Whether you’re selling a single product or have multiple SKUs available in your e-commerce platform; if your conversion rates aren’t 100%, then there’s always room for optimization. A platform called Optimizely enables businesses both small and large to deliver the best customer experience online, across all user devices for the web and mobile. Most use the platform to increase their site engagement activity and conversions, but its visual editor could also be used to make instant changes to a website page without the help of an IT professional.
With its user friendly interface, anyone can easily setup new experiments and find the best performing version of their website property. At any time during a live experiment you can login to see how visitors are interacting in real-time with your different variations. For a deeper integration and greater insights, Optimizely can pass data into Google Analytics and other Analytics platforms.
There are 3 types of experiments you can run:
A/B Test: The most popular and least complex, this test-type helps split test 2 to 4 page versions against each other. You can either make changes to the current landing page or even redirect traffic to an entirely different page URL for an apple to oranges comparison. Once enough visitors have been funneled to these pages, the results of each version can be compared. A/B testing in Optimizely can also be used to test engagements with a single element such as but not limited to textual changes on a call-to-action button, button color, stock images used or a different slogan.
Multivariate Test: Similar to an A/B test but includes more variables in order to test changes to multiple sections of a page. One use case is testing out changes to a pages header copy, sales copy and length of a lead generation form. While it generally takes more time to declare a winner with multiple combinations, it can be super helpful for testing a redesign of a page.
Multipage Test: Test changes to multiple pages or funnel. You can create multiple designs for a set of pages in the customer funnel. It’s important to keep the experience for the user to see only one design style throughout all styles to maintain consistency. Similar to simplicity of A/B testing, it’s best to only test a few variables at a time in the multi-page experiment otherwise it will take a long time to declare a winner with numerous variations. One example is testing the placement of a checkout button placed all throughout category and individual product pages. It’s also recommended to keep the number of variations even for every page variation that is part of the experiment; otherwise the data collected would be difficult to interpret.
While in the Visual Editor when setting up variations for any experiment, users can edit or insert elements (text, HTML, styles and images) as well as move, resize and remove items. Per the screenshot below, experiments can target specific visitors and track engagements i.e. clicks or pageviews.
While an experiment is live, Optimizely tracks every unique site visitor and displays a different version to these bucketed users dependent on the experiment type and its settings. Not only do you have the option of showing your experiments to a controlled group of visitors, you can allocate a desired percentage of traffic to a certain variation. If you already have a good performing variation getting a lot of traffic yet not comfortable losing out on more than half of your traffic to a newer variation, you could test it out in smaller scale to gain a sampling of insight before giving too much weight of the valuable impressions.Once Optimizely helps declare a winning variation, you can immediately allocate 100% of the visitor traffic to the particular variation for as long as it’s needed until the technical team can deploy those changes officially.
Those are just high level insights into using the Optimizely platform. Happy Optimizing!