There is a lot of advice about spam filters, and much of it is guesswork. The reason is simple: mailbox providers (the services that run inboxes) publish guidelines for senders, but they keep their filters secret and change them often. This exercise sorts popular beliefs into myths and facts.
In this exercise, fact means the statement fits what large providers say in their own sender guidelines, such as Google's guidelines for email senders, or how filtering is generally known to work. Myth means the statement is wrong as it stands, usually because it is far too absolute.
What is actually known
Filters combine many signals. No single feature decides on its own in normal cases.
The sender's reputation and the reactions of recipients (complaints, replies, moving mail out of spam) carry a lot of weight.
Providers ask for authentication (SPF, DKIM, DMARC), a low spam complaint rate, an easy way to unsubscribe and a steady sending volume.
Content still counts: misleading wording, phishing patterns and links to bad domains can all hurt. But words are one input among many.
The exact weights are secret. Anyone who claims to know the precise formula is guessing.
The tasks
Why each statement is a myth or a fact
Myths
A single word sends any email to spam: content is analysed, but reputation, authentication and recipient reactions weigh more. The word "alone" makes it a myth.
Authentication guarantees the inbox: it is required and it helps, but it only proves who sent the mail. Spammers can authenticate their own domains too.
Every extra link makes it worse: there is no published link limit. Where the links point matters more than how many there are.
An image hides the text from filters: it was a spam trick many years ago; filters adapted. It also makes the email useless for many readers.
"This is not spam" helps: filters do not believe what an email says about itself.
Once in the inbox, always in the inbox: each message is judged again, and reputation changes with every send.
Facts
Complaints hurt: Google publishes a spam rate limit (below 0.1 percent, never 0.3 percent) for this reason.
Bounces hurt: many addresses that do not exist show a list that was bought, guessed or never cleaned.
Sudden jumps are risky: providers ask senders to increase volume slowly. See the guide to sender reputation and warm-up.
Reactions can count more than words: providers name complaints and engagement in their guidance; they publish no word lists.
An easy unsubscribe protects you: people who can leave easily are less likely to complain.
Bad link domains harm an email: filters check where links lead.
Teacher notes
How to run it (25 minutes). Start with a show of hands: who has heard that the word "free" sends email to spam? Then 15 minutes for the three tasks, alone or in pairs. Use the last 5 to 10 minutes to read the explanations together and to discuss how to judge advice.
Common mistakes.
Swapping one myth for the opposite one: "content never matters" is as wrong as "one word decides".
Treating authentication as the finish line instead of the starting point.
Trusting advice because it sounds technical. Ask learners to name the source of each tip they found online.
Follow-up questions.
Find a deliverability tip on a blog or social network. Is it a myth, a fact or "it depends"? What source does it give?
Why do providers not publish their filters? (Spammers would adapt at once.)
Continue with the Contoso case study, where learners diagnose a real-looking spam folder problem.