Not every advertisement is spam.
Not every unwanted message is illegal.
Not every legal commercial message is wanted.
Those three facts make spam surprisingly hard to measure honestly.
Spamhaus uses a deliberately narrow operational definition: unsolicited bulk email. Its blocklist documentation says it does not classify mail based on whether the advertised content itself is legal; it focuses on whether the traffic fits its unsolicited-bulk definition. See the Spamhaus Blocklist policy.
U.S. law uses a different frame. The FTC’s CAN-SPAM guidance explains that commercial email generally does not require prior opt-in consent under federal law, provided senders meet the law’s requirements and honor opt-outs. See FTC guidance on CAN-SPAM.
So a message can be lawful commercial email and still be unsolicited.
A recipient can also consider an old newsletter unwanted even though they originally subscribed.
“Commercial” is a content category, not a verdict
A receipt is commercial in the broad everyday sense that it came from a business, but it is not equivalent to a cold sales pitch.
A subscriber may actively want a weekly store circular.
Another person may hate the exact same message.
A useful study therefore needs categories that describe what actually happened.
For example:
- Was the message commercial, transactional, personal, political, fraudulent, or something else?
- Did the recipient knowingly subscribe?
- Was the message expected from this sender?
- Was it sent in bulk?
- Had the recipient previously opted out?
- Did the recipient personally classify it as unwanted?
- Was it reported as spam to the provider?
Those variables can overlap without being interchangeable.
Recipient expectation matters
Provider complaint metrics are one useful signal because they capture an explicit recipient action.
Gmail’s Postmaster system measures user-reported spam rates for senders. See Gmail’s sender guidelines FAQ.
But even that misses plenty of unwanted mail.
Many recipients simply delete a message. Some unsubscribe. Some ignore it. Some never see it because a filter removed it first.
So “reported spam” is not the same thing as “all spam,” and neither is the same thing as “all commercial email.”
Good measurement admits uncertainty
A strong Dead Internet Theory study should resist emotionally convenient definitions.
If every advertisement is counted as spam, the result exaggerates abuse.
If only illegal mail is counted, the result ignores enormous volumes of unwanted but lawful contact.
If only complaint-button presses are counted, the result depends heavily on user behavior and provider interface design.
The better approach is to state the definition first and report ambiguous cases separately.
Spam Empires are worth studying precisely because unwanted attention exists on a spectrum from obvious criminal botnet traffic to a legitimate retailer that simply kept talking after the customer stopped listening.
The measurement gets stronger when the vocabulary stops pretending those are all the same thing.