Exit Intent: What the Trigger Detects, and What It Costs a B2B Page

Exit Intent: What the Trigger Detects, and What It Costs a B2B Page

A visitor arrives on a campaign landing page from an email you wrote, reads the offer, decides they want to think about it, and moves the cursor toward the address bar. A panel slides over the page asking for their email address in exchange for a checklist.

They already have an offer from you. They already know who you are. The panel has interrupted somebody in the middle of deciding, to ask for a thing they were already being given, and the cost of the interruption is charged against a visitor who was expensive to produce.

That is the B2B version of a mechanism designed somewhere else, and knowing where it came from explains what it does badly here.

What the trigger actually is

On desktop, exit intent is a cursor heuristic: JavaScript watches the position and velocity of the mouse and fires an exit intent popup when the cursor crosses the upper boundary of the page toward the close button, back arrow or address bar. Wikipedia's entry on the pattern puts it as a popup shown when "the cursor moves outside the upper page boundary", triggered by a snippet that measures "the speed and direction of the mouse", and the popup vendors describe the same mechanism. Exit intent pop-ups and exit intent popups are the same mechanism under two spellings, and the published conversion figures quoted for either come from the same ecommerce population.

That is an inference, not a detection. Moving the cursor upward is compatible with leaving, and it is equally compatible with reaching for a bookmark, switching to another tab to check something you just read, or moving the mouse out of the way. The trigger has no access to intent. It has access to a movement that correlates with one.

On mobile there is no cursor at all, and this is where the mechanism becomes something else entirely.

Mobile exit intent is a different product wearing the same name

Mobile exit intent is a different product wearing the same name: with no cursor to read, vendors substitute back button taps, upward scrolling, inactivity and tab switching, none of which is exit specific. Because a touch screen produces no cursor track toward a browser chrome that is mostly not on screen, exit intent mobile implementations have to infer from ordinary interaction instead.

Look at what those four events have in common: none of them is exit specific.

Scrolling up is what a person does when they want to re-read the headline. Inactivity is what a person does when they stop to think, or answer a message, or put the phone down for a minute. A tab switch is frequently a person going to check something in order to come back. On desktop the trigger at least fires late, at the moment somebody is reaching for the exit. On mobile it fires on ordinary reading behaviour, which means it interrupts engaged visitors rather than departing ones.

The design guidance the vendors themselves publish reflects this. A full screen panel on a phone occupies the whole page, which makes a mistimed trigger considerably more disruptive than the same mistake on a laptop.

DesktopCursor heuristic

• Cursor velocity toward the top of the viewport
• Fires as somebody reaches for close, back or the address bar
• Also fires on reaching for a bookmark or another tab
• Late enough to be arguably free

MobileBehaviour proxies

• Back button tap
• Scrolling upward
• Inactivity after a set interval
• Tab switching

What those also meanThe overlap problem

• Scrolling up to re-read the headline
• Pausing to think about the offer
• Checking something in another tab and returning
• Being interrupted by a notification

The published conversion figures come from a different business

The number quoted almost everywhere for this pattern is around 2.81 percent. Wisepops' own exit popup guide carries the sentence "Exit popups convert around 2.81% of visitors on average, with top campaigns reaching 19.63%", and a separate figure of 5.77 percent for popups carrying a single signup field.

Those are honest figures about the population they were drawn from, and that population is overwhelmingly ecommerce. What the published landing-page numbers actually count, and why they do not transfer onto outbound traffic, is the same problem one level up in average landing page conversion rate. The same guide explains the mechanism as "catching visitors mid-exit forces them to pause and consider the offer", and the examples beside it are discount codes, coupon claims, cart recovery and product recommendations. Two of the top results for the term are Shopify app listings. The strongest published use case is cart abandonment, which is a state that does not exist in a considered B2B purchase.

Three things stop those numbers transferring.

The offer is not available. An exit popup converts by putting something in front of somebody at the moment of leaving that is worth more than leaving. In ecommerce that is a discount on a thing they already put in a basket. In B2B there is no basket, no price on the page, and no discount to give.

The conversion is not the same event. That published 2.81 percent conversion on a consumer store is a purchase or a newsletter signup from a person who may buy in the next week. A form fill on a B2B page is a name entering a pipeline that resolves months later, and its value is decided by whether that person can buy at all.

And the single field advice runs against B2B qualification. The finding that one field converts best is real and it is a direct consequence of the form being a price the visitor pays, an argument made in full in landing page lead generation. A one field popup optimises the number of submissions, which is exactly the metric that produces a list of students and competitors when nobody has said what a submission should be worth.

What that row can be asked afterwards is decided at the moment of capture, because an exit-intent pop-up submission and a demo request are indistinguishable in a CRM six months later unless a source and a capture date were stored as fields beside it, which a standalone exit-intent tool has no way to supply.

The interruption is charged to the wrong visitor

Here is the part specific to a page fed by outbound rather than by search or ads.

A visitor who arrived from a cold message is the most expensive visitor on the page. Somebody built the list, verified the address, wrote the message and sent it, and that person read it and chose to click. They arrive knowing what was promised, because you told them, and they are usually there to decide whether to reply.

