Email Spam Trigger Words: Do They Still Matter for Deliverability?

Search for "spam words" and you will find lists of hundreds of terms with the promise that avoiding them guarantees the inbox. It does not. Yet ignoring the topic entirely is also a mistake, because wording still influences both automated filters and the human who decides whether to click "Report spam". Here is a realistic picture, with the numbers that are actually documented.

How spam filters actually decide

A mailbox provider combines many signals into one decision. The weights below are qualitative: providers do not publish them, but their own guidance consistently puts reputation, authentication and complaints first.

Signal	Weight	Examples
Sender reputation	Very high	Domain and IP history, complaint rate, spam-trap hits
Authentication	Very high	SPF, DKIM and DMARC alignment
Recipient engagement	High	Opens, replies, moves out of spam, deletions unread
List quality	High	Bounces, invalid and role addresses
Content patterns	Low to medium	Link and image ratios, deceptive formatting, certain phrases

Two filtering philosophies coexist. Provider-scale filters (Gmail, Yahoo, Outlook) rely heavily on machine learning over whole messages and on user feedback. Rule-based filters that companies run on their own mail servers, such as Apache SpamAssassin, assign a score to each rule that matches and classify a message as spam when the total reaches a configurable threshold. SpamAssassin's documentation states that the default threshold, required_score, is 5.0 and calls that setting "quite aggressive". A single flagged word usually adds a fraction of that, so one "free" does not sink a message; several patterns together can.

Why the words still matter a little

• Humans are the real filter. People mark as spam what feels like spam. Hype language raises complaint rates, and complaints are a top reputation signal: Gmail asks bulk senders to stay under 0.10% and never reach 0.30% (see the requirements checklist).
• Stacking is the problem. One promotional word is normal. A subject line with a money claim, an urgency phrase, all-caps words and three exclamation marks looks like bulk junk and trips several rules at once.
• Some phrases are strongly associated with scams (lottery, inheritance, "act now" with guarantees), so they are worth avoiding on principle.
• Corporate rule-based filters are everywhere. Your B2B recipients' mail servers often run them, and they react to patterns that consumer providers ignore.

Myths that waste your time

• "One spam word sends the whole email to junk." Not in a modern filter. A word is one weak feature among thousands of signals, and rule-based filters sum small scores.
• "A spam-checker score of 0/10 guarantees the inbox." It only says your wording does not match the rules the tool knows. Reputation and authentication are invisible to it.
• "Misspelling the word (fr33) fixes it." Obfuscation is itself a spam pattern, and it makes you look less trustworthy to human readers.
• "Spam words only matter in the subject." Body text, link text, the sender name and even the ratio of images to text are all evaluated.
• "Transactional emails are exempt." They are filtered too; a password reset full of promotional language can land in junk.

Context decides: the same word, two outcomes

Filters, and readers, judge the whole message and the relationship. "Free trial" from a SaaS product someone signed up for last week is expected; the same phrase in an unsolicited message to a purchased list is a red flag. Likewise, words that are normal in your industry (a pharmacy writing about prescriptions, a bank writing about loans) will be scrutinised more closely because spammers imitate those sectors, so your authentication and reputation have to carry more of the weight. This is why a word list alone cannot be right for every sender.

The categories to watch

Category	Typical phrases	Safer approach
Money and offers	"100% free", "cash", "earn money", "extra income"	State the real benefit and conditions: "Free shipping on orders over $50"
Fake urgency	"Act now", "final notice", "last chance", "today only"	Use true deadlines: "Registration closes Friday at 5 pm"
Exaggerated claims	"Guaranteed", "no risk", "winner", "once in a lifetime"	Support claims with specifics or evidence
Pushy calls to action	"Click here", "buy now", "open immediately"	Describe the action: "See the full report"
Scam signals	"Dear friend", "lottery", "this isn't spam"	Never use; they match known fraud patterns

Formatting matters as much as vocabulary: ALL CAPS words, repeated exclamation marks, a fake RE: or FW: prefix on a message that is not a reply (deceptive subject lines are also an offence under laws such as CAN-SPAM), and strings of symbols such as $$$.

