Words are a signal, not a verdict
Let's be honest up front: in 2026, no single word decides where your email lands. Gmail and Yahoo lean overwhelmingly on authentication (SPF, DKIM, DMARC), sender reputation, engagement, and your spam-complaint rate — the things covered in the Gmail & Yahoo sender requirements. A well-authenticated sender with a clean list can write “free” all day long.
But content scoring hasn't disappeared — it has moved to accumulation. SpamAssassin-style filters add up small scores from many signals, and a message that stacks bait phrases on top of ALL-CAPS, exclamation runs, and an image-only layout still tips over — especially from a sender without an established reputation. This site's checker works the same way: a single mild phrase scores as safe, a couple of them draw a warning, and a genuine pile-up (or one heavyweight phrase too many) gets flagged as spam.
The trigger-word list, by category
These are the phrases the checker scans for — a deliberately conservative list of the classics that real filter rules target. Phrases marked strong carry weight on their own; the rest only matter when several stack up in one message. Matching is case-insensitive, and smart-quote tricks (“you’ve won” with a curly apostrophe) are folded back to the plain form before scanning.
Prizes and winnings
- you've won — strong. The classic prize-scam opener; almost no legitimate newsletter needs it.
- congratulations — mild, but a staple of lottery and giveaway scams.
- free prize — mild; “free” plus “prize” is a scam-mail signature.
Money and finance
- $$$ — strong. Stacked dollar signs read as get-rich bait to both filters and humans.
- wire transfer — strong; the payment channel of advance-fee fraud.
- no credit check — strong; a predatory-lending staple.
- make money — mild get-rich-quick language.
- extra income / extra cash — mild; same family.
- work from home — mild; burned into filters by years of job-scam mail.
Urgency and pressure
- act now — mild manufactured urgency.
- buy now / order now — mild hard-sell pressure.
- limited time — mild; fine once, suspicious in a chorus.
- click here — mild, and bad link copy anyway (see the rewrites below).
Free offers and guarantees
- 100% free — mild; the overclaim is the problem, not the word “free”.
- risk-free — mild; matched with or without the hyphen.
- guaranteed — mild; any form of “guarantee” counts.
- no obligation — mild sales-pitch boilerplate.
- exclusive deal — mild promotional bait.
Health and pharma
- viagra / cialis — strong. The canonical pharma-spam terms; either one alone draws a flag.
- weight loss — mild; heavily abused by miracle-product mail.
- miracle — mild; the adjective of choice for health-claim spam.
The formatting signals filters score
Word choice is only half of content scoring. The checker — like the filters it mirrors — also scores how the message is built. These signals add to the same accumulated score as the phrases above.
ALL-CAPS shouting
Words of four or more letters in full uppercase read as shouting. Three or more of them in one message adds to the score; six or more adds twice as much. Common acronyms (HTML, HTTPS, RSVP, ASAP, JSON) are exempt, so technical writing isn't penalized.
Exclamation marks
A run of three (!!!) anywhere, or six or more scattered through the body, is a flag. One genuine exclamation is fine — enthusiasm by volume is not.
Emoji pile-ups
Three or more loud emoji in the body reads as promotional shouting. The occasional emoji is harmless; a row of money bags and sirens is a signal.
Link count
More than 20 links in one message is link-heavy and draws a warning. Worse, a link that points at a raw IP address instead of a domain (http://203.0.113.9/…) is treated as an outright spam signal — legitimate senders link to domains.
URL shorteners
Links through bit.ly, tinyurl.com, t.co, and their kin hide the destination, which is exactly what phishing mail wants — so filters penalize the pattern wholesale. Link directly to your own domain instead.
Image-to-text ratio
An email that is essentially one big image — images present but almost no live text — is treated as spam: it renders blank with images blocked, and image-only mail is a classic filter-evasion tactic. Very little text relative to the number of images (roughly under 60 characters of text per image) draws a warning. Always send real text alongside your images.
The subject line
The same logic applies up top. Bait words in the subject (“free”, “winner”, “urgent”), a fake Re: or Fwd: prefix on a first-contact message, and doubled exclamation marks all get flagged — and a deceptive subject line also violates the CAN-SPAM Act, which makes it a legal issue on top of a deliverability one.
What to write instead
The goal is not to launder your vocabulary — it is to sound like a human with something specific to say. Concrete beats hype on every axis filters measure:
- Swap manufactured urgency for real deadlines. “Act now — limited time!!!” becomes “Sale ends Thursday at midnight.”
- Link descriptive text, not “click here”. “Read the full changelog” tells readers and filters where the link goes.
- Make claims concrete instead of guaranteed. “Guaranteed results!” becomes “9 of 10 readers finish setup in under five minutes.” Numbers beat superlatives.
- Say what free means. “100% free, no obligation!” becomes “Free — no card required.”
- Emphasize with bold, not caps-lock. One bolded phrase carries more weight than a SENTENCE OF SHOUTING, and costs you nothing with filters.
- Never obfuscate. F-R-E-E, “v1agra”, and smart-quote tricks are themselves classic spam patterns — filters normalize them away, and human readers trust you less.
Frequently asked questions
Do spam trigger words still matter in 2026?
Yes, but as a minor signal. Gmail, Yahoo, and Outlook weigh authentication (SPF, DKIM, DMARC), sender reputation, engagement, and spam-complaint rates far more heavily than any word choice. A pile-up of bait phrases still hurts — especially for senders without an established reputation — but no word list substitutes for genuine opt-in and clean sending infrastructure.
Will a single word like “free” send my email to spam?
Almost never. Modern filters score by accumulation, and so does this site's checker: one mild phrase in an otherwise normal email scores as safe. Trouble starts when several trigger phrases stack up, or when a heavyweight phrase like “you've won” or “wire transfer” appears alongside shouty formatting.
Should I strip every trigger word from my newsletter?
No. Rewriting “free shipping” into an awkward euphemism costs you clarity and gains you almost nothing. Fix the pile-ups and the formatting signals — ALL-CAPS, exclamation runs, image-only layouts — and keep the odd promotional phrase where it reads naturally. Never obfuscate words (F-R-E-E, v1agra) to dodge filters: that pattern is itself a classic spam signal.
Do filters check the subject line too?
Yes. Bait words like “free”, “winner”, or “urgent” in the subject, fake “Re:” or “Fwd:” prefixes, and doubled-up exclamation marks all draw flags. And under CAN-SPAM, a deceptive subject line is a legal problem, not just a deliverability one.
How do I find out which trigger words my email contains?
Paste your HTML or upload the raw .eml into the free checker on this site. It scans for the exact phrase list in this guide plus ALL-CAPS density, exclamation runs, emoji, link count, URL shorteners, and image-to-text balance — then tells you what to fix in plain English.
Score your email against this exact list
You don't have to cross-check this page by hand. Paste your newsletter HTML or upload the raw .eml and the checker scans for every phrase above — plus ALL-CAPS density, exclamation runs, emoji, link count, shorteners, and image-to-text balance — and gives you a plain-English verdict with fixes.