Set up trust badges test for ecommerce
Set up trust badges test for ecommerce
1. Select a trust badge or set of trust badges that you would like to test.
Trust badges provide a way for ecommerce sites to address the common concerns of online shoppers and make them feel more comfortable completing a purchase. To decide which type of trust badge to test, put yourself in your customers’ shoes and consider their concerns at each stage of the journey on your site: Security trust badges show visitors that a site is secure. Some indicate that the site is free of viruses and malware, some indicate that any data submitted on the site is secure. Examples of security trust badges include the TrustedSite Certified Secure trustmark and the McAfee SECURE trustmark. Payment trust badges are intended to make a shopper feel safer about inputting their credit card information on a site. Some examples include PayPal, Google Pay, and Visa Secure. Endorsement trust badges show that an outside organization has verified the site is legitimate. The BBB Accreditation Seal is an example of an endorsement badge. Some providers like TrustedSite offer a set of badges that address a variety of shopper concerns. Search for case studies of the trust badges you’re interested in to see if they have a track record of performing well on sites like yours. Consider your budget, your audience, and business goals when making your final decision on which badges to test. Start a free trial with your trust badge of choice. If a free trial is not available, start with a monthly subscription instead of annual, to keep costs low until you have the results of your A/B test.
2. Set a goal for what you would like to learn or achieve by testing trust badges on your ecommerce site.
For example, whether: A trust badge increases conversions on your site. Multiple trust badges increase conversions on your site. A new trust badge works better than a current one. Trust badges increase revenue for your business. Trust badges convert more new visitors that are less familiar with your business. Trust badges keep visitors on site for a longer session. Consider your budget and the cost of the trust badge service. What ROI would you want to see? Use these considerations to create your goal. For example, your goal might be the following: By testing trust badges, we hope to see a 5% increase in conversion rate. The additional revenue brought in would help us achieve an ROI 10x the cost of the trust badge service.
3. Review your analytics funnel visualization and navigation summary reports, to figure out where users are dropping off or exiting the site.
Trust badges can reduce friction. To figure out where to test trust badges, find current high friction areas. Open your website analytics reports and make a list of pages where users are frequently dropping off or exiting the site. Analyze the root cause of that friction. Is it caused by trust issues? Can you counter it with a trust badge? If you find that you need different trust badges than you’d intended to test, amend your plan now. For example, say users are dropping off at the stage of the checkout where they need to enter their credit card information. Would a trust badge that shows credit card data is secure, encourage them to input their information and complete the purchase?
4. Pick high friction pages for testing the trust badges.
Decide the exact location on the page that each badge will be displayed, and whether you will test multiple locations on a particular page. Here’s a sample plan using TrustedSite’s suite of trustmarks:
5. Create an experiment plan based on your analysis and learnings.
Your experiment plan should include: The problem you’re trying to solve like new or unfamiliar visitors not converting, or repairing existing trust issues. Proposed solution. Hypothesis. Placements. Experiment design, for example, what placements are in which variation and why. Screenshots of the control. Screenshots of mockups of the challenger variations. Metrics and goals the experiment will be measured against. Documentation of which user groups, device types, and traffic that will be included in the experiment. Estimated duration of test. Most sites will need to test for at least 2-6 weeks, depending on traffic volume and the experiment design. Prepare this information in a slide deck or PDF format.
6. Code out the experiment with a web developer or an agency like CXL.
For more complicated websites, JavaScript-coded experiments often work more consistently and have fewer issues in the A/B testing tool, compared to using the tool’s visual editor. Design the code for the experiment to be flexible in the order elements might load on the page, and the time loading might take. The A/B testing tool may load the experiment code more or less quickly for different users. Check that the trust badge placements do not conflict with other dynamic elements on the page, and that they look appropriate on mobile devices. A high quality experiment may involve placements that transition between locations or change size responsively, depending on the properties of the window.
7. QA the test to make sure everything is working and then launch the test.
Check placements on your machine and run cross-browser testing. Always submit a test transaction through each variation before launching the experiment, to ensure that the experience is functioning as intended.
8. Check results weekly, on the same day of the week that the test launched.
Check especially that the addition of the trust badges hasn’t had a significant negative impact on sales. Checking results on the same day of the week will help to negate any fluctuations in traffic or conversions that occur over a week. Don’t be too quick to end the experiment if the results aren’t what you expected. You need to give the test time to have enough traffic pass through the site.
9. End the experiment when you’ve reached a significant result.
A significant result could be achieved in 2-6 weeks depending on your traffic volume. A more advanced way of ending an experiment is to close the experiment to new visitors and let the experiment continue to run until most customers have completed their buying cycle. This can take several days or weeks, depending on your audience.