Marketers love to cite "statistical significance" even if they don't understand what it means. (Hint: Not a "significant" result that includes statistics.) Significance testing can also be difficult, even for experts who can confuse statistical significance with "substantial significance." Many studies mistakenly cite important gray areas, such as these findings published in peer-reviewed journals. "Escaped significance" (p = 0.07) "Not completely significant" (p = 0.085) "Significantly very close at the 10% level" (p = 0.104) advertisement Continue reading below Even your online test calculator may confuse you: Screenshot of false analysis of statistical significance Statisticians discuss the benefits of statistical significance, but marketers often think of it as tracking confidence levels, unknowingly moving p-hacks and data towards false conclusions ().
If you even know what the significance ghost mannequin effect service level is in the first place). At best, statistical significance indicates the potential for results if the null hypothesis is true (that is, by chance). advertisement Continue reading below I'm not sure if the conclusion about the cause of the result is correct. Nor is it guaranteed to be of practical importance (whether the results are noticeably large). So how does this affect your quest to correctly interpret PPC data? It simply means that you have the burden of planning and knowing your risk tolerance from the beginning. If you accidentally launch it and then blindly insert the numbers into the formula to derive insights, you're less likely to learn something meaningful or "important." Garbage in, garbage out. Two trash cans: Garbage In Garbage
Out Conclusion The paradox of "data-driven decision making" is that data does not make decisions. advertisement Continue reading below You need to make decisions about your data. You may not have an academic background in statistics, but you can hold yourself in the quantitative aspects of digital marketing like you: Develop critical thinking skills. Learn about important calculations. Dive deeper into the data story. Understand sample size constraints. In addition to your own self-confidence (and vocabulary), your insights, reports, and relationships with clients will improve dramatically. Other resources: How to measure the success of your PPC campaign with and without conversion data How Data Science Enhances Performance Marketing and SEM How to Optimize Other Marketing Channels with Paid Search Insights