Add-to-cart rate is one of the strongest signals of how well your product page converts. If shoppers don't add items to their cart, they won't reach checkout, no matter how much traffic you drive.
Instead of guessing what works, you can run an A/B test on your product page (PDP) with GemX Template Testing to validate which version actually drives more add-to-carts. This guide walks you through the full setup from picking what to test to applying the winner.
What to Test on Your Product Page
The product page is where shoppers make their buy-or-leave decision. Small changes here often produce an outsized impact on the add-to-cart rate.
Common high-impact elements to test:
-
Add-to-cart button: copy, color, size, or placement
-
Sticky add-to-cart bar: especially on mobile
-
Product images: gallery layout, image vs. video, zoom UX
-
Trust signals: reviews, badges, or guarantees above the fold
-
Pricing display: compare-at price, savings %, urgency text
-
Variant selector: dropdown vs. color/size swatches
Pro tip: You should only pick one change per variant. If you test multiple changes at once, you cannot define which change actually moved the metric.
How to Set Up an A/B Test on PDP to Increase ATC Rate
Step 1: Build your hypothesis
Every test starts with a clear hypothesis. You can use this format to ensure a clear and testable hypothesis:
| "If we [change], we expect [metric] to [direction] because [reason]." |
Example: If we move customer reviews above the add-to-cart button, we expect add-to-cart rate to increase because shoppers will gain trust before deciding.
The key is always back your hypothesis with data, such as page analytics, heatmaps, or session recordings, which can show you where shoppers drop off or hesitate on your PDP.
Learn more: How to Use Page Analytics to Spot Bottlenecks
Step 2: Create a Template Testing experiment
Once you know what to test, set up the experiment in GemX.
From your GemX dashboard, click Create new experiment.
Select Create template experiment to use Template Testing feature.
1. Under the Winning metric box, set Conversion as your primary goal metric.
|
Why choose Conversion instead of Revenue? GemX lets you choose between two winning metrics: Conversion or Revenue. For add-to-cart tests, choose Conversion — it's the most stable metric tied to actual purchases, and it reaches probability to win faster than Revenue. You'll still see Add to Cart, click-through rate, and other metrics in the Performance detail panel as supporting data, but the winner is decided by your selected winning metric. Pick Revenue instead when you're testing pricing, bundles, or upsells — changes that directly affect order value. |
Learn more: How to Choose the Winning Metric (Primary Metric) in GemX
2. Under the template section, you'll assign the testing templates for the Control and the Variant.
Click Select a template in the Control box.
You can choose Product page as the page type to filter product page templates only.
Pick your live product page template as the Control.
Continue to select a product template to assign it to the Variant. You can choose from existing templates, or duplicate from the Control.
Important note: Please note that the "Create Variant based on Control" option is only available if the Control template is built with GemPages.
Then, you can name your experiment clearly (e.g., PDP – Reviews above ATC button).
Learn more: How to Create a Template Testing Experiment
Step 3: Edit your variant
With your experiment created, apply your hypothesis to the Variant.
Open the Variant in your theme editor, or in GemPages, if your PDP is built with it.
From here, make only the change defined in your hypothesis.
Once done, save and preview the variant to ensure it render correctly on both desktop and mobile. Then you can publish it and go back to GemX for further configuration.
Important note: If your PDP is built with GemPages, see Connect GemX with GemPages for the correct editing flow.
Step 4: Set traffic split & audience
Configure how visitors are distributed across your test.
By default, the traffic split is set to 50/50 split between Control and Variant. You can change it by dragging the slider.
Moreover, you can segment your traffic with optional targeting: device type, country, traffic source, or new vs. returning visitors.
Tip: For PDP tests, you should target only product page traffic to keep results clean.
Step 5: Launch & monitor
Now you're ready to go live. Click Start experiment to launch your experiment.
Then, you can access the experiment analytics window to view live data.
Click View all metrics to open the Performance detail panel. From here, you can track add-to-cart rate, sessions, and probability to win for each variant.
Pro tip: Please wait until you reach statistical significance, typically 14+ days or 1,000+ sessions per variant.
Don't stop the test based on early results, as peeking too soon causes false positives and bad decisions.
Step 6: Apply the winner
When your variant reaches 95%+ probability to win, you can lock in the result.
Open the experiment from your dashboard and click Make Winner to apply the winning template.
Then, the progress will be automatic, and you don't have to take any further actions. The winning variant becomes the live product page for 100% of traffic.
You can document the insight and use it to inform your next PDP test.
Learn more: How to Apply the Winning Template
FAQs
