Choose the first meaningful moment
The shopper should normally understand the page before a promotional modal interrupts it. The right trigger is the earliest event that demonstrates enough context or intent for the offer to be useful.
Google also advises against promotional interstitials that obscure page content when users arrive. Review the current Google interstitial guidance when using immediate full-page experiences, especially on mobile.
Immediate display: visible, but most interruptive
Immediate display can fit a storewide operational announcement or deliberate test, but it is usually a poor default for a promotional modal. The visitor has not yet read the product, collection, or landing-page promise.
- Do not make immediate full-screen mobile display the default.
- Warn when immediate display is paired with every refresh.
- Load after campaign eligibility is known rather than blocking the page parser.
- Consider a small banner for information that truly must appear immediately.
Time delay: simple and predictable
A delay gives the page time to render and the shopper time to orient. It is easy to explain and test, but elapsed time alone does not prove engagement: a tab can remain open while the shopper is elsewhere.
Start with a modest delay appropriate to the page, then compare outcomes. Product pages may establish intent faster than a long editorial guide. Avoid publishing a precise universal benchmark as if every store, traffic source, and offer behaves the same.
Scroll depth: evidence of page engagement
Scroll triggers are useful when the offer should follow meaningful browsing. They work best on pages with enough content to reach the threshold naturally.
- Use a threshold that exists on short mobile pages as well as desktop pages.
- Do not count automatic layout movement as shopper scroll.
- Use passive listeners and avoid expensive work on every scroll event.
- Recalculate carefully when images load and page height changes.
- Do not use scroll depth on a page where the CTA is already visible above the fold.
Exit intent: useful on desktop, different on mobile
Desktop exit intent can respond when a pointer leaves the document near the browser edge. Touch devices have no pointer departure, so a mobile implementation needs a separately described fallback such as engaged upward scrolling or another configured trigger.
Do not label an unrelated mobile event as identical exit intent. Tell the merchant what signal is actually used, and test false positives caused by normal navigation gestures.
Timing cannot compensate for excessive frequency
A well-timed popup still becomes disruptive if it opens on every eligible page. Choose whether it may appear every refresh, once per session, once per visitor, or after a repeat period.
| Situation | Reasonable starting approach |
|---|---|
| General promotion | Engagement trigger with once-per-session frequency. |
| Short operational announcement | Small, clearly dismissible format with limited repetition. |
| Product-specific offer | Eligible product paths plus scroll or modest delay. |
| Potential abandonment | Desktop exit intent with an explicitly different mobile fallback. |
| Testing only | Every refresh in a controlled test context, never forgotten in production. |
Test the business outcome honestly
- Choose one primary outcome before publishing.
- Record delivery, dismissal, CTA click, conversion signal, and offer-assisted revenue separately.
- Keep a holdout or A/B comparison when enough traffic exists.
- Check mobile and desktop separately.
- Watch bounce, conversion, and storefront performance—not only popup click rate.
- Run long enough to avoid reacting to a handful of sessions.
Observed association is useful, but it should not be described as guaranteed causal lift. The product keeps trigger, frequency, and variant decisions explicit so merchants can interpret the result.