How to Set Up An A/B Test

This guide walks you through setting up, configuring, and managing an A/B test in Digioh so you can compare two or more Campaigns and let real visitor data pick the winner.

What you'll learn

• How A/B testing works in Digioh
• How to set up an A/B test, including traffic controls
• Best practices for getting a valid, trustworthy result
• How to track performance and pick a winning Campaign

How A/B testing works

Digioh randomly buckets each site visitor into one of the Campaigns in your test, according to the traffic split you configure. A visitor's bucket is sticky across sessions, so they'll always see the same variation for the life of the test. Most tests compare two Campaigns, but you can add more for a multivariate test.

Steps to set up an A/B test

All of the settings below live on one page - you don't move between separate screens.

1. Prepare your Campaigns. Create the versions you want to test using the duplicate feature (found under a Campaign's “More options” menu in the Campaign list) - for example, one with a 10% discount and another with a 20% discount. You can also A/B test headlines using Digioh's AI headline generator.

2. Create a new test. Select A/B Testing → Create New Test in the top navigation.

3. Name your test. Give it a short, memorable name (e.g., "10% vs 20% off").
4. Set your overall traffic allocation. Use the Overall Traffic Percent slider to control what share of your site traffic is included in the test at all; visitors not selected won't see any Campaign in the test (useful as a holdout group if you're worried a variation could hurt conversion). For high-traffic sites, 10% is often enough to reach significance; for lower-traffic sites, set it to 100% to reach significance faster.
5. Choose your Campaigns and split their traffic. Under Campaign 1, Campaign 2 (and beyond, using + add another Campaign), select the Campaigns you want to test and set each one's Traffic Percent. An even split (50/50, or 33/33/34 for three-way tests) reaches a statistically significant result the fastest. You can also weight it unevenly - for example, sending a new “challenger” only 20% of traffic against an established Campaign at 80% - if you want to limit a variation's potential impact while it proves itself over a longer period. Keep in mind: a Campaign held to 20% of impressions needs roughly 5x the conversion rate of the others just to contribute equally to your results, so judge test outcomes by total conversions, not just conversion rate, whenever the split isn't even. Note that a Campaign can only belong to one A/B test at a time, and a test can't go live until every Campaign in it is turned on and published.
6. Set your test conditions (optional). Check Automatically Start/Finish A/B Test to end the test automatically after a number of impressions (e.g., 50,000 per version) or after a date range (e.g., the next 30 days), whichever comes first. When a test finishes automatically, Digioh picks the winner based on whichever conversion rate is larger - submissions/impressions or redirects/impressions.

7. Enable revenue attribution (optional). Check Track Revenue for this A/B Test to compare attributed revenue between variations once your test is running. This requires Google Analytics 4 (GA4) connected to your Digioh account first.
8. Save and publish. Click Save A/B Test, then Publish from the A/B test page when you're ready to go live.
9. Track performance. Once your test is running, expand its row on the A/B Testing → Current Tests list (use the chevron on the right) to see real-time impressions, submissions, and redirects for each Campaign, plus a link into more detailed Digioh analytics. If you enabled revenue attribution in step 7, you can compare attributed revenue between variations here too.
10. Pick a winner. Expand the test's row on the Current Tests list and select the trophy icon next to the Campaign you want to promote. This turns off the other Campaign and finishes the test - otherwise, every Campaign in the test could keep showing to visitors at the same time. Or, if you set automatic test conditions in step 6, Digioh automatically picks the winner based on conversion rate once those conditions are met, promotes it to your live campaign, and turns off the other Campaign (you can always override this afterward by turning campaigns back on or off yourself).

Testing tips

• Test one change at a time. If you change multiple things about the design or conditions between Campaigns, you won't know what caused the result - or worse, one change might cancel out another and make it look like neither had any effect.
• Decide your conversion metric up front, and find it in Digioh analytics. Digioh doesn't define what a “conversion” is for you - it could be emails collected, redirects, quiz completions, or something else specific to your business.
• Judge condition changes on total conversions, not just conversion rate, whenever the traffic split isn't even (see step 5 above).
• Always start from fresh Campaign duplicates - see the callout in step 1. This applies to every Campaign in the test, not just the new variation.

That's it

Your A/B test is live. Need a hand? Reach out to support@digioh.com.