How to A/B test your popups (and find a real winner)

A/B testing replaces opinions with evidence. Instead of guessing whether a headline or offer works better, you show each to a slice of traffic and let conversions decide. Here's how to do it well.

What to test (one thing at a time)

• Headline — the single biggest lever
• Offer — discount vs free shipping vs content
• Call-to-action wording
• Format or timing (e.g. exit-intent vs delay)
• Two-step (yes/no) vs a direct form

Change one variable per test. If you change three things and conversion improves, you won't know which one did it.

Set up the test

1. 1

   Create Variant A and Variant B

   Keep everything identical except the one element you're testing.

2. 2

   Split traffic

   Send a set percentage to each variant — 50/50 is standard.

3. 3

   Pick the success metric

   Use conversion rate (clicks or submissions ÷ impressions), not raw clicks. A variant with more impressions can have more clicks but a worse rate.

4. 4

   Let it run

   Wait for enough conversions per variant before deciding — ending early on a small sample produces false winners.

In NounDesk, enable A/B on a campaign, set the traffic split, and the analytics show each variant's conversion rate side by side so you can promote the winner.

Reading the result

Look at conversion rate, not totals, and make sure the gap is meaningful and stable — not a day-one blip. Once a variant clearly wins, ship it as the new baseline and test the next idea. A/B testing is a habit, not a one-off.