Opt-out

Collecting addresses and mailing them was the default for a long time. Business cards from a conference, addresses left behind on a download form, and lists bought from somewhere else all ended up in one file, and the send button got pressed against it.

The cost of that arrangement landed entirely on the receiving side. Ten messages a day meant hunting for ten different unsubscribe links, each one usually set in small gray type at the bottom of the message. Clicking it often produced a login prompt, or two more screens asking why you were leaving. So people stopped unsubscribing and started marking messages as spam instead, and once those complaints accumulated against a sending domain, ordinary mail from the same company started landing in spam folders too. Reputation earned that way takes weeks of careful low-volume sending to repair, and there is no button that undoes it.

Regulation moved the same direction. The principle that commercial messages need permission collected in advance took hold in Europe under the General Data Protection Regulation (GDPR), and similar rules followed in other jurisdictions. So at the stage where a list gets built, opt-in became the working default in most markets.

That did not remove opt-out; it relocated it. What used to be a way of assembling a list became the mechanism that lets somebody who already consented leave whenever they want.

The sender sends first, and the moment a recipient indicates they do not want this, they come off the list. Five things have to be in place for the arrangement to function honestly.

1. The way out has to be visible. In an email marketing message, one line at the bottom of the body is the standard place. Set in six-point gray, it satisfies the form and nothing else — people will report the message as spam rather than hunt for it.
2. It has to finish in one action. Clicking the link should complete the removal without a login. Adding two confirmation screens, or requiring a reason before proceeding, is a design that blocks while appearing to comply, and that pattern has a name: a dark pattern.
3. Removals have to be stored permanently. The list that holds them is usually called a suppression list. Without it, next month's upload puts the same person back in, and somebody receiving a second copy of a message they already refused answers with a complaint rather than another unsubscribe.
4. The scope should be divided. Offering "stop everything" alongside per-type controls lets someone turn off the weekly roundup while continuing to receive payment failure notices.
5. Processing has to be fast. Even with sends already queued, a removal should take effect within a day.

On the recipient's screen, this approach shows up as a setting that arrives already switched on, which makes the default the whole decision. A new user whose marketing preference starts enabled is on opt-out; one whose preference starts disabled is on opt-in. A pre-checked consent box on a signup form looks like consent but is functionally opt-out, which is why it does not count where prior permission is required.

The most common place is the footer of messages sent to people who already consented. Even a list built entirely through opt-in needs a way out, so the two are not alternatives to choose between — they sit back to back.

Notification settings inside the product are built the same way. Separating email, push and in-app messages into individually switchable types lets someone reduce the volume without leaving the product. The practical test is whether a person who wants less can find the right control in under ten seconds. Where the only option is all or nothing, every complaint about volume turns into a decision about whether to keep the product at all.

The same structure appears on the data side of marketing. The clearest example is the choice not to have your activity used for targeted advertising. Requirements differ sharply by region, though. Where GDPR applies, tracking needs consent beforehand, which is why a cookie consent screen appears before anything loads. In the United States, the federal CAN-SPAM Act permits sending commercial email first as long as recipients can opt out afterward, so opt-out remains a valid default there. Because the same campaign has to be designed differently depending on where recipients live, the applicable marketing privacy rules get checked before the list and the copy are built.

"This campaign can't run on opt-out. It's promotional, so it only goes to the consented list."

At the stage of collecting permission, this approach is not available.

"You have to click the unsubscribe link three times before it processes. That turns into spam complaints."

Making the exit hard produces a worse signal than the exit itself.

"Let's do per-type controls instead of one global unsubscribe. People are losing payment notices too."

A single switch is blocking messages the recipient actually needs.

At Pagedeck, Ravi reviews the suppression list in the sending tool every month. New uploads are filtered against it automatically, and since that was set up nobody has received the same message twice.

The bar rises with the enterprise customers Sofia works with. Some ask for a contract clause stating that marketing email will not go to their employees' addresses, so those accounts get flagged as excluded from sends entirely. Grace audits the consent records quarterly, and any list where there is no record of when and with what wording permission was given does not get used.

The most common error is treating opt-in and opt-out as a choice between two options. They look like an opposing pair, which invites that reading. But they cover different moments: consent governs getting onto a list, and opt-out governs getting off it. Miss that and you arrive at the conclusion that a consented list does not need an unsubscribe link. Mail sent on that basis accumulates complaints until the sending domain itself becomes unusable. The fix would have been one line of text at the bottom of the message, which is what makes the omission hard to defend after the fact.

The second error is treating a low removal rate as the goal. Lowering numbers is usually the right instinct, so this one is easy to fall into. But suppressing the unsubscribe rate does not make those people disappear — it moves them to the spam button. Spam complaints erode sender reputation, which lowers deliverability, which means mail stops reaching the people who did want it. An unsubscribe is closer to the list cleaning itself. What deserves reducing is not the number of exits but the sending frequency and content that produce them.