Exit-intent popups that don’t feel like a trap

Exit intent is the moment a visitor has decided to leave: the cursor heads for the tab bar on desktop, or the thumb flicks upward toward the address bar on mobile. A popup that fires here interrupts nothing, because the visit was already over. That is what makes it the highest-leverage trigger in the toolbox — and also the easiest to turn into an ambush.

The difference between the two is not the mechanism. It is who sees it, what it offers and how it behaves when the answer is no.

Why exit intent works at all

A timed popup competes with whatever the visitor was doing. An exit popup competes with nothing. The visitor has finished browsing, hasn’t bought, and is about to become a statistic you can’t reach. Making one clear offer at that moment costs you nothing you were going to keep anyway.

It also self-selects a useful audience. People who leave from a product page after two minutes were interested enough to look and not convinced enough to buy. A small reason to stay in touch — a discount, a restock alert, a fit guide — lands differently with them than with someone who bounced from the homepage in three seconds.

Why it so often fails

Most exit popups fail for one of five reasons, and none of them is the popup format:

• It fires on everyone. Including people who already subscribed last week and people who just bought. They didn’t leave because they needed a discount. They left because they were done.
• It fires everywhere. On the cart, on the checkout, on the returns policy. A popup over the checkout is not a save; it is a distraction from the sale you were about to make.
• It’s dishonest. “Wait! Are you sure?” is a guilt trip, not an offer. Fake countdowns and “only 2 left” badges nobody checks train visitors to distrust the whole store.
• It’s twitchy. Desktop exit detection tuned too aggressively fires when someone reaches for the back button or a bookmark. The visitor was not leaving. Now they are.
• It never stops. The same wheel, every visit, forever. By the third time, the dismiss button is the only thing being clicked.

Five rules for an exit popup people forgive

1. Offer something specific

The headline should contain the offer, not a plea. “Leave with 10% off, on us” beats “Don’t go yet!” because the visitor can evaluate it in a glance. If you have nothing worth offering, don’t show a popup — a restock notification, a size guide or a “save your cart” link is a perfectly good exchange for an email address.

2. Treat desktop and mobile as different triggers

There is no cursor on a phone. Mobile “exit” is a proxy — a fast upward scroll, a scroll back to the top, a long idle. It is noisier, so the mobile version should be gentler: a slide-in or a teaser rather than a full modal, and a longer suppression window. In Airo, triggers are scoped per device precisely so that “mouse leaves the page” on desktop and “scrolls up quickly” on mobile can be two rules with two sensitivities.

3. Cap it hard

Once per visitor per two weeks is a reasonable default. Once per session is the floor. A visitor who dismisses the popup has answered; suppress it for longer than you think you should. The cost of one missed opportunity is small. The cost of being the store that nags is not.

4. Exclude the people who already said yes

Subscribers, recent purchasers and anyone who submitted a form on this visit should never see an acquisition popup again. This sounds obvious and is the most common mistake we see, because most tools make “has already subscribed” a hard condition to express. Make it the first rule you add, not the last.

5. Stay off the transactional pages

Cart, checkout, account and policy pages are exempt. If you must show something on the cart page, make it about the cart — “you’re $12 from free shipping” — and put it in a bar, not a modal.

An exit popup is a last word, not a last stand. If it reads as though the store is afraid of the visitor leaving, the visitor will leave faster.

Copy that respects the answer

Two lines matter more than the rest: the headline and the dismiss link. The headline carries the offer. The dismiss link carries your character. “No thanks, I’ll pay full price” is a well-worn joke that mildly shames the reader; “No thanks” is fine; “Maybe later” is honest if you actually plan to show it later. Pick one you would be comfortable saying to a customer in person.

Keep the form to one field. If you need a phone number as well, ask on the next step, after the email has landed — a two-step capture converts better than a two-field form and lets you keep the email even if the phone never comes.

How to know it’s working

Three numbers tell the story:

• Submit rate — submissions divided by impressions. A well-targeted exit popup usually beats a timed one here, because the audience is self-selected.
• Dismissal rate — the early-warning signal. If it climbs week over week while submit rate falls, you are showing it to the wrong people or too often.
• Attributed revenue — the only number that justifies the popup’s existence. An exit offer that generates leads who never buy is a cost centre with a nice chart.

Then test the thing you are least sure about. In our experience that is almost always the offer, not the design: “free shipping” against “10% off”, a restock alert against a discount, one field against two. Run the variants to statistical confidence, promote the winner, and move on to the next uncertainty.

The checklist

Do those nine things and the exit popup stops being the thing people complain about and becomes what it should have been all along: a polite last word to someone who was interested enough to look.