Exit-Intent Done Tastefully

The exit-intent popup has a terrible reputation, and it earned it. You know the pattern: your cursor drifts toward the tab bar and the page lunges at you with a modal. Usually it begs. “Wait! Don’t go!” Sometimes it insults you, with those decline buttons that say “No thanks, I hate saving money.” The whole genre feels like a store clerk grabbing your sleeve on the way out, and most advice about it splits into two useless camps: the growth camp says use it because it works, and the taste camp says never use it because it’s gross. I think both camps are skipping the interesting question, which is what the leaving visitor actually deserves from you.

Start with what exit intent really is. It is a signal, detected on desktop by the cursor moving quickly toward the top of the viewport, that this visitor has decided your page is not worth more of their time. That decision is information. Something on the page failed: the price was unclear, the product seemed wrong for them, they got what they came for, or they were never going to buy and just wandered in. An exit prompt is your one chance to respond to that failure, and the reason most exit popups feel gross is that they don’t respond to it at all. They just repeat the original ask, louder. The visitor said no to the newsletter box in the sidebar, and now the same newsletter box is blocking the whole screen. Nothing changed except the aggression.

The tasteful version follows a different rule: the interruption must contain something the page didn’t already offer. An interruption is a cost you impose on the visitor. It is only justified if you pay for it with new information. A different framing of the offer counts. A genuinely smaller ask counts, like a one-page summary instead of a sales call. An honest question counts: “Before you go, what were you looking for?” with a text box, which converts a lost visitor into a sentence of research about why you lost them. What doesn’t count is the same pitch in a bigger font. If your exit modal could be deleted and its content moved into the page with no loss, the modal was never adding anything. It was only subtracting patience.

Tone matters more here than anywhere else on your site, because of when the message arrives. This is the last thing the visitor reads before forming their final impression of you. A confident product says something like: no problem, here’s a link to the docs if you want to evaluate us properly later. A desperate product says WAIT and starts discounting. Oddly, the panicked discount often costs you twice. The visitor who was leaving takes the coupon and still leaves, and the visitor who was about to pay full price learns that your list price is fiction and your real price appears if you feint toward the exit. Training your best prospects to fake leaving is a strange business model, but plenty of sites run it.

There are also mechanical rules that separate tasteful from hostile, and they are not subtle. Once per visitor, ever, or at most once per long stretch of time, enforced with a cookie or local storage. A close button that is visible, big enough to hit, and instant. The escape key must work. And on phones there is no cursor to watch, so “exit intent” on mobile is usually a guess based on scrolling or the back gesture, and the guess is wrong often enough that the polite move is restraint. A modal that misfires while someone is reading is worse than no modal at all, because it interrupts the one visitor who was actually engaged. Beyond rudeness there is a practical edge: intrusive interstitials on mobile can hurt how search engines assess your page, and an unclosable overlay is an accessibility failure for anyone navigating by keyboard or screen reader. When I scan pages with GazeSite, overlays like this show up in the screenshot exactly as a visitor experiences them, sometimes covering the very call to action the page wanted clicked, which has a certain dark comedy to it.

The deeper point is that exit intent is a technology for having the last word, and having the last word is a responsibility. Most sites use it the way a bad conversationalist does, to restate their case as the other person reaches for the door. The better use is the one good hosts know instinctively: make leaving pleasant. Answer the question they came with if you can. Give them something small and genuinely free to take along. Leave the door visibly open. Visitors remember how a site treated them at the moment it had nothing left to gain, and some of them come back because of it. The popup isn’t the problem. Begging is.