Spurios relationship- two or more variables are related but no causation Binary bias- tendency to want things to be black and white
What data should we be rethinking?
Example Q: Why did CTR go down? A: Because they were getting very few clicks before. Clicks had gone up 5,000%.
Website traffic falling can be a great thing. Example of client who ranked for something that wasn’t relevant. When they lost the ranking, the accuracy of their analytics went up.
What does clickthrough rate really tell you? What does average position really tell you? What does the number of ranking keywords really tell you?
These help you visualize some data about your keywords:
Cultivating a learning culture
Ask questions:
- What leads you to that assumption?
- What might happen if it’s wrong?
- What are the uncertainties in the analysis?
- What are the disadvantages?
- How do you know?