Not all data is equal, but in decision making, you often hear all data being given equal weight. I am going to show you the 3 factors you should use when you evaluate how good a type of data is. We will also talk about how bias can be part of any data set and how it creeps into different types of data.
3 Factors
• Correlation vs. Causation
• Our own ideas vs. What the actual visitors think
• Bias – Type of data can have bias built into it
Types of Data Sets
• UX & Heuristics – Design best practices, your own evaluation of the visitor experience
• User Research / Usability Studies – People or could be end users, but just a few of them, also easily biased
• Analytics & Heatmaps – actual end-user behavior, Correlative data, no bias in the data just in how it is interpreted
• Customer Feedback / Surveys – Actual voice of the end-user, lots of bias because of vocal minorities
• A/B Testing Data – Actual end-users tested, Causal data, no bias in the test