Test for statistical significance

[StevenAdema]

from https://www.reddit.com/user/forgottendinosaur/:
What you can do is run a difference-in-difference or a regression discontinuity analysis. You can also play with the bandwidth (how much time you have on either side of the series introduction). A plain t-test here makes a strong assumption of equal slopes on either side of the cutoff. There might also be a delay in the effect (assuming there is one) - a rise in popularity of the opening after a certain number of days passed, episodes watched, or popularity of the show.