5 Tips for Ecommerce Split Testing
Last week, CoLab’s content and SEO expert, Jess Spate, co-hosted a webinar with Jirafe’s CEO Amit Shahon on ecommerce customer acquisition.
The topic was split testing, an optimization strategy we highly recommend and use regularly with clients like Valuate.
In a split test, a slightly different version of a website is shown to a random sample of visitors. A classic example of a split test is changing the button color on an ecommerce product page.
Split testing is only one technique used in a CRO (Conversion Rate Optimization) strategy. But it’s an effective one. Google and Amazon rely heavily on split testing to increase their conversions. Amazon famously increased their annual revenue by $300 million (yes, million) by adding the option to check out as a guest. (The $300 Million Button.) Every time you shop online, you’re likely on the receiving end of a split test.
First, the tips:
- Test a multitude of colors.
- Tailor messaging according to what the customer wants.
- Identify points where you can add trust signals.
- Ensure a friction-free navigation process.
- Adjust your testing strategy according to your site traffic.
The importance of color
Tests that seem to be about color are sometimes really about contrast and visibility.
Google famously tested 41 shades of blue to arrive at the blue used on their search result pages. The reason online marketers continue testing colors is because there is no single best color for conversion. There is no one color that is always best for a “buy” button. It depends on context and audience.
Get the message right
Similar to colors, the best choice in messaging depends on the context and audience. People don’t always behave logically, which is why it’s important to test.
Here are the results from recent email split tests we conducted for four different clients. The higher performing message is marked ✔.
|Shop now||Own it! ✔|
|Contact us||Talk to an engineer ✔|
|Subscribe||Subscribe to hear more ✔|
|Get a free demo ✔||Schedule a free demo|
When considering messages to test, think about what the customer actually wants. If they see what they want, they’re much more inclined to go forward. Ask, “What do I want to do other than submit this form?” Do they want to join, to buy, or to get involved? Reflect their unique desire in the messaging.
Trust is difficult to establish online. As Jess points out, people trust other people more than they trust text.
A recent project involved adding trust signals to a client’s landing page. Including a single testimonial and a few real user photos lifted signup rates from 0.81% to 4.81%.
Adding social media buttons — with share counts — is one way to potentially build trust. Be sure to have a strategy for increasing your share count before putting the numbers on display. A low number of shares can breed distrust, resulting in an adverse effect on conversions, according to a recent case study from Visual Website Optimizer.
Identify friction points
In ecommerce, friction points are problems in the checkout process that keep a potential customer from completing a purchase. The result is your dropoff rate. Dropoff rates typically range from 30% to 70% of visitors.
Like color and messaging, testing friction points with small variations can have a measurable impact. We’ve seen a project with an 11% conversion lift just by adding visual emphasis to the site-wide “checkout” button.
Adjust according to traffic
For clients with large traffic — over 20,000 unique visits per month — we often run two or three split tests in parallel. In this case, you can be constantly testing.
If your site is smaller, your testing opportunities are limited. If you have 1,000 visits per month, it might take up to 2 months for a single test to come to maturity. You will need to figure out what the best practices are, implement those, and test more selectively.
Watch the full webinar: Ecommerce Customer Acquisition Decoded