Exit Intent Popups: The Complete Guide

What an exit intent popup actually detects

What it catches is narrow: a cursor sliding up toward the tab strip, the back button or the window's close control, plus the occasional harmless false positive from someone reaching for their bookmarks bar or a second monitor placed above the first.

What it misses matters more:

• Every touch device. No pointer, no event, no popup. Ever.
• Keyboard closes. Cmd/Ctrl + W and Alt + Tab leave no pointer trail.
• The back button, if it is clicked without the cursor first crossing the top edge — and on many layouts it is nowhere near the top edge.
• The visitor who simply stops. Someone who reads two lines, loses interest and sits there is leaving in every sense that matters, and exit intent has no idea.
• Exits through the sides or the bottom of the window — dragging to a second screen on the right, or down to the dock or taskbar.

Why the leaving moment converts

Every popup is an interruption, and the cost of an interruption depends on what it interrupts. A popup three seconds into a page covers the paragraph the visitor came for. A popup that lands as they head for the tab bar covers nothing — the session was over anyway. Three things follow from that:

• The attention cost is close to zero. You are not competing with the content; you are the last thing on screen.
• The visitor has context. They have seen the product, the price, the article. The offer lands on someone who understands it — unlike the popup that fires before they have read a word.
• You already lost them. The alternative outcome for an exit popup is a bounce, so any conversion it produces is genuinely incremental in a way a homepage-on-load popup's is not.

The flip side: you get exactly one shot and no build-up. Nobody reads an exit popup twice. That is why the copy rules below are stricter than for any other trigger, and why the offer has to be legible in about two seconds.

Exit intent vs time delay vs scroll depth vs page count

Exit intent is one of six automatic triggers, and picking the wrong one is the most common reason a well-designed popup underperforms. Here is what each one actually sees.

Two more exist and are worth knowing: clicks count fires after N clicks anywhere on the page, which is a decent proxy for "this person is engaged but not finding it", and inactivity fires after N seconds with no mouse move, key press, scroll or touch. Inactivity is the closest thing to exit intent that works on a phone, because a finger that stops moving is a real signal.

What to offer on an exit intent popup

The offer has to do two things: answer the reason they were leaving, and be deliverable on the spot. The second half is where most exit popups quietly fail — a promise that arrives in an email tomorrow is not an answer to someone whose cursor is already on the close button.

Notice that only two of those five are discounts. A discount is the default exit offer because it is easy, but it is also the one that trains people to trigger the popup on purpose and costs you margin on buyers who were going to convert anyway. If you want a broader look at what other kinds of offers look like in practice, our roundup of website popup examples shows the patterns side by side.

A live overlay on a real storefront. Note the second field: a mobile number is a much bigger ask than an email address, and on an exit popup it is usually the thing to cut.

When a plain discount is not enough

If your exit popup has plateaued, changing the mechanism often does more than changing the number. A prize game turns the same discount into something the visitor plays for: they spin, scratch or drop, the claim screen collects their email, and submitting it is what reveals the prize. The outcome is decided server-side so it cannot be rigged from the browser, and you set the odds per prize. Our spin to win popup guide covers the prize table and the failure modes in detail.

The same discount, delivered as a game. Losing segments are part of the design — they are what make the winning ones feel earned.

How to write an exit intent popup

You have roughly two seconds and one screen. Five rules cover almost everything:

1. Put the offer in the headline. Not the brand, not "Don't go!", not a question. The headline is the offer, in plain words, with the number in it.
2. One field. Email is nearly free once someone has decided to convert. A phone number, a name and a dropdown are three more decisions on a screen that has no time for any. If you genuinely need more, collect it later — that is exactly what progressive profiling is for.
3. Label the button with what they get. "Claim my 20% off" beats "Submit" every time, because the button is the last thing they read before deciding.
4. Give them a visible way out. A real close button plus a text decline link. It sounds counterproductive; it is the difference between a dismissal and a bad impression.
5. Deliver on the thank-you screen. Show the code right there, not "check your inbox".

A worked example, line by line

Here is a complete exit popup for a coffee store's product pages. Every line is paste-ready:

Three things to note. The eyebrow does the "you are leaving" work so the headline does not have to waste words on it. The decline link is written in the visitor's voice, which makes dismissing feel like a choice rather than a defeat. And the code is on the thank-you screen because ChilliPopup does not send email — the welcome message comes from your own email tool, later, from the addresses you collect. If you want to go deeper on the wording itself, the popup copywriting guide breaks down headline and button patterns.

The Design step. The Main and Thank you tabs above the preview are two separate screens — and the little monitor and phone icons next to them let you check both layouts before you publish.

The display rules to pair with exit intent

The trigger decides when. Everything that decides who, where and how often lives in the display rules, and this is where an exit popup either behaves or becomes the thing people complain about.

