All Latest 620 A/B Tests

In this experiment, the footer was simplified with multiple elements being removed (catalog request, secondary links, and a guarantee). Additionally, a more prominent phone number was also displayed.

In this shopping cart experiment, a tall footer was completely removed. The footer contained elements such as: app download links, a newsletter signup, secondary web site links, trust symbols, social media icons and company contant information. Impact on sales was measured.

In this experiment, a count down timer was added near the top of a checkout page. The timer was only shown before 1pm and clarified that the serivce (contract cancellation) will be initiated on the same day if users act before a cut off time. Impact on completed payments was measured.

In this experiment, a product comparison table was added in the middle of a product detail page. The comparison table contained products from the same class or family of products. Clicking on the photo thumbnails also allowed customers to visit the specific detail page. Impact on adds to cart and transactions was measured.

In this experiment, top category navigation (with fly out categories) was removed from the shopping cart of an online store. Impact on sale transactions was measured.

In this experiment, bestselling product colors were highlighted with a bestseller badge on product listing and product detail pages.

In this experiment, the B variation showed the relative discount in price from within the most recent 12 months as a percentage point. A tooltip was also shown which explained this message on hover. 

In this experiment, an additional purchase trigger was added - the ability to start by chosing a gift announcement option. In the control, this same question was asked further in the purchase funnel (after adding to cart). In the variation, this question was surfaced earlier as an alternative way of starting the purchase. Impact on total sales was measured.

In this experiment, five clickable feature elements were surfaced on the homepage. Clicking on any of these 5 elements launched an Instastory-like short 10 second video in full screen. These videos explained the key features of the software being offered. Impact on lead generation was measured.

In this experiment, three testimonials were appended at the bottom of landing pages of a contract cancelation service (paid). These testimonials were also shown throughout the complete signup funnel (4 more steps). Impact on progression (step 2) and final completed purchases were measured.

In this experiment, four TrustPilot reviews were appended at the bottom of landing pages of a contract cancelation service (paid). These reviews were also shown throughout the complete signup funnel (4 more steps). Impact on progression (step 2) and final completed purchases were measured.

In this experiment, a floating sticky bar was added on product pages. The sticky elements only appeared after users scrolled beyond the fixed buy box  area that is relatively high on the page (visible on the control screenshot). The sticky bar contained three elements: a button to configure up to three product choices, a flavor selection pulldown, and the add-to-cart button. 

In the variation, when users clicked on the flavor pulldown three things happened. First, they were anchored back to the top of the buy box. Second, the floating sticky disappeared. And third, the flavors pulldown automatically expanded (overlapping the original primary add-to-cart button).

The control did not have any of the sticky behaviors.

Impact to total sales was measured.

This was a larger leap experiment with numerous changes to the header part of a design program landing page. One of the key changes however was a shift from a single to multiple call to actions for lead generation. In the control, all potential leads would first funnel through a single syllabus download flow. In the variation, users were given three visible choices: download syllabus, webinar signup and/or book a live call with admissions. Impact on overall generated leads was measured, as well as paid enrollments.

In this experiment, an additional and alternative text based call to action link was added on real estate properity listings. Instead of only having "Ligar" ("Call"), "Enviar mensagem" was also appended ("Send Message"). This additional link lead to a lead-gen form. 

In this experiment, the addition of a secondary CTA for lead generation in the nav and the hero were tested. The additional button invited users to watch a gated demo video (approx 12 minutes). Clicking the CTA triggered a pop-up form collecting some basic contact information before the user could access the video page. Total leads were measured with lead form submittions. 

In this experiment, social proof and scarcity messages were shown on a course enrollment landing page. When students were starting to sign up at the beginning of each month (with greater availability), a simpler "X students have already enrolled in the next cohort" message was used.

Later in the month when fewer spots were available, a more scarce message was used with the following copy "ONLY X SPOTS available in the next cohort".

In both cases, the numbers were accurate and dynamically updated.

In this experiment, a question about a customer's reason for purchase was broken out into a separate step and moved earlier in the process. In the control, this question was asked in the final checkout step along with a plan selection (Step 4 of 4). In the variation, this question was shifted as a standalone first step (Step 1 of 5). Impact on completed purchases was measured (for a contract cancellation service in this case).

In this experiment a dynamic canned response was added to a lead form. The contents of the copy reflected a handful of user choices from other form field selections. Impact on leads / form submissions was measured.

In this experiment, a historically higher price point was added as a crossed out anchor. The control only showed the current price. The variation showed the current price with the higher price crossed out. Impact on sales was measured.

In this experiment, up to 5 recently viewed product pages would appear on the left hand navigation. The idea was to make it easier for users get back to what they were looking at in case they were browsing. These recently viewed products were not visibile in the control. Impact on adds-to-cart and completed transactions was measured.