An exit panel intercepts that person to collect an email address you already have.

That is the whole objection, and it is arithmetic rather than taste. The visitor is already identified. The thing you want from them is a reply or a booked conversation, not a second consent to a lower value asset. Trading a decision in progress for a lead magnet download converts somebody from a possible conversation into a name on a nurture list, which is a downgrade dressed as a conversion.

For a page fed by search, where most visitors are anonymous and most will never return, the calculation genuinely runs the other way. That difference in visitor state by traffic source is the thing to reason from, and it is the same split that decides what a landing page has to do at all.

Before you install one

• ✓Yes: The page is fed by anonymous search or paid traffic
• ✓Yes: The offer on the panel is more valuable than the one on the page
• ✓Yes: You can measure booked conversations, not just submissions
• ~Depends: The panel is suppressed for anyone arriving from a campaign link
• ✕No: The page is fed by outbound and the visitor is already identified
• ✕No: It is enabled on mobile with the same trigger settings
• ✕No: You budgeted for it on a published ecommerce conversion average

What the vendors sell, and how the market splits

Shopping for exit intent popup software therefore reduces to a question the feature lists do not answer, which is whether the tool can be told not to fire. Two checks settle it before any trial. Ask whether display rules can key on a URL parameter, which is what lets you suppress the panel for traffic arriving on a campaign link. Then ask whether the tool can read contact identity from the system your forms already write to, which is what lets you suppress it for people you have already met.

The tools in this category divide into two groups worth telling apart.

The first is popup platforms, where the exit trigger is one display condition among many alongside scroll depth, time on page and inactivity. OptinMonster markets its exit technology as a named, trademarked feature and the category has grown up around consumer and store use cases.

The second is marketing platforms where exit forms are a component of a larger suite. HubSpot sells exit intent forms as part of its marketing product, which is the version most B2B teams will actually encounter, because it arrives switched off inside software they already own rather than as a separate purchase.

The second shape is the safer one to experiment with, and not because the technology differs. It is because the form submission lands in a system that already knows who the person is, which makes it possible to suppress the panel for known contacts and for anyone arriving on a campaign link. A standalone popup tool typically does not have that context, so it shows the same panel to a stranger and to the prospect you have been working for a month.

Measuring it without flattering it

If you run one, the metric to publish is not the panel's own conversion rate. That number is computed over people who were leaving, so it looks free by construction and always reads positive.

Two numbers are needed beside it. The count of submissions that reached a qualified conversation, which is the only test of whether the captured names were worth capturing. And the change in the page's primary conversion over the same period, because the honest question is whether the panel added conversions or intercepted people who were going to convert anyway on the way to reconsidering.

The general form of that error, where a rate improves while the outcome count falls, is set out in conversion rate. It applies with unusual force here because the denominator is chosen by the tool.

There is a related trap on the list side. Names captured this way have declared less than a form fill on the page itself declares, and they land in whatever system your marketing email runs through. Mixing that population into a sending domain that also handles outbound is the failure mode set out in email marketing lead generation, and it is worth deciding before the first panel goes live rather than after the first complaint.

Where the effort usually belongs instead

The problems an exit panel is asked to fix sit upstream of the panel. A visitor leaving without converting has usually either landed on a page that did not continue the promise that brought them, or reached a page whose offer asks for more than the visit justifies.

Which of the two it was is answerable by asking, and the same trigger will carry a one-question exit survey instead of an offer: the answers arrive from people with no relationship to protect, and ten of them naming a missing price or a missing proof point is a fix somebody can make this week, where four hundred captured addresses is a backlog with no owner.

Both are cheaper to fix than to intercept. The message match check takes an afternoon: put the source and the destination side by side and confirm the headline restates the promise. The offer ladder question is whether the page's single ask is proportionate to where the visitor actually is, and where a site leaks the visitors it already has, the leak points in order are covered in lead generation website.

The deeper version is about which stage of the funnel the asset is serving, because a panel is a middle of funnel instrument bolted onto a page that is usually doing a different job. B2B content marketing funnel is the argument for matching the asset to the question the reader is actually asking.

The short version

Exit intent is a heuristic on desktop and a set of loosely related behaviour proxies on mobile, where scrolling up, pausing and switching tabs all fire the same trigger that ordinary reading produces. The conversion figures in circulation, around 2.81 percent on average with 19.63 percent for the top decile, are published by popup vendors and drawn from campaigns whose offer is a discount and whose abandoned object is a shopping basket.

On a B2B page fed by outbound, the panel interrupts the most expensive visitor you have in order to collect something you already hold. If you run one, suppress it for known contacts and campaign traffic, leave it off on mobile until the mobile trigger settings have been read, and measure qualified conversations rather than the panel's own rate.

When the shortage is qualified conversations rather than form fills, see what a first campaign produces against your own market.

Conversion figures above were read from Wisepops' own exit popup guide in August 2026, and the trigger mechanics from the popup vendors' own product and help pages. Verify current figures at source before relying on them.