In the world of e-commerce, the ability to measure, analyze, and act on data-driven insights can have major impact on revenue for your business. Understanding how users interact with your website—and where potential pain points may be—provides the online retailer with clear vision to effectively boost traffic, minimize cart abandonment, provide a consistent cross-device shopping experience, and ultimately increase conversions. Sounds easy…but how can these integral insights be derived?
It all begins with solid analytics. Gamma Partners provides a comprehensive set of analytics and marketing services to customize and augment e-commerce site implementations and/or migrations. Analytics, SEO, PPC, data feed management, product campaigns and tag management services are some of the most common optimization tools. At some point in the “discovery” process, the focus turns toward examining core elements of the website and assessing how “tunable” it is. One of the most powerful components utilized at this stage is A/B testing, which helps to determine: 1) is there a problem based on the analytics data, and 2) can the problem be reasonably solved by performing one or more controlled adjustments?
In a recent engagement with a large multi-channel apparel supplier, Gamma was tasked with addressing considerable issues surrounding overall site usability and seemingly stunted conversion rates. Director of Business Insights and Analytics, Adam Pickel, recalls, “What we found was that the end user was required to go through a very non-intuitive sequence of steps in order to recover a cart across multiple devices. We also identified a number of massive gaps in terms of drop-off from the conversion process on mobile…for example, we saw a 40% drop-off from the check-out step to the log-in step alone.”
THE A/B TESTS
1) Countdown Timer – this “visibility” test was developed to support analytics’ revelation that uptake was not high enough whenever the client was running promotions. Promos were not called-out explicitly enough across the website, and promo codes were being under-utilized. “They were really relying on the user to dig a promo code up from an email, or go back to the home page to recover it and then input it at checkout”, Pickel said, “so by adding the countdown timer we saw a large increase in overall usage of the promotion codes as well as a significant lift in conversion.”
2) Checkout Flow Analysis – analytics showed significant user drop-off as well as barriers to conversion:
- 30% of mobile users started the checkout process from the cart page vs. 40% for Desktop
- Mobile users were 93% more likely than desktop users to drop off at the login step on the checkout page
- Mobile users were 57% more likely than desktop users to drop off at the shipping step of the checkout process
- Mobile users were 24% more likely than desktop users to drop off at the payment step compared to desktop users.
- As a result, Mobile conversion rate was only 1.80% compared to 4.64% for Desktop
A/B testing confirmed that mobile conversion rates could be positively impacted simply by moving the “checkout” button higher on the cart page. A series of more granular refinements were also made to the login, shipping, and payment pages, which resulted in an overall reduction of button/text clutter and more efficient use of mobile page “real estate”. Additional tests for desktop eventually led to cart page layout and functionality adjustments that have also shown positive results.
3) “Marie Antoinette” Image Test – this collaborative effort with the site designer proved extremely insightful (if not overly intensive to execute). Based on a JCPenney study that found removing the model’s head in product photos subsequently increased conversion rates, Gamma pushed out a three-way test containing: a) pure product shots, b) standard model shots, and c) the headless model shot. The test yielded surprising results in terms of increased click-through to product pages and boosting overall conversion rates.
It should be noted that, even though the image test was limited to just 50 products, Gamma still had to account for a) the default color option which was selected on the website, b) different device sizes, and c) the variety of pages that any individual product could potentially show up on. While A/B testing can provide deep insights, accounting for multivariate options—especially as it relates to responsively designed sites—can also add a significant amount of time and effort to the overall optimization initiative.
The combination of sound analytics and testing has provided additional insights to allow the client to enhance and extend the shopping experience for users. For example, knowing that 30% of mobile users—who had started the checkout process at some time—will ultimately convert on desktop offers several creative solutions. There are always ways to help users “complete what they started” and streamline the conversion process. Moving forward, we anticipate most websites in general—and e-commerce websites in particular—will be well-served by data-driven systems that facilitate cross-device play and conversions.
The A/B testing process enables Gamma Partners to “audition” optimizations for the client without the burden of upfront resource expenses or heavy implementation effort. Additionally, it provides the client with a very accurate projection of what resulting ROI would be once recommended changes make the transition from virtual to actual. Pickel emphasizes, “When we use analytics-derived insights that tell us that there are big problems in specific areas, we can intuit that any change will have a much bigger impact than if we’re just going with our gut. A/B testing basically closes the loop between analytics and action.”