Poptin With Causality Engine: From Popup Leads to Shop Orders

The short answer: Poptin will tell you what share of views a popup turned into leads. It will not tell you whether a single extra order followed. Poptin, the popup, form and email tool for small shops, counts 2 kinds of conversion, a submitted form or, on a poptin without fields, a button click, and divides them by views (conversion tracking article). No help article reports revenue or orders. That makes this a partial fit, and the honest pairing is about building the missing link yourself: a Shopify coupon per poptin, a GA4 event, and then Causality Engine, which reads GA4's export per channel, with a free first finding at app.causalityengine.ai/start.

Disclosure: I founded Causality Engine and own it, and this page recommends it. Every claim about it links to Causality Engine's own pages.

Where the popup's number ends and the shop's begins:

1. Turn leads into orders on paper. For example, 1,860 leads at an assumed 20% first-order rate is 1,860 × 0.2 = 372 orders. The 20% is a guess, and the coupon report below replaces it.
2. Use the coupon Poptin already creates in Shopify. A Shopify Coupon made under Grow, Coupons also appears in Shopify under Discounts (Shopify coupon article), where its uses are counted against real orders.
3. Keep the signup out of the sales export. Poptin's GA4 guide fires a sign_up event and has you mark it as a conversion (GA4 article); GA4's attribution paths report lumps all key events together unless you pick purchase alone (Google's report help).

What each side records

What Poptin counts, exactly

The rule is short. A poptin with fields scores a lead each time the form is submitted; one without fields scores a click when its button is pressed. Conversion rate is leads, or clicks, divided by views. A Link Button counts only if its Count As Conversion checkbox is ticked, and a design with the Submit button removed counts nothing (conversion tracking article).

The reports article offers scheduled conversion and statistics reports, daily, weekly or monthly, with an Include click conversions option; untick it and click conversions drop out of the total. The manual export is an XLS with visitors, views, conversions and percentage per poptin.

Poptin's email side reports opens, clicks, unsubscribes, complaints and bounces per campaign (campaign metrics article). No revenue there either.

So Poptin claims less than most tools in this series. It claims a lead. Treating that lead as a sale is the reader's step, not Poptin's.

What Poptin sends to GA4

A popup brings no visitor, so no tag is needed on the popup itself. GA4 hears from Poptin only through the Conversion Code box, under Emails and Integrations in the poptin's design, where the GA4 article pastes a gtag call that fires sign_up on each conversion. Several codes can sit side by side, for Google Ads or Meta as well (conversion codes article). The guide's custom-event step that follows is easy to get wrong, so check in GA4's Realtime report which event name arrives before marking anything as a key event.

Two things from Poptin's own pages touch channels. Traffic Source Targeting decides who sees a poptin by where they came from: Direct Visit, Social Networks, Search Engines, AI Engines, Google Ads (detected by GCLID) or a specific URL (traffic source article). Its Direct is Poptin's own reading of the referrer, not GA4's channel. And the welcome emails Poptin sends: no article says their links carry UTM tags, so add them with Google's URL builder or GA4 may book those visits as (direct) / (none), which Google's direct traffic page attributes to links without a clear referral source.

Worked example: one coupon, four claims

For example, with figures invented to show the sums, not taken from any shop: a skincare store on Shopify runs one exit poptin for 30 days. Poptin's report shows 41,300 views and 1,860 leads, a conversion rate of 1,860 ÷ 41,300 = 4.5%. The Shopify coupon WELCOME10, created in Poptin, appears on 412 orders. The store books 2,950 orders.

Meta claims 1,410 of them, Google Ads 1,190 and the email platform 640: 1,410 + 1,190 + 640 = 3,240 orders, which is 3,240 - 2,950 = 290 more than the store took. Those 412 coupon orders sit inside the claims too, since every buyer arrived from somewhere before the popup appeared.

The 412 is the closest thing to a Poptin sales figure, and it is still not a caused figure: some of those buyers had a cart full before the discount appeared.

A causal read leaves the popup out and works on the visits. Per its how-it-works page, it splits Direct back to the channels that sent those buyers, sets each channel's caused share beside last-click, gives each a data-health score and names a next step to test in one region first. If Paid Social's caused share comes out far under Meta's 1,410, the coupon count cannot defend it. Run the regional test before moving money.

The free check: switch the poptin off on random days

Poptin documents A/B tests between poptins, with the winner being the version with more conversions (A/B article). No version is "nothing", so the test can pick a better popup but cannot price the popup. The holdout is by hand:

1. Take 42 days. Draw 21 at random from a hat and mark them off days.
2. On off days, switch the poptin off before midnight; change nothing else.
3. From Shopify, record orders divided by sessions for every day.
4. Compare the average of on days with off days, and subtract the WELCOME10 discount given on on days.

For example, 2.84% on days against 2.71% off days, on 9,600 sessions a day, is (0.0284 - 0.0271) × 9,600 = 12.48 extra orders a day before the discount is paid for. Daily figures are noisy, which is why the days are drawn at random and why 42 is a minimum.

That stands on its own. When the question is the channels behind those sessions, download the Attribution paths CSV with only the purchase key event selected and drop it on the start page, where the preview runs in your browser for nothing. Causality Engine, which I founded, sells the full read for €99 once, excluding VAT, refundable for 30 days (pricing, checked on 11 November 2026).

Limits on both sides

• No popup in the read. It reports channels, so which poptin or offer did the work stays a Poptin and Shopify question.
• A GA4 purchase and a year of it. The read needs 12 months or more of history, and a channel under 1% of credit on both methods is called too small to act on (how it works).
• No sales in Poptin. Until a coupon or a GA4 event joins the two, Poptin's dashboard and the shop's orders do not meet.

By the A/B article's rule, the version that collects more conversions wins, so the bigger discount will beat the smaller one nearly every time, and the margin that difference gives away shows up in Shopify's discount report, never on the Poptin screen that called the winner.

See also

• Poptin: its two meters, plans and complaints
• How to use Poptin: an exit popup that sends a welcome email
• Poptin on the marketplace: Poptin's affiliate program terms
• Causality Engine: the product, its price and its limits
• How to use Causality Engine: six steps from export to decision
• Causality Engine on the marketplace: the affiliate program, as its owner states it
• Poptin vs OptinMonster: one bill for popups and email, or two
• How to export GA4 data for attribution: getting a clean file out

FAQ

Do Poptin and Causality Engine share data?

No. Neither has a connection to the other. Causality Engine takes a CSV exported from GA4 and reads the channels inside it (integrations page); Poptin's part ends at the sign_up event and the coupon codes in Shopify.

Does a Poptin A/B test track revenue per version?

Not in its documentation. Versions rotate per visit, a visitor keeps the same version for the whole visit, and the winner is the one with more conversions (A/B article). Give each version its own Shopify coupon and the discount report splits the orders.

Should sign_up be a key event if I run a causal read?

It can be, as long as the export is filtered. GA4 aggregates every selected key event in the attribution paths report (Google's report help), so choose purchase alone before downloading.