Client : Anonymized for Privacy (Gift Products Online Store)
Split-Testing Platform : Optimizely
Site Type : E-commerce (Multi-Product)
Test Approach : Variable Cluster A/B Split Test (Multi-page)
Objectives : Increase in sales conversions and revenue per unique visitor
Commentary and Critique on Original Site (Control)
This eCommerce stores sells a large variety of products suitable to be purchased as gifts. These products are the newest gifts on the market, incorporating unique, engaging and highly desirable gifts for anyone and everyone.
Based on previous test results as well as feedback from customer support, there is indication that customers had trouble finding suitable products as gifts and they need help. We also gathered that few people scrolled down to the bottom of the homepage.
For such an eCommerce store it is important that potential customers are able to easily identify and purchase what they need. If they are having a hard time finding what they need, they will just leave the store for another.
We identified some possible areas of improvement for conversion testing, based on 4 main areas: initial experience, value proposition, friction and anxiety. Below we highlight some of the major issues we looked at.
As mentioned earlier, in the control, customers have trouble finding suitable gift products. We hypothesize that if the site improved on this aspect, their conversion rates would see an increase.
In the treatment page, we tested our hypothesis and changed the top navigation and side navigation to improve the focus. We designed inspirational paths to help potential customers browse the shop more effectively. We based our design on surveys, customer support as well as information from previous tests.
To reinforce their core value proposition of variety, we added a headline in the treatment that conveys the message above the inspirational path banner. We used reverse text as the attention maps show that this design captures more attention.
The marketing copy below the headline serves to draw the visitor into the shop experience and strengthen the value proposition.
As mentioned earlier, we designed inspirational paths to help visitors browse effectively as there are many products on the site. We categorized products based on by the person you are gifting, by occasion, by price and by personality. This can help reduce friction as it allows them to find exactly what they need and make a purchase.
Newsletter Feature and Product Display
We reduced friction on the treatment page by removing the newsletter subscription function as it might distract potential customers from the primary goal - to browse and buy.
We reduced the number of products displayed to 3 so as to reduce friction. Having too many options can overwhelm the visitors and complicate and distract the potential customers from the main goal of purchasing. When information is presented to them in a manner that is difficult to digest, it reduces the likelihood of them taking action.
To reduce anxiety, we added a very obvious “Shipping and Returns” link on the top navigation bar. Shipping and returns is important to the potential customers and by letting them know clearly how much their shipment cost and introducing the 365-day 100% money back guarantee policy would counter their concerns.
The “checkout” button in the control was changed to “view cart” in the treatment. In actual fact, the button functionally leads to the cart and not directly to checkout. It intuitively makes sense to label the button clearly based on what it actually does. By doing so, anxiety of potential customers is reduced if they want to view their cart before proceeding to checkout immediately.
To further reduce anxiety, we added credibility icons by showing where the eCommerce store had been featured on. This would provide social proof and a peace of mind for potential customers viewing the page.
Running the Test
We used Optimizely as the A/B split test platform for this website, and conducted the A/B split testing on several template elements on the site clustered as a single variable.
The combined changes made to the site resulted in an increase in both conversion rate and revenue per visitor.
The conversion rate for the treatment improved by 12.8% in comparison to the control at 95% level of confidence.
The revenue per visitor increased by 16.2% in comparison to the control at 95% level of confidence.