One of the biggest challenges when analyzing data is recognizing the difference between correlation and causation. According to the Merriem-Webster Dictionary, the differences occur when you answer the question “Why?”
- correlation is “the relationship between things that happen or change together”
- causation is “the act or process of causing something to happen or exist”
To say it like a statistician, post hoc ergo propter hoc (after it and therefore because of it).
The problem is, when a statistician says this, they are being sarcastic and pointing out a fallacy. Tyler Vigen has a popular website of Spurious Correlations full of charts showing the problems with connecting data that correlates for no apparent reason. One example shows the coincidental correlation between per capita consumption of mozzarella cheese and civil engineering doctorates awarded in the US.
Why Should We Always Ask Why?
It is very easy to look at correlations in data analysis and assume there is causation. But closer analysis may show a factor that changes the way you develop your strategy. For instance, when Google changes an algorithm it has an effect on your results, but you need to do more thoughtful research and analysis to figure out which factors are actually causing the dynamic.
The ability to filter out the spurious correlations and isolate the real cause takes experience and a human analysis. This is vital when dealing with the huge waves of data in PPC management. There are many correlations that show up, but deciding which ones are connected, how they connect, and why they happen takes an expert. Without expert analysis, you might end up wondering if eating mozzarella cheese will help your civil engineering career.
For more help on analysis and PPC management, visit reciprocalconsulting.com/pay-per-click.php.