Online environments present data miners with a unique mix of challenges and advantages for data collection and analysis. Here’s the bad news: Web data formats can be difficult to import and manipulate in data-mining applications.
Systems that serve web pages are often poorly integrated with sales tracking systems, making it hard to identify connections between the visitor’s experience and the resulting actions. Web designers and webmasters don’t always have testing in mind when they develop designs or select web technology. Even if no big technical challenges exist, people can be reluctant to open web platforms for experimentation.
There’s good news, too. Special tools are available that greatly simplify the process of serving web pages properly for testing, and provide analysis capabilities as well. So, no matter what tools you generally use for data mining, consider using a specialty tool for web page testing. (You can find information about these by searching for terms such as A/B testing tool, multivariate testing tool, and split testing tool.)
Just one thing: To test online, you first must have cooperation with the people responsible for your organization’s website. If you don’t have that now, it’s time to open that discussion.
Tools designed for data mining are not necessarily ideal for A/B testing of email or web pages, and you may encounter challenges integrating the two. But you may be able to sidestep those issues altogether. Many email service providers offer built-in A/B testing capability. If you use any of the major email services, you probably have this available right now.
For web design testing tools, use your favorite search engine and the keywords A/B testing or multivariate testing.