From the course: Marketing Attribution and Mix Modeling
Unlock the full course today
Join today to access over 23,400 courses taught by industry experts.
A/B testing for statistical significance
From the course: Marketing Attribution and Mix Modeling
A/B testing for statistical significance
- [Instructor] If you make a change and performance goes up, was it your change that moved the needle or was it a coincidence? The only way to know for sure is to run an AB test holding all other variables equal. It's important to do this because correlation does not equal causation just because ice cream sales and shark attacks happened in the summer at the same time, does not mean that one causes another. AB testing is important in order to allow you to control for those variables. So for example, if you randomly assign 50% of the people who visit your website into bucket A, and 50% to bucket B, and bucket B sees a different designed to version A, you know that it's only that change that was the difference. You can AB test lots of different things in marketing. Your ad creative, your ad targeting, or even the channels that you run. You can AB test your website design, your content, the conversion flow, also your email…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Contents
-
-
-
-
-
(Locked)
A/B testing for statistical significance3m 37s
-
(Locked)
Bandit testing: Optimizing for results over accuracy4m 32s
-
(Locked)
Geo and lift testing to prove incrementality3m 2s
-
(Locked)
"How did you hear about us?": Surveys and panel studies2m 8s
-
(Locked)
Working with multiple attribution methods3m 26s
-
(Locked)
-