Exit intent fires when a visitor's behaviour suggests they are about to leave, which means it interrupts only people who were already going. That is the strongest argument for it, and it does not work on phones, where most visitors are, so the reach is smaller than the concept suggests.
How it works
On a desktop, the cursor moving rapidly toward the top of the window is treated as a signal that the visitor is heading for the close button or the address bar.
That is a reasonable inference and it is only a guess. Someone reaching for a bookmark, another tab, or a browser control triggers it identically.
The reason it appeals is the timing. A message interrupting someone reading is a cost; a message appearing as they leave costs nothing they had not already decided to give up.
The mobile problem
There is no cursor on a phone, so there is no reliable exit signal.
Implementations substitute proxies: scrolling up quickly, a period of inactivity, or a back gesture. None of these are exit intent, and each fires on people who were still reading.
Since most visitors to a small business site are on a phone, this means exit intent reaches a minority of the audience, and the mobile version is an interruption rather than a farewell.
The defensible position is to run it on desktop only and accept the smaller reach, rather than to approximate it badly on mobile.
When it is worth trying
Where you have something genuinely useful to offer
A checklist, a guide, a seasonal reminder, or a price list. Something with standalone value that a person who was leaving might actually want.
On research pages rather than service pages
Someone reading an explanatory article was not about to book. Offering to send them something is proportionate. On a service page, the same interruption competes with an action they might still take.
Where the buying cycle is long
Renovations, major installations, and anything researched for months. Capturing a contact from someone six months out has real value, and no other mechanism reaches them.
When it is not
- On a page where the visitor can act now, since you are interrupting a live decision to offer a lesser one.
- Offering a discount, which trains people to leave in order to be offered one.
- With nothing to give, where the popup is a newsletter signup with no stated benefit.
- On a phone, for the reasons above.
- Where enquiry volume is already fine, since the marginal gain is small and the annoyance is real.
The parts people get wrong
Frequency. Showing it on every visit to a returning visitor is the fastest way to make a site unpleasant. Once per person, remembered, is the correct setting and it is frequently left at the default.
Dismissal. A visible close button of reasonable size, closing on the escape key, and closing on a click outside. Anything harder reads as a trap.
Delay. Not firing at all in the first several seconds, since someone leaving immediately was never engaged and interrupting them achieves nothing.
Accessibility. Keyboard focus should move into the dialogue and be returned when it closes, and it should be announced properly. This is routinely neglected and it makes the popup a trap for anyone not using a mouse.
Measuring it honestly
The reported conversion rate of an exit popup is misleading on its own, because it counts only the people who saw it.
What matters is whether total enquiries rose. A popup converting at a few percent of the people who see it looks impressive and can coincide with no change in enquiries at all, if the people it captured would have made contact anyway.
The honest test is running it for a period, then removing it for a comparable period, and comparing total enquiries rather than popup conversions. Few businesses do this, which is why the case for these is weaker than the reported numbers suggest.
The alternative worth considering first
Everything exit intent tries to achieve can be attempted with a static offer at the end of a page, which interrupts nobody.
Someone who reached the bottom of an article has read it, and an offer there reaches an engaged reader without any interruption or accessibility complication. It converts at a lower rate and costs nothing in annoyance.
Trying that first is the sensible order, and if the end-of-page offer produces nothing, that is useful information about whether the offer is wanted rather than about the delivery mechanism, which is the same distinction drawn in urgency that is honest and urgency that is not.
Frequently asked questions
What is exit intent?
A message triggered when a visitor's behaviour suggests they are about to leave, usually the cursor moving rapidly toward the top of the window on a desktop.
Does exit intent work on phones?
Not reliably. There is no cursor, so implementations substitute proxies such as scrolling up or inactivity, which fire on people who were still reading.
When is exit intent worth using?
On research pages where the visitor was not about to book, where you have something genuinely useful to offer, and where the buying cycle is long enough that a contact months out has value.
Should an exit popup offer a discount?
No. It trains people to leave in order to be offered one, and it devalues the price you quoted on the page they were reading.
How do I measure it honestly?
By whether total enquiries rose, not by the popup's own conversion rate. Run it for a period, remove it for a comparable one, and compare total enquiries.
Is there an alternative?
A static offer at the end of the page, which reaches an engaged reader, interrupts nobody, and avoids the accessibility complications. Worth trying first.
West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.
Considering a popup to catch people leaving?
We would try the end-of-page version first, which converts more slowly and costs nothing in goodwill.
Start a Conversation