Every proactive trigger watches exactly one signal. You pick it under When it fires, and it is the single most consequential choice on the form, because it decides not just whether the chat opens but how soon.

This article goes through all eight: what each one actually measures, the range you can set, the default, and where each one can silently never fire.

Fires after a visitor lingers on a matching page.

A timer starts when the page loads and the trigger fires when it runs out, provided the rest of your rules still hold.

Good default territory: 30 to 60 seconds on a product page is long enough to mean genuine interest and short enough to catch someone before they leave. Under 10 seconds tends to feel like an interruption rather than an offer of help.

Fires once the visitor scrolls past a set point.

The case where it cannot fire: if a page is too short to scroll, there is no scroll to measure and the trigger never fires there. A scroll-depth trigger set to fire on Any page will therefore work on your long product pages and do nothing at all on a short policy page. That is not a fault, but it does mean scroll depth is the wrong signal for a site of short pages.

Fires when the cursor moves to leave the tab.

The classic last-chance signal. It watches for the pointer travelling up and out of the top of the browser window.

The case where it cannot fire, and it is a big one: exit intent never fires on a phone. It is disabled below a screen width of 768 pixels, because a touch device has no cursor and therefore no such thing as a pointer heading for the close button. This holds even if you set Device to Mobile or Any under Who sees it. A mobile-targeted exit-intent trigger is a trigger that will never fire, and nothing on the page will tell you so. If mobile matters to you, use Inactivity or Time on page instead.

Fires when items sit in the cart untouched.

This is the one genuine cart-watching signal in Arbyn. It checks that the cart has items, that the cart's value clears any minimum you set, and that the cart has not changed for the length of time you chose.

A timing detail worth knowing: the cart is checked on a repeating poll roughly every 15 seconds rather than the instant a cart changes. So there can be up to about 15 seconds between a shopper's cart going quiet at the threshold and the trigger noticing. Choose thresholds in minutes and this never matters.

Sensible thresholds, and why. Under 2 minutes is usually too soon: a shopper reading a size guide in another tab has not abandoned anything, and a chat window opening on them is an interruption. Around 5 to 10 minutes is the range where "they have stopped" is a fair reading. Beyond about 30 minutes you are mostly catching people who have already left the tab open and walked away, which is a weaker moment but still a real one if your discount is worth it.

Fires when the cart passes a value you set.

Use it for the offers that only make sense on a large basket: free shipping, a gift with purchase, or simply a human offering to help before someone spends a lot.

Fires on a repeat visit to your store.

This is the signal that genuinely means "came back later", because it enforces the gap you set. It is worth contrasting with the Visitor filter under Who sees it, which is a different control with a much looser meaning. That distinction is covered in its own article, and it catches almost everyone out at least once.

The case where it cannot fire: the last-visit time is stored in the visitor's own browser. If your store uses a consent banner and the shopper has not accepted marketing cookies, that time is never recorded, and a returning-visitor trigger will not fire for them. It is also per browser and per device, so the same person on a phone and a laptop looks like two different visitors.

Fires when one product is viewed several times.

A strong buying signal, and an underused one. Someone who has come back to the same product three times is deciding, not browsing.

The case where it cannot fire: same as Returning visitor. The count lives in the visitor's browser and depends on marketing consent, and it does not follow them between devices.

Fires when the visitor goes idle on the page.

The catch-all attention signal: no mouse movement, no typing, no touch, no scrolling, no clicks for the time you set.

This is the mobile-friendly alternative to exit intent. It has no cursor requirement, so it works everywhere.

Set it carefully. Inactivity is the fastest-firing signal on this list at its lower settings, and because triggers do not resolve collisions by priority, a short inactivity threshold will consistently beat every other trigger you have configured. If you want a considered signal to win, do not put a 5-second inactivity trigger beside it.

Which signal should I start with? Cart abandoned at 5 minutes, on a shopper whose cart has items. It has the clearest intent behind it and the least risk of feeling random.

Why does my exit-intent trigger get no mobile fires at all? Because exit intent is disabled on phones by design. There is no cursor to watch. Use Inactivity for mobile.

Can one trigger watch two signals? No, one trigger watches one signal. Build a second trigger, and read the article on how often and collisions before you do, because two triggers do not queue politely.

Does Time on page count time across my whole site? No, it restarts on every page load. For a whole-visit measure, use Returning visitor or Repeat product views.