All Latest 617 A/B Tests

This experiment ran on desktop only, on all our regular product listing and search result page. In the control, clicking on products would open them up in the same page, whereas in the variation, the links opened in a new tab. Impact on adds-to-cart were measured.

This experiment ran on desktop only, on all our regular product listing and search result page. In the control, clicking on products would open them up in the same page, whereas in the variation, the links opened in a new tab. Impact on adds-to-cart were measured.

In this experiment, more popular search categories (pills) were shown on the homepage based on analyzed search volume. Impact on searches and email subscriptions were measured.

The A version showed a non-sticky checkout button at the bottom of a flyout cart. The B variation added a sticky bar with a price summary and checkout button, keeping it always visible. Impact on sales was measured.

A global flyout drawer was added that consolidated sale promotions into one centralized location, replacing the customary placement of promotions throughout the site.

The control had a clickable link "Use discount code" which would reveal an input field for a coupon code. The variation removed this link and input field. Impact on purchases was measured. (Desktop only)

The control had a clickable link "Use discount code" which would reveal an input field for a coupon code. The variation removed this link and input field. Impact on purchases was measured. (Mobile only)

A floating Add to Basket button was added to a product page. Impact on sales was measured.

In this experiment, the sorting link (defaulting to popular) was swapped with the filter one. In the control, the sorting appeared on the left with the filter on the right, whereas in the variation these two were flipped. Impact on adds to cart and sales were measured.

In this experiment, a larger vs smaller font size for the yearly price was tested. Impact on annual subscription clicks was measured. (The experiment has been inverted to fit the pattern.)

The control had a clickable link "Use discount code" which would reveal an input field for a coupon code. The variation removed this link and input field. Impact on purchases was measured. (Desktop only)

The control had a clickable link "Use discount code" which would reveal an input field for a coupon code. The variation removed this link and input field. Impact on purchases was measured. (Mobile only)

In this experiment, different reassurance messages were shown at the bottom of the add-to-cart widget on a product page. The variation emphasized full refunds for faulty or misrepresented items.

In this experiment, two changes were made to the checkout page: (1) the minimum basket size requirement was made more visible using a floating element, and (2) an additional threshold was introduced to encourage customers to add more items by offering a free gift at a higher spend level. The impact on sales was measured.

In this experiment, the shopping cart screen displayed a floating “Checkout” button with common credit card icons that directed users to the checkout page when clicked. In the control, the button was positioned inline, further down the page. Impact on progression to checkout and total sales were measured.

In this experiment, a pricing table (an upsell of product protection plan coverages) was tested against a more explicit yes/no toggle for including the plan. Impact was measured on adds to cart, orders, and upsell rate, using an 80/20 traffic split.

In this experiment (1) the zip code field position was moved up, right below the last name. And (2) entering the zip code would populate the state and city using an autofill API call. Impact on orders completed was measured.

In this experiment (1) the zip code field position was moved up, right below the last name. And (2) entering the zip code would populate the state and city using an autofill API call. Impact on orders completed was measured.

In this tightly triggered experiment, users who (1) did not complete a purchase and (2) visited a different product page saw a button in the top-right corner labeled “Don’t Forget Your Club.” Clicking this button resumed the checkout process at the point where they left off, without requiring the same information to be re-entered.

In this experiment, two types of UI filters for metal color choices were shown at the top of product listing pages. One variant only used category labels, while the other variant combined images with labels to reinforce the categories. Impact on filter usage, adds to cart and sales were measured.