How to Use Wisepops to Measure a Popup's Real Revenue

The short answer: below about 100,000 eligible visitors in two months, a Wisepops control-group test usually ends without a verdict, because its help centre needs roughly 50,000 visitors per variant to detect an uplift of around 10% (results guide); above that, the test beats any dashboard figure, so set a revenue goal, assign it to the campaign, create an A/B test with a 50/50 control group judged on revenue per visitor, leave it untouched for up to eight weeks, then turn the result into a ratio of incremental to attributed revenue for daily monitoring.

Disclosure: I may earn a commission through some links on this site. It does not change what this page says.

Wisepops says the hard part itself. Its revenue impact article states that attribution alone cannot tell you whether a campaign caused extra revenue, and that in its experience 20% to 80% of attributed revenue is truly incremental, depending on the campaign type. That range comes from the vendor, and it is wide enough to change a budget decision. The way to find your own number is a group of visitors who never see the campaign. For channels and plans, see the Wisepops glossary entry.

Check these before you start:

• A plan with control groups. The A/B testing article ties control groups and detailed results to the newer plans, and the results guide names the Custom plan, so confirm that your account shows the control group toggle. On Shopify, the $299 a month plan lists A/B testing (App Store listing, checked on 26 October 2026).
• Enough eligible traffic. Eligible means visitors your targeting would show the campaign to. For example, 3,000 visitors a day with an exit-intent popup that fires for 40% of them gives 1,200 eligible visitors a day; two groups of 50,000 need 100,000 / 1,200 = 84 days, past the two months the results guide sets as a maximum.
• Revenue that reaches Wisepops. The Shopify app tracks orders on its own; other sites need a goal that carries a revenue value.
• A baseline from your own shop system. Note last month's revenue per visitor before launch, so the result has something to sit beside.

Step 1: Set the attribution model

In Settings, open Goals and choose the currency and the attribution window, per the goal tracking guide. The default window is 5 days after a click, or the same session if you prefer (attribution article). The experiment's revenue per visitor counts all revenue in each group, whatever the window; the window shapes the attributed figure you will calibrate in Step 9.

Step 2: Create a goal with revenue

Open Goals in the left menu and click Add goal. A URL goal suits a thank-you page with a fixed value. A JavaScript goal carries the real order amount, and its snippet ships with a placeholder, revenue: 100, that you replace with a static or dynamic value; Google Tag Manager can fire it, per the goal tracking guide. Shopify stores skip this step.

Step 3: Assign the goal to the campaign

In the popup editor, click the Goal icon in the left menu and pick the goal. A campaign holds one goal at a time, goal tracking has to be switched on for it, and conversions that happen before the goal is saved are never attributed, according to the same guide.

Step 4: Create the test with a control group

Click the A/B button to the right of the campaign, enable the control group, keep the split at 50/50, set the success metric to Revenue per visitor, click Done and check that the experiment is published, as the revenue impact article shows.

Pick that metric deliberately. The results guide warns against CTR when a control group exists: the control sees no campaign, so its CTR is zero and any variant shows an infinite uplift. The metric is fixed once the test runs.

Step 5: Decide how it ends

The A/B testing article offers two endings: you conclude by hand, or the winning variant rolls out to 100% of traffic automatically, with an optional end date. Against a control group, conclude by hand. You want the number first, and a look at it before anything rolls out.

Step 6: Leave it alone

Do not edit variants once data comes in; the A/B testing article says to publish changes only after the test concludes. The experiment dashboard opens 24 hours after launch and refreshes daily. Keep the whole test under eight weeks, because Wisepops' experiment design article says longer tests drift with promotions, launches and seasonality.

Step 7: Read revenue per visitor

Open the Experiment dashboard, click See results, scroll to All metrics and choose the Revenue tab (revenue impact article). Compare the control group, labelled Baseline, with the exposed group. An uplift shown in grey has not reached significance. The two articles set different bars, 90% confidence in the revenue article and 95% in the A/B testing article, so wait for 95%. A winner is declared on the primary metric only, per the results guide, however strong a secondary metric looks.

Step 8: Let the AI summary draft the readout

The Wisepops homepage lists an AI Summary on every test. Use it as the first draft of the note to your team, not as the verdict. Before it goes out, a person checks three things: the revenue-per-visitor result carries the significance mark, the control group followed the same targeting as the campaign, and the test window held no sale or product launch.

Step 9: Turn the result into a ratio

Incremental revenue is the difference in revenue per visitor times the visitors exposed. For example, if 50,000 exposed visitors each brought $0.11 more than the control, the campaign added 50,000 x $0.11 = $5,500; if the campaign dashboard attributed $10,000 over the same weeks, the ratio is $5,500 / $10,000 = 55%. The revenue impact article recommends this use: once a campaign's ratio is known, watch attributed revenue daily and rerun the experiment from time to time.

Why the default number needs the test

The campaign dashboard leads with attributed revenue, credited to a tracked click within five days, and a large attributed figure is what keeps a subscription renewed; the control group is the one setting that can shrink it. That is a reason to run it, not a reason to distrust the tool.

Where Wisepops says tests go wrong

From the results guide and the stats comparison article:

• Different targeting per variant. When the rules differ between groups, the visitor split drifts from your allocation and selection bias creeps in.
• Old campaign history. A test created on a running campaign leaves its earlier data in the campaign dashboard; only the experiment page counts from the split.
• Checking against Google Analytics. Wisepops reports in UTC, filters few bots and counts one click per visitor per campaign, so its numbers will not match GA or Shopify.
• Two campaigns, one order. When a visitor clicks two campaigns, the last one clicked gets the conversion.

When the question goes beyond the site

A holdout answers it for one onsite campaign. Later, if you want the same exposed-against-unexposed comparison running all the time across channels, a causal attribution tool like Causality Engine does it. Disclosure: I founded it.

FAQ

How long should a Wisepops experiment run?

Until revenue per visitor reaches significance, and no longer than about two months. The revenue impact article says a revenue result typically takes one to several weeks; the results guide sets the two-month ceiling to keep seasonality out.

Why does Wisepops show more revenue than Google Analytics?

They credit sales differently. Wisepops gives the order to the last tracked click inside your window, in UTC, while GA and Shopify may credit email or ads under their own models, and on Shopify the Wisepops figure is net of discounts, per the stats comparison article.

Can I change the success metric after launch?

No. The results guide says it is set when the experiment is created; stop the test and create a new one with the right metric.