Rewrite examples

Before	After	Why it is better
URGENT!!! 100% FREE GIFT - ACT NOW	Your free sample pack is ready to claim	No caps, no stacking, clear benefit
RE: your order (not a reply)	Your October order: shipping update	No deceptive prefix; honest and specific
Make money fast with this guaranteed system	3 ways small teams reduce churn this quarter	Specific, no guarantee claim
CLICK HERE to win cash!!!	Take the 2-minute customer survey	States the action and effort
Limited time!!! Don't miss out on huge savings	Your 15% code expires Sunday	A real number and a real deadline replace hype

🚫Spam Words CheckerFind known spam-trigger phrases in your email with highlighting, categories and a declared risk levelOpen the tool →

A scoring model you can reason about

The checker on this site is deliberately simple and transparent so you can reason about it. It matches a dictionary of phrases in five categories (money, urgency, claims, call to action, scam), each phrase with a weight between 0.5 and 3, counts each phrase once, lets a longer phrase hide the shorter ones inside it (so "100% free" is not also "free" and "100%"), and maps the sum to low, medium or high risk. That is the same shape as a rule-based filter's scoring, which is why it is useful for catching a pile-up. It is not, and cannot be, a prediction of what Gmail will do with your message.

How to use a spam-words checker properly

1. Treat it as a lint tool, not an oracle. A low score does not guarantee inbox placement.
2. Look at the pile-up. One flagged term is fine; three categories at once is a red flag.
3. Read the sentence in context. "Discount" in a retailer's newsletter is normal; the same word in an unsolicited cold email is not.
4. Fix the foundation first. Authentication, list hygiene and unsubscribe handling matter more than any word list.

💬Email Subject Line TesterDeclared-formula score for subject lines: length, mobile preview, spam and personalization, with multi-line comparisonOpen the tool →

Other content signals worth a quick check

Signal	Why it matters	Practical rule
Image-only emails	Filters cannot read the text, and readers with images blocked see nothing	Keep real text and alt attributes for all key content
Broken or mismatched links	Link text showing one domain while pointing to another looks like phishing	Make the visible URL match the destination
URL shorteners	Widely abused, so often penalised	Use your own branded domain for tracking links
Missing plain-text part	Multipart messages are the norm for legitimate mail	Always send a plain-text alternative
Missing unsubscribe link	Required by law and by Gmail and Yahoo for bulk senders	Visible footer link plus the List-Unsubscribe header
Hidden text and tiny fonts	Classic keyword-stuffing technique	Never hide copy with colour or size tricks

Setting up the unsubscribe header correctly is covered in the List-Unsubscribe guide.

What to do instead of memorising a list

• Authenticate your domain and keep complaints below 0.1%.
• Write the subject line you would be comfortable receiving, then trim the hype.
• Keep subject lines honest and aligned with the content (see writing an effective subject line).
• Test two versions by clicks, not opens (see the A/B testing sample size guide).
• Read the headers of a test message in a corporate mailbox to see whether a rule-based filter scored it (how to read email headers).

Frequently asked questions

Do spam trigger words still affect email deliverability?

Slightly. Reputation, authentication and engagement dominate. Wording matters mostly because stacked hype raises complaint rates and can trip content rules in rule-based filters.

Is the word "free" a spam word?

It is a weak signal on its own. It becomes risky when combined with capitals, exclamation marks, urgency and money claims.

Do exclamation marks and all caps hurt deliverability?

Excessive use is associated with spam and can increase complaints. One exclamation mark is rarely a problem; strings of them and shouting words are.

Should I put "not spam" in my email?

No. Phrases like "this is not spam" are themselves a classic spam signal.

How can I test whether my email will land in spam?

No tool can promise placement. Check authentication, send seed tests to your own mailboxes and monitor complaint and engagement data after sending.

References & further reading

1. Apache SpamAssassin — Mail::SpamAssassin::Conf (required_score and scoring)
2. Google — Email sender guidelines
3. Google — Email sender guidelines FAQ
4. FTC — CAN-SPAM Act: A Compliance Guide for Business
5. M3AAWG — Sender Best Common Practices