Popups

Storefront interventions that fire on real behavioural signals — with a held-back control group, so the reported revenue is incremental rather than gross.

A popup that appears after five seconds is a guess applied to everyone. A popup that appears when a shopper has rage-clicked the variant selector three times is a response to something that actually happened.

Popups is the intervention layer built on the data the tracker already collects. Same frustration, hesitation and intent signals that drive heatmaps and suggestions — used live, in the session, while the shopper is still there.

Building an intervention end to end — component, trigger, conditions, preview, publish.

Component types

Eight ready-made components, all responsive, all rendered in your palette and typography via Brand DNA:

Component	Form	Typical job
Promotional popup	Centred overlay	Discount offer, lead capture, announcement
Sticky announcement bar	Persistent top bar	Store-wide sale, free-shipping threshold, urgency
Exit-intent rescue	Overlay on leave intent	One last offer to a session that is going
Cart nudge	Bottom-corner card	Free-shipping progress, cart reminder
Social proof toast	Slide-in notification	Recent-purchase signals
Combo / bundle offer	Overlay	Cross-sell, raising average order value
Countdown timer	Urgency bar	Flash sale, time-limited promotion
Product badge	Label on product cards	"Selling fast", best-seller highlighting

Behavioural triggers

This is what separates Popups from a popup app. Triggers read state, not elapsed time:

Shopper state	Detected from	Sensible response
Frustrated browsing	Rage clicks, dead clicks, erratic cursor movement	Offer help, or a small discount for the trouble
Price hesitation	Long dwell on a product, variant switching, no add to cart	Surface a code, or address the specific objection
Cart abandonment risk	Items in cart plus stalling or leave intent	Free-shipping nudge, or completion prompt
High intent	Multiple product pages, deep scrolling, return visit	Welcome back, items still available
Scroll passers	Scrolling past a key section without engaging	Re-surface the missed CTA

Manual triggers are available too — exit intent, scroll depth, time on page, cart value — for the cases where you already know what you want.

Building one

Pick a component

Or start from an AI suggestion. Popups reads your behavioural data and proposes interventions with the reasoning attached — "38% of product-page visitors rage-click the size selector; show a size guide with a small discount" — which you can accept, edit, or ignore.

Write the content

Headline, body, promo code, button text. Or generate it — the copy is conditioned on your extracted brand voice, and imagery can be generated and background-removed in the editor.

Choose the trigger

Behavioural or manual. This is the decision that determines whether the nudge is an intervention or an interruption.

Set conditions

Which pages, frequency per visitor, schedule window, and suppression rules so repeat visitors are not hammered. Frequency capping is not optional politeness — an over-fired nudge trains shoppers to dismiss on sight and poisons the measurement.

Preview, then publish

Desktop and mobile preview in the editor, then live. If the content includes a promo code, Popups creates the matching Shopify discount for you rather than making you set it up separately — and tears it down if the nudge is paused, so a paused nudge never leaves a live redeemable code behind.

Attribution: net, not gross

The number Popups reports is incremental

A share of eligible sessions is deliberately held back as a control group and shown nothing. Reported revenue is the difference between the treated group and that holdout — so it excludes the orders that would have happened anyway.

This matters because the easy version of this feature is to sum the revenue of every session that saw a nudge and call it "nudge revenue". That number is always larger and always wrong. A shopper who was going to buy, then saw a toast, is not revenue the toast created.

Available per intervention and in aggregate:

Live dashboard — impressions, clicks, conversions today
Per-intervention analytics — impressions, CTR, conversion rate, revenue trend
Attribution view — incremental revenue per intervention, by month
A/B testing — traffic split automatically between variants, with impressions, clicks, conversions and revenue per variant and a called winner

The same net figure feeds the dashboard's earned-this-month headline and the Impact page, so the three surfaces cannot disagree with each other.

Plan limits

Active interventions are capped per plan: 2 on Free, 5 on Growth, 15 on Pro, unlimited on Custom. A/B testing starts at Growth. Free can build and preview but not publish.

If a plan changes downward — a downgrade, or a failed charge — the newest interventions up to the new cap stay live and the rest are paused, not deleted. You are told which ones stopped, and their discounts are torn down with them. They come back on upgrade.

Limits

A nudge is a response to friction, not a fix for it. A discount shown to someone rage-clicking a broken selector rescues that session and leaves the selector broken. Nudge the symptom while you fix the cause.
Incremental attribution needs volume. The holdout comparison is a difference between two groups; on low traffic that difference is noisy. Early numbers on a small store are directional.
Triggers fire on observed behaviour, so they inherit its ambiguity. Long dwell on a product page might be price hesitation or a phone left face-up on a table.
Frequency capping is your responsibility. The controls are there; the defaults are conservative; a nudge set to fire on every page view will still do that.
Interventions render on your storefront theme. A heavily customised or unusually structured theme can position a component imperfectly — which is what the desktop and mobile preview is for, before publishing.

Cart Overview

Where the abandonment is, so the nudge targets the real step.

CRO Report

Sizes the leak a nudge is rescuing, so you know what it is worth.

Impact

Nudge revenue joins fix revenue in one measured monthly figure.

Common questions

Will this hurt my Core Web Vitals?

Nudges render client-side after the page is interactive and are not part of the initial layout, so they do not contribute to LCP or CLS on first paint. Storefront Speed will show you the real numbers on your theme.

Can I stop nudges from stacking on top of each other?

Yes. Frequency capping and priority rules mean a shopper sees at most the highest-priority eligible nudge, not every rule that happened to match.

Do nudges work on mobile?

Yes, with mobile-appropriate triggers — exit intent is desktop-specific, so on mobile it is replaced by signals like rapid upward scroll, back-navigation intent and dwell patterns.

How is attribution calculated?

Sessions eligible for a nudge are split, with a portion held back as a control. The revenue difference between exposed and held-back sessions is what gets reported, rather than crediting the nudge for every order that followed it.