All Collections
1:1 Product Recommendations
Best Practices
Optimize Your Store Performance Using A/B and Multivariate Testing
Optimize Your Store Performance Using A/B and Multivariate Testing

Test, learn & optimize. Use mathematical proof to optimize your personalized experience for higher performance.

Written by Will Wadman
Updated over a week ago

A/B/n test your recommendations box placement, number of boxes, and types of boxes to see what layout or setup drives the highest number of conversions, throughout your customer journey on your site.

For information on how to use the A/B testing feature in case you want to try it out yourself, check out our get started guide here. Contact your Customer Success Manager if you need any help with setting up an A/B test.

We’ve got a ton of great ideas to get you optimizing your store performance using A/B and multivariate testing. Please pick the area you’re interested in and chat with us for more information.

1. Best recommendation type and placement per page

We’ve written an entire guide on what type of recommendation works best for each page and what is the best placement per box.

But if you're cautious about putting your recommendations where you haven’t before? Simply select a portion of your site traffic and run an A/B test to see how it impacts your sales and conversions.

Example: A clothing boutique could A/B test placing recommendation boxes at the top of a product page, to the side, below the add to cart button, below the description, and below product reviews to see where their recommendations perform the strongest to drive up AOV.

2. Best order to stack product recommendations

A/B/n testing is your friend when it comes to testing stacked recommendation boxes. If you’re worried about overwhelming your customers with choice, try experimenting with how many boxes you stack, or what other content blocks you put between your recommendations (such as reviews or product information).

Example: A furniture store could highlight other items in a ‘Log Cabin Style’ collection based on a user viewing a rustic wooden table, while serving up a second row of ‘Top Trending Products’ that skews toward their bestsellers in this same collection, or across the entire site.

3. What to recommend when a customer signals exit intent

Consider A/B/n testing your exit intent pop up offers to see what keeps customers staying put – A discount? Email signup? Bestselling products? Frequently bought together? There’s more than one type of offer to serve up so don’t be discouraged if your first one doesn’t keep customers shopping.

Example: A skateboard store could promote a free shipping code in an exit intent pop up along with other boards from the same brand as the product page someone is about to exit.

4. What Hero content works best per audience segment - Content Personalization users only

You can use Content Personalization to change the image, text, CTAs, and any other HTML element on a webpage per visitor. You may want to change the order of your home page when, for example, you want to show different content on the desktop vs mobile to a specific user (e.g. returning shopper from an email campaign). You can even replace a section with ‘nothing’ and have that section disappear for certain customers or devices.

A/B testing is the most powerful tool to optimize your content personalization campaigns per audience segment. You may want to consider doing a few A/B/n tests with a small portion of your visitors (e.g 5% of traffic) to make what type of hero image, message or CTA on the homepage creates more customer interaction per audience segment.

Example: Test two different featured collections (e.g. sales vs. new arrival collections) on the home page for your First-time Buyers visitors to see which one created more click-throughs and revenue or reduces bounce rates.

Did this answer your question?