, "free" and urgency words. Our article on cold email subject line examples and rules has copy-ready subject lines.

Is there a spam trigger words list for 2024 or later?

Lists titled with a year, like "spam trigger words 2024", are mostly the same vendor lists with a new date on them. No mailbox provider issued a new word list that year. What did change in 2024 were the sender requirements Google and Yahoo started enforcing for bulk senders. According to Google's Email sender guidelines, these cover authentication with SPF and DKIM, a DMARC record, one-click unsubscribe for bulk mail, and keeping the reported spam rate low. Those rules sit on the sender side, not the word side. If you update anything because of a new year's list, update your DNS records and your email list hygiene first. Our guide to DKIM, DMARC and SPF for cold email senders covers the setup order.

How to check a cold email for spam signals before sending

1. 1 Read the draft aloud and cut every sentence that would also fit a mass promotion: urgency, guarantees, prizes, money claims.
2. 2 Send the draft to a test inbox at Gmail and open "Show original". Check that SPF, DKIM and DMARC all show PASS and that DKIM is signed by your own domain, not someone else's.
3. 3 If you have SpamAssassin installed, run the saved message through it with spamassassin -t. Read the X-Spam-Status header and report, which list the score and every rule that matched. Fix the rules you can control.
4. 4 Check that every link points to the domain its text names, and that the message has a plain-text part.
5. 5 Check your DNS records with an SPF, DKIM and DMARC checker before the first real send.

A low score in a test doesn't predict inbox placement. The receiving server decides that, using reputation data the test can't see. Use the free SPF, DKIM and DMARC checker for the DNS step.

Where EmailFlow OS fits in

EmailFlow OS is built for B2B cold outreach with personalised AI video. You record one video and upload it to EmailFlow OS. Then you connect your Brevo account and your sending domain and import your prospect list into EmailFlow OS. EmailFlow OS turns that single recording into a separate personalised version for each recipient, not one per company. Each version goes out from your own inbox, through your own Brevo account and your own sending domain, never through a shared platform sender. So the reputation a filter weighs is your domain's, and the DNS records it checks are the ones you control. That doesn't make wording irrelevant. The text around the video still goes through the same content checks, so the rules above still apply. You can read how personalised video email from your own Brevo account works on the EmailFlow OS home page.

Questions people ask

? Does the word "free" send emails to spam?

Not on its own. In a scoring filter like Apache SpamAssassin, a single match adds part of the points needed to reach the default 5.0 threshold set in its documentation. "Free" becomes a problem when it's repeated, set in capitals, placed in the subject, or combined with other signals such as failed DKIM or a poor domain reputation.

? Do Gmail and Outlook use spam word lists?

Neither publishes one. Both run machine-learned filters that weigh sender signals, such as SPF, DKIM and DMARC alignment, domain history and complaint rates, alongside the content. Vendor word lists approximate the content side only, so they can't explain a spam-folder problem caused by authentication or reputation.

? What is a good SpamAssassin score for a cold email?

Lower is better. Anything at or above required_score, which defaults to 5.0 according to the Apache SpamAssassin documentation, gets tagged as spam by servers that keep the default. A clean one-to-one email usually lands well below that. Read the list of matched rules, not just the total, because each rule tells you what to change.