APPROACH
Performance testing is the process of designing statistical tests on web elements to find the optimal combination to maximize ROI. It’s one of the most over-looked and effective ways to improve website performance. From landing pages and layouts to button colors and call to actions, let us help you make the most of your website traffic.
Once a client’s analytics have been analyzed and we’ve identified web elements and bottlenecks with the most improvement potential, we develop design alternatives. Rather than just picking an idea everyone likes, we say “That might work: let’s test it.”
This scientific approach ensures we find the most effective solution and allows us to continually find ways to improve performance and communication, leading to higher ROI across all marketing channels.
A/B TESTING
We use A/B testing to optimize independent web elements. This could be a simple test such as button text or a headline, or something more complex like completely different landing page designs. Typically, A/B tests are used first on drastic layout changes to determine the most effective overall design, then moving to smaller page elements.
Software such as Google Optimize, Optimizely, and Visual Website Optimizer are great tools to help distribute traffic and analyze data. We’ve used these tools and many more to organize and run performance tests (but we can run them manually and crunch the numbers ourselves too!) and are happy to use whichever tools fits your needs best.
MULTIVARIATE
Multivariate testing is used when a test needs to find the “best combination” of web elements. With multivariate testing, we are able to simultaneously test the relationship of variables such as headlines, images, offers, and prices to find out which combination is most effective, and how much each element affects performance.
This format in particular benefits from testing software. We recommend implementing software to help with the testing and analysis process. Our role is helping to determine effective alternatives, designing effective tests, setting up reliable tracking, and analyzing performance to determine winners and other opportunities.
REPORTING
What good are A/B and multivariate tests without the results? We’ll show you the statistical testing process and report on performance as the results become significant. Using insight gathered from the performance tests, we’ll also suggest follow-up testing that may continue performance improvements.
All major performance testing software includes reporting, but we extend this analysis by helping to interpret results as well as extending insights by looking at other data such as Google Analytics and CRM data.