How to Run a Website Experiment with Intellimize

When it comes to website experimentation, it’s best to think of it as a cycle. Why? Consumers' needs, wants, and preferences are constantly evolving which means yesterday’s winning variation might not work for today’s market. As marketers, it’s our job to keep up and consistently deliver relevant experiences that compel shoppers to convert. To do so, we must approach website experimentation as an ongoing, iterative process.

Keep reading to learn how to run a website experiment with Intellimize in four simple steps.

‍

Research

‍The first step to any well-executed experiment is deciding where and what to test. As a starting point, take a look at your highest volume pages or pages that are seeing the most drop-off (which you can identify in Google Analytics or a similar tool). The best elements to experiment with are generally the most straightforward, including homepage headlines, hero images, and navigation structure. For example, Impact Dog Crates first experiment was changing their homepage hero from an image of a German Sheppard to a Golden Retriever. In two weeks, their conversion rate doubled.  

After you’ve experimented with the low-hanging fruit, take the time to get to know your visitors' needs and habits by reviewing heatmaps and reading customer reviews. These will help you identify which website elements you should experiment with.

Additionally, research what other companies are showcasing on their websites. Although you should keep in mind that what works for your competitors may not work for you, this competitive intel can help inspire your next round of website experiments.

We’ve streamlined this part of the research process with one-click playbooks which you can access directly in Intellimize. Just choose from our library of playbooks, and we’ll help you launch an experiment that’s worked for brands like yours. Here are just a few of the playbooks you’ll find in the app:

• Improve click-through with non-committal CTAs
• Infuse social proof into your headlines
• Simplify your pitch with plain language

Plus, rather than calling an experiment quits after finding a winning variation, use the insights you gain as a jumping-off point for your next experiment. Although this may sound intimidating and time-consuming, with Intellimize it’s easy.

Build

‍Once you’ve done your research and decided what to test, it’s time to build your experiment. Using Intellimize you can choose between three types of experiences: Test, Personalize, and AI Optimize. To find a single winner of your experiment select “Test.” To pre-determine which variations a specific audience sees, choose “Personalize.” If you’d like to optimize for conversions, pick “AI Optimize,” which uses our proprietary AI trained on over 10 billion web pages to pick which version of your site works best for any given visitor.

No matter which type of experience you run, you’ll be able to build it easily using our codeless Visual Editor. Using ​​a simple "WYSIWYG" interface, you can add, delete, refine, and rearrange elements. You can also refresh your headlines, calls to action (CTA), and any other page copy you’d like. Visual Editor comes equipped with our AI copy suggest feature which can help you beat writer’s block as you create new copy variations that reflect your specific brand voice.
‍

‍Launch

‍Once you’re happy with the variation(s) you’ve created in Visual Editor it’s time to launch your experiment! Just click the “launch all” button and your experiment will start running on your website.

Yup, it’s that easy!

‍Analyze

‍The type of experiment you’ve chosen to run, your site traffic, and several other factors will impact how long it takes to see experiment results you can act on. With our Test experiment type, you should wait until your results are statistically significant before declaring a true winner and baking it into your site.

If you’ve chosen to use AI Optimize, there’s no need to wait for “stat sig.” Because optimization happens in real time, you’ll be able to identify positive and negative trends much faster than with traditional A/B testing.

No matter what type of experiment you choose to run, be sure to review the results to decide whether you should keep the variation running, pause it, or bake the winning change into your site. This type of analysis, which is made easy with Audience Insights, will not only help you identify what’s working on your site but will also inspire your next experiment.We hope this four-step process serves as a foundation as you run your experiments with Intellimize. Happy testing!

‍