Exit-intent popups aren't evil. Bad ones are.
Exit-intent popups have a bad name, mostly earned by lazy ones. Here's when they help visitors, when they hurt, and how to build one people don't resent.
You’re reading an article. Your cursor drifts towards the tab bar because you want to check something, and the screen goes dark. A box appears: “WAIT! Don’t miss out!” The close button is a pale grey cross in the corner. The alternative to signing up is a link that says “No thanks, I don’t like saving money.”
Everyone has met this popup. Most people who work in marketing have, at some point, shipped it. It’s why “exit-intent” has become shorthand for a cheap trick, and why plenty of teams refuse to use one at all.
We think that refusal is a mistake. Not because the popup above is defensible (it isn’t), but because the idea underneath it is sound, and throwing it out leaves something useful on the table.
What exit intent actually is
Strip away the bad examples and exit intent is just a signal: this visitor looks like they’re about to leave. On desktop it’s usually the cursor heading for the top of the window. On mobile, where there’s no cursor, it’s typically inferred from things like a fast scroll back up or a period of inactivity.
A signal isn’t good or bad. What matters is what you do with it. Interrupting someone with a generic discount offer is one response. Asking whether they found what they were looking for is another. The trigger is identical; the experience is completely different.
Why the bad ones are bad
The popups people hate tend to share a few traits:
- They’re irrelevant. A newsletter offer on a page where someone was comparing prices. A discount on a page about a service that’s quoted individually.
- They’re manipulative. Confirm-shaming copy, hidden close buttons, fake countdown timers.
- They’re repeated. The same box on every page, every visit, after you’ve already dismissed it twice.
- They ask for too much. An email address, in exchange for nothing in particular, from someone who’s halfway out the door.
Notice that none of these is about exit intent itself. They’re about relevance, respect and value. A popup that shows on page load with the same problems is just as irritating; it’s simply less memorable because it doesn’t catch you mid-gesture.
When leaving is a useful moment
Think about why people leave a page. Some got what they came for: they read the article, found the opening hours, checked the price. Let them go. Others leave because they didn’t get it: the answer wasn’t there, the form looked like a chore, the price raised a question nobody answered.
That second group is worth talking to. They had enough interest to arrive, and something stopped them. A good exit prompt is the equivalent of a shop assistant noticing someone heading for the door with a puzzled look and asking, “Did you find what you were after?” Nobody minds that. Plenty of people appreciate it.
Some examples of exit prompts that respect the visitor:
- On a pricing page: “Not sure which plan fits? Tell us roughly what you need and we’ll suggest one.”
- On a property listing: “Want to know if this one’s still available, or see similar homes nearby?”
- On a car listing: “Got a question about this car? Ask, or message us on WhatsApp.” (Motormetry’s demo dealership puts WhatsApp on its listings, if you want to see the idea in place.)
- On a long guide: “Want this as a checklist you can come back to?”
Each of these is specific to the page, offers something the visitor might actually want, and asks for little. That’s the whole trick. If you’re tempted to write a single prompt for the entire site, read our piece on page-aware conversations first.
Rules we’d follow
If we were setting up exit intent on a site today, these would be the house rules:
- Match the page. Different pages get different prompts, or no prompt at all.
- Offer before you ask. Lead with an answer, a shortlist, a booking slot. The contact details can come once there’s a reason to give them.
- Make leaving easy. A clear close button, an escape key that works, and copy that doesn’t sulk if you say no.
- Show it once. If someone dismisses it, don’t bring it back on the next page or the next visit for a good while.
- Leave the converted alone. Someone who already booked or enquired shouldn’t be asked again.
- Keep it light on mobile. A full-screen takeover on a phone is far more intrusive than on a desktop. A small slide-in or a chat bubble is usually enough.
Make it a conversation, not a form
A big reason exit popups underperform is that they present a form to someone who has just shown they don’t want to fill one in. A conversational prompt does better: one question, a couple of quick-reply options, and a path that adapts to the answer. “Did you find what you were looking for?” with “Yes, thanks” and “Not quite” buttons is something almost anyone will answer. “Not quite” opens a conversation; “Yes, thanks” closes the box politely.
This is where exit intent fits naturally with tools that already hold conversations, like Serviceform’s conversational forms and AI agents: the exit moment becomes a chance to answer the question that was about to drive the visitor away. The same thinking appears in why your contact form loses leads: fewer fields, one step at a time, and a reason to keep going.
How to judge whether it’s working
The obvious metric, popup submissions, is the wrong one on its own. A manipulative popup can produce plenty of submissions and still cost you, in unsubscribes, in irritated visitors, in brand goodwill nobody tracks.
Look instead at:
- What the captured leads turn into. Do they book, reply, buy? Compare with leads from other sources.
- Dismissal rate. If nearly everyone closes it straight away, it’s the wrong prompt for that page.
- Effect on the rest of the visit. Did people who saw it go on to view fewer pages? That suggests it’s pushing them out rather than holding them.
A simple hold-out (show it to most visitors but keep a slice who never see it) helps answer whether it’s adding anything at all.
Our view
Exit intent is a timing tool. Like most timing tools, it amplifies whatever you put behind it. Put a cheap trick behind it and you get a faster, more memorable cheap trick. Put a genuinely helpful question behind it and you get a chance to rescue a visitor who was about to leave without the answer they came for.
The popups people hate deserve the reputation. The idea doesn’t.