Simplifying the Cart Page Increases Revenue 11.7%
"It seems that perfection is reached not when there is nothing left to add, but when there is nothing left to take away"
- Antoine de Saint-Exupery
In the eCommerce space, new features are regularly added to sites. Usually this is done because of design trends, market shifts, or in an effort to differentiate yourself from the competition. Over time some sites fall victim to "feature bloat", where there are too many elements on the page competing for attention and detracting from the page's primary purpose.
A client came to us looking for ways that they could improve their order conversion rate. They were new to the concept conversion rate optimization (CRO) and had not previously performed any split testing or rigorous data analysis, so this was an opportunity to come in completely fresh.
One method we use to devise first-run experiments is to dive into a client's Google Analytics account and evaluate their conversion funnel. We noticed a rather large pageview discrepancy between their cart and checkout pages. The client decided this was a ripe opportunity to run their first A/B test. Thus, we were presented with the question: "Why is there a high drop-off rate on the cart page?"
Screenshot of the original cart page:
Using In-Page Analytics, we looked at where shoppers were clicking on the cart page. Although the "Go to Checkout" button was the most heavily clicked item on the page, several other elements on the page also had a high level of heat, particularly the top navigation, and left column content.
Because of this user behavior, we wondered if perhaps the shoppers skewed towards multi-item purchases, so we took a look at average order size. The overwhelming majority of orders were single item purchases. Due to the discrepancy between user navigation behavior and purchasing habits, we hypothesized that: "Shoppers are getting distracted by an overload of features that aren't relevant to the page's primary purpose. They are navigating away from the cart page and then abandoning their cart."
"Simplifying the page layout by removing low relevancy elements and presenting a stronger, more focused call-to-action will decrease drop-off and abandonment." To accomplish this, we made several alterations to the page:
- Removed the top navigation and utilities from the header
- Removed the left column (which contained redundant site search, catalog quick order, weekly special email sign up form, and secondary thematic navigation)
- Removed item recommendations
- Increased the width of the cart item grid
- Increased the size "Continue Shopping" and "Go To Checkout" call-to-action buttons and gave them a more semiotic design
- Separated the call-to-action buttons and positioned them in a way that is more relevant to their purpose (left = "back" for Continue Shopping, right = "forward" for Go To Checkout)
- Duplicated the call-to-action buttons above the cart items grid
Screenshot of the simplified cart page layout:
Performed a 50:50 A/B split test. Variation A was the baseline: 50% of traffic sees the original page. Variation B was the comparison: the other 50% of traffic sees the simplified layout.
The experiment ran for 1 month during a non-holiday, typical traffic period. The results verified the hypothesis. Version B saw a 3.8% increase in order conversion and an impressive 11.7% increase in revenue.
- Version A - 34.84% conversion rate
- Version B - 36.15% conversion rate (+3.8% improvement)
- Version A - $50.56 per visitor
- Version B - $56.46 per visitor (+11.7% improvement)
Another interesting takeaway from this experiment was that Version B caused an engagement focus on the "Continue Shopping" button.
Continue Shopping Button Engagement
- Version A - 11.22% conversion rate
- Version B - 25.48% conversion rate (+127.1% improvement)
For those shoppers interested in doing additional shopping, instead of being presented with a variety of navigation options, they were presented with a focused call-to-action that would direct them to a single location which can be more carefully controlled and merchandised. We surmised that this was a contributing factor to the big revenue jump and plan to do additional testing on this space.
Be careful when deciding what features go on which pages. Each page on your site has a purpose, and that should be the primary focus of the page above all else, lest the page falls victim to feature bloat. Sometimes there can be a lot of contributing factors:
- Marketing says those cross-sells are really important
- Customer Support is getting calls; says we need a warning message
- The HiPPO's pet project
- An inexperienced designer and/or developer
This can be disastrous on a business critical page like the cart page. Allowing these situations to happen and persist can damage your bottom line. The key is to test the impact of these changes. Stay simple. Stay clear. Stay focused.
Nice improvement and a much nicer looking page.
Are you still working with this client?
I have a few ideas that I think could boost conversions further, maybe you could test a version built up from ideas on this thread?
Optimization Co-ordinator at Widerfunnel
Yes, we are actively performing tests for this client. We have a testing plan prepared with about 15 additional, unique tests planned so far, not counting possible iterations that may emerge from the results of a completed experiment. Right now we're heavily focused on surface-level testing; cleaning up and refining already existing functionality. We feel that for this client, these types of experiments will provide the best return for the lowest cost; they have some development/code hurdles to overcome first. That impacts their development costs, so we want to focus on testing changes that won't have a large development overhead cost to implement should an experiment be a success. Once we've exhausted those avenues we will then begin to evaluate new features/functionality.
I'd definitely be interested in hearing what ideas/iterations other people can come up with. How do you come up with experiment ideas?
One way that we get experiment ideas is to do office brainstorming sessions over a lunch meeting. Everyone loves a "free" lunch, but we have an admission fee of 1 test idea. If someone else ends up presenting the same idea first, you'd better have a backup! This setup encourages creativity and participation. On average, we get at least 5 new ideas per session.
I am very scientific with my hypothesis generation. I run a number of analysis techniques over a page to locate issues/problems/hurdles/inconsistency/clarity and then I move on to processing analytics data. Every no and then I love through some A4/A3 printouts onto a wall and scribble all over them with some 'creative' based ideas just to mix things up a bit.