Opt-in

Contact details used to be something you simply picked up. A list bought from somewhere got blasted with texts. A business card collected at a booth went straight onto the newsletter list. Sending cost the sender almost nothing, so the arithmetic said send more.

For the people receiving it, the arithmetic ran the other way. Dozens of messages nobody asked for piled up daily, and unsubscribing meant hunting for small type hidden in a different spot in every message. Plenty of those links didn't work. Eventually people stopped filtering individual senders and started ignoring email altogether, and mail providers began routing suspicious senders into spam wholesale. Mail from companies that had collected permission properly got buried with everything else. The cheap channel had priced itself out: the more that went out, the less of any of it got read.

Opt-in solves this by reversing the order. Rather than sending first and letting people ask to be removed, you collect a yes before anything goes out. Many jurisdictions have written that requirement into law for promotional messages, with fines attached. In practice, that makes opt-in less a strategy you choose than the baseline you start from. Reversing the order also moves the cost of a bad list onto the sender, up front, instead of onto every recipient afterward.

Opt-in holds up when three pieces exist: the place consent is collected, the record it leaves, and the way back out.

1. Build a place that asks. A checkbox on the signup screen is the usual form, and it has to start unchecked. Consent collected through a pre-checked box doesn't count as consent, on the reasoning that a box nobody touched records no decision at all.
2. Split what's being agreed to. The terms you must accept to use the service and the marketing consent that lets you send promotions cannot share one checkbox. Bundled, anyone signing up takes the marketing along with it, and what they did was get past a gate rather than make a choice.
3. Keep the record. Store who agreed, when, on which screen, and against which wording. If a complaint arrives months later, that record is the only thing the company has to show. Keep the fact of consent without the wording and nobody can say what was actually agreed to.
4. Decide whether to double opt-in. A confirmation message goes to the address given, and the subscription is final only once the link inside it is clicked. The extra step costs roughly ten percent of signups, but typos and other people's addresses get filtered out, which lifts deliverability. It also answers a later claim of never having signed up, since the confirmation click is a second piece of evidence.
5. Leave the exit open. Every message carries an unsubscribe link, and pressing it has to take effect from the next send. A narrow door in and a wide door out is the basic shape of opt-in. An unsubscribe that takes three days to apply reads, to the person who pressed it, as no unsubscribe at all.

It attaches to every channel where a company speaks first.

1. Email marketing — a newsletter signup form is the most typical opt-in there is. Pressing subscribe is itself the consent, so no separate checkbox is needed.
2. Push notifications — the operating system puts up a permission dialog and collects consent on your behalf. A refusal is hard to reverse, so many teams show a screen explaining why the permission is useful before triggering the dialog.
3. Send times — some countries require a second, separate consent for promotional messages delivered late at night. Teams that stored consent as one undifferentiated flag have to go back and re-ask the whole list when a rule like that applies to them.
4. Tracking and personalization — collecting browsing behavior in order to show different content to different people needs consent of its own. Europe's GDPR is the best-known example, and the consent banner is where it lands.

One more number comes with it: the share of registered users who agreed to receive marketing. At 20 percent, a thousand new signups means a campaign that can reach two hundred of them. Raising that share by making the checkbox less visible invalidates the consent itself, so the thing to work on is the sentence next to the box and where in the flow it appears.

"Is this list opted in? If not, we can't send to it."

Check the consent record before anything goes out.

"The signup form bundles terms and marketing into one checkbox. Please separate them."

Bundled consent doesn't count as consent, so split the boxes.

"The list shrank after we turned on double opt-in, but open rate went up."

Address quality improved, so a larger share of recipients actually read it.

At Pagedeck, Ravi filters every campaign list by consent status before building the campaign. Not all free-plan users, only the ones who agreed to marketing, and for anything scheduled at night he checks the separate late-hours consent too. Grace audits quarterly that those records actually exist — during one migration between sending tools, the consent timestamps dropped out entirely. A permission is only as good as the record that survives the next tool change.

Opt-in isn't only the marketer's job. A designer draws the checkbox, an engineer stores the value, and legal reviews the wording. Miss one of the three and the other two don't help.

The most common belief is that signing up permits anything. Someone typed in their own address, so it feels like blanket permission. But consent is scoped by purpose. A payment receipt or a password reset is service communication, not promotion, and needs no separate consent. A discount announcement or a new-feature push is promotion and does. What separates the two is the purpose of the message, not which tool happened to send it. Blur that boundary and one message raises the complaint rate for the whole list, after which even the mail you have to send lands in spam.

The other belief is that asking for consent shrinks the list and costs you reach. On the raw count, true. The people removed, though, are the ones who were never going to open anything. A campaign that went to 10,000 people and got 500 opens becomes one that goes to 2,000 and gets 480: open rate climbs from 5 percent to 24 percent while the number of actual readers barely moves. The comparison that decides anything is readers, not recipients. Fewer dead addresses also improve how mail providers view the sender, so a larger share of the same mail reaches the inbox at all.