Identify what you are trying to improve and the full scope of what you are testing. Set goals. List challenges. Set a testing schedule that doesn’t interfere with other tests.
After deciding on potential improvements and tests, focus on one improvement at a time. Test colors, fonts, taglines and headings, ordering of content, a single image, icons, etc. When you focus on one specific element, your results become more refined, making a larger impact.
When experimenting, make sure to test one element at a time. Whether you are testing a single page or a user flow, it’s important to test one specific element or section at a time. Testing more than one variable at a time can lead to confusing results about what affected your visitors. Testing one thing at a time helps focus on your results, getting to your goal faster.
One of the best ways to improve your testing is to use javascript triggers. Optimizely, and many other A/B testing platforms, should have the ability to create better event tracking using triggers that are called using javascript. This enhances your testing ability by narrowing down to specific events on your website or app. Typical events such as clicking on the cart button, reviewing the abandonment rate for forms, quantifying popular products, and triggering events for more dynamic properties like user roles, purchase amounts, and more! Be on the lookout for our future article on javascript triggers!
Optimizely, specifically, has a feature called extensions, that allow developers to build out testing capabilities beforehand, building an extensive interface for not so tech savvy users. Extensions are best utilized for creating duplicate tests, such as A/B testing seasonal headlines, or holiday background images. You can create extensions that are very easy for anyone to change, with all of the metrics set up and ready to go! We plan on writing an article specifically about this as well, stay tuned!
Experiments should have about 3 - 4 variations, including the original. It is important to keep the variation decisions limiting and unique. Keeping variations completely unique allows for better results. You can always create a new experiment narrowing down different variations in a different experiment. For example, let’s say you want to test a headline. You have a few options. You could test the location on the web page, the color, the font, or the size. You certainly would not want to test them all at once. It’s better to break the experiment up in stages. Start with an experiment that captures the best position, and then in the future, create experiments based on the style.
There are tons of resources out there! If you are looking for help improving your digital products through testing, using Optimizely or another testing suite, feel free to contact us! We are here to help you improve your digital products.