All Latest 498 A/B Tests

This experiment was triggered by a small segment of users who completed the first step of checkout funnel (submitted email, name, phone number), but dropped from checkout and returned to cars catalogue 7 or more minutes later.

In the control, users didn't see any cart icon nor function to resume their checkout flow.

In the variation however, users saw a filled shopping cart icon with resume functionality. Clicking on the icon would guide and redirect users to their latest abandoned stage of their checkout flow.

In this experiment, people who saw an offer (in an email or popup) would see a more visible site wide reinforcement of their earned discount being active. In the control, the discount was only shown during checkout. In the variation, it was shown throughout the web site on the homepage and product detail pages.

In this experiment, popular products were shown at the top of the homepage. Impact to sale was measured.

In this experiment, users that did not complete a purchase and came back to the homepage were shown two different experiences. In the control, users would see the homepage (essentially having to restart the purchase flow). Whereas, in the variation, users would be shown a "Welcome Back" summary view with the completed steps shown as completed and a quick resume button to the next incomplete step. Impact on completed sales was measured.

In this experiment, instead of showing 12 products per screen (with pagination), the variation showed all 42 products on a single screen. Impact on adds to cart and completed orders was measured.

In this experiment, recently purchased products were appended at the top of the homepage. The test ran for loggedin users only. Impact on add-to-cart, sales and revenue was measured.

In this experiment, 9 Twitter card style testimonials were appended onto the homepage of Rollbar. These were image / screenshots recreations without links to the actual tweets.

In this experiment, very short form testimonials (with a reference to over 3 million customers) were replaced with 3 more elaborate ones. These elaborate or benefit testimonials contained: highlighted statements, star reviews, emphasized location, tag summaries and photos of the purchased product. The control also contained a 3 testimonial carousel interaction.

This test appeared at the bottom of a longer homepage with additional product listings above.

In this experiment, a video testimonial was added mid page onto a signup / lead form page.

In this experiment, three video testimonials were added mid page onto a signup / lead form page.

In this homepage experiment a number of changes were introduced - with perhaps the most prominent one being the replacement of a video component with customer review badges. Additional copy changes included reinforcement of the latest version number (v11) throughout the page, as well as a dedicated (v11) section in the middle of the page. Impact on demo signups was measured.

In this experiment, copy was added which showed the number of experts (in a given category) a person may gain access to after filling out a lead form. The context of this is a lead-gen landing page where people are seeking experts for legal purposes. Impact on initial progression (of a multi step form) and completed leads was measured.

In this experiment, supporting university logos were appended near the primary call to action for additional credibility. These logos were placed around the middle of long landing pages on Expert Institute's web site (where experts for legal advice are searched). Impact on total leads 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, 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, 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.

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, 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 single step popup modal was tested against a 4 step sign up funnel. Users entered the experiment on the homepage and the behavior of the two buttons, book a demo and get a quote, was adjusted. Impact on leads was measured.