• Device: desktop. Set it explicitly. The trigger cannot fire on a phone anyway, but being explicit keeps the campaign honest and makes the analytics readable. (The player treats a viewport of 768px or narrower as mobile.)
• Page targeting. Six URL operators — contains, does not contain, is exactly, is not, starts with, ends with. Use one rule to include (contains /products/) and one to exclude (does not contain /checkout). Every enabled rule is an AND gate.
• Audience. Everyone, new visitors or returning visitors, where you define "new" and "returning" in days — 30 by default. A first-order discount belongs to new visitors.
• Frequency. Once a day, every visit, every session, once every N days, once every N visits, or all pages all the time. Once every 7 days is the sane default for an exit offer: a dismissal is remembered for a week rather than repeated on the next page.
• Show again after conversion — turn it OFF. This is the single most important switch on the screen. With it off, anyone who submitted that campaign stops seeing it. Leave it on and your own subscribers meet the signup popup forever.
• Schedule and date window. A weekday-and-hour schedule and a start/end date, for offers that expire.

Mobile: what to run instead of exit intent

This is the section most exit intent guides skip. If a large share of your traffic is on phones — and for most stores it is — an exit intent campaign on its own reaches only part of your audience. The fix is not a cleverer trigger. It is a second campaign.

Build a duplicate of your exit popup, set device targeting to mobile, and swap the trigger:

• Page scroll at 50–60%. The best all-round substitute. Someone who has scrolled past halfway has consumed the page, which is the same "I am nearly done here" state exit intent is guessing at. On short pages that do not scroll, no scroll event fires and the popup never opens — use a time delay there instead.
• Time delay of 20–30 seconds. Blunt but reliable. It is counted from page load, always, so treat it as "seconds on this page", not "seconds in the visit".
• Inactivity of 30–45 seconds. The closest mobile analogue to exit intent: no scroll, no tap, no touch means attention has gone somewhere else.
• Page count of 3. Good for stores, where browsing three products is a real intent signal. It counts pages within the browser session, including the current one.

If you want a true mobile exit signal, script it

There is no built-in "mobile exit" trigger. The closest usable proxy is a fast upward scroll — the flick back to the top that usually precedes a tab close — and you can wire that up yourself with the JS API in about twenty lines:

(function () {
  // Read the flag first, so it also guards the next page in the same session.
  var fired = false;
  try { fired = sessionStorage.getItem('exit_shown') === '1'; } catch (e) {}
  if (fired) { return; }

  var last = window.pageYOffset;
  window.addEventListener('scroll', function () {
    var y = window.pageYOffset;
    // jumped up by 400px or more, and we are near the top again
    if (!fired && last - y > 400 && y < 200) {
      if (!window.ChilliPopup) { return; }   // pixel not ready yet
      fired = true;
      try { sessionStorage.setItem('exit_shown', '1'); } catch (e) {}
      window.ChilliPopup.close();            // never stack overlays
      window.ChilliPopup.open('YOUR_POPUP_UUID');
    }
    last = y;
  }, { passive: true });
})();

When an exit intent popup is the wrong tool

Four situations where the right move is to not run one:

• Checkout and payment pages. A visitor moving their cursor toward the top on a payment page might be checking the padlock icon, not leaving. Interrupting a purchase in progress is the one place a popup can cost you the sale it was meant to save. Exclude those URLs explicitly.
• Mobile-dominant traffic with no companion campaign. If most of your visitors are on phones and you only build the exit version, you have built a campaign for the minority. Build the mobile one first, then add exit intent for desktop.
• Support and documentation pages. Someone trying to solve a problem does not want 10% off. If you want something from them here, ask a one-question survey about whether the page helped.
• When you have nothing to offer. An exit popup with no offer is a goodbye message with a form attached. Fix the offer before you fix the trigger.

There is a timing case too: a one-weekend sale rarely justifies building an exit popup from scratch, because by the time you have enough data to judge it the campaign is over. Build the exit popup once, keep it running, and change the offer inside it when the season changes.

How to tell whether it worked

Every popup gets its own analytics page: Views, Unique views, Clicks, Submissions, Conversion and Field inputs, an interactions-over-time chart, an event breakdown, a shows-by-device split, a Collected fields breakdown of the top values per field, and a Collected input data table of the raw submissions.

Views, unique views, clicks, submissions and conversion in one row — and a shows-by-device split that will tell you immediately whether your exit campaign is behaving.

Four things to actually look at, in this order:

• Shows by device. Start here. An exit intent campaign should be showing almost entirely on desktop. If you see meaningful mobile shows, something else is triggering it — check whether a second trigger is enabled.
• Views versus site traffic. If views are tiny, the problem is the rules, not the copy. Too narrow a URL rule, too aggressive a frequency cap, or a conversion gate doing its job on a list that already converted.
• Conversion. The dashboard divides submissions by views, so read it as "of everyone who saw this, how many finished". Compare a campaign against itself over time — never your exit popup against your time-delay popup, because they are shown to different people in different states.
• Collected fields and the raw submissions table. Junk addresses and obvious throwaways tell you the offer is attracting people who want the code and nothing else. That is a signal about the offer, not the trigger.

To show the arithmetic — and to be clear, these numbers are invented for the example: a popup with 4,000 views and 300 submissions reports a 7.5% conversion. Add a stricter URL rule so it only shows on product pages and views drop to 1,600 while submissions hold at 260 — conversion jumps to 16.3%, and you kept 87% of the leads. That is a better campaign, and the headline conversion number moved for reasons that have nothing to do with the copy. Always read conversion next to views.

Build your exit intent popup in ChilliPopup

About thirty minutes, six steps, no code. If the pixel is not on your site yet, start with the one-line install guide — the same script serves every campaign you ever build.

Related reading

Frequently asked questions