What data was collected? P2+ viewership for every quarter hour of NXT and AEW Dynamite television programs, going back to October 30 in the case of AEW and November 6 in the case of NXT. P18-49 viewership was also collected for most but not all of the weeks in this timeline. Therefore we’ll rely most on the P2+ metric. This data was collected by Cory (@voidtoaster1992) from issues of the Wrestling Observer Newsletter.
What do P2+ and P18-49 mean? These metrics represent Nielsen’s estimate of the number of viewers watching the program during a given quarter-hour. P2+ means all people viewing over the age of 2. P18-49 means all viewers between ages of 18 and 49. The latter demographic is often considered more valuable to advertisers.
So what did you do with the data? I tried to use math to make a fair assessment about which quarter-hours exceeded expected performance and which personalities’ appearances coincided with viewership expectations being exceeded, or the opposite.
How did you do that? There are factors that may skew a quarter-hour’s viewership. For example, the opening quarter of both programs is often one of the most viewed, as is the segment at the start of the second hour. Likewise, the 7th segment is often among the least viewed of either program.
How did you adjust for those inherent advantages and disadvantages? For every data point (every quarter-hour), I determined an “Expected Viewership” based on what the average viewership of the program would have been without the given quarter-hour and what the average viewership of the given segment number would have been over the entire timeline. Overall, I used a 3-step process to determined “Expected Viewership” for all data points:
1. Find the degree to which each data point deviates from the mean: For all data points, find the % to which a given data point was greater or less than the average of all quarter-hours for the given week’s program. I also found the +/-% for the average for the entire week’s program compared to the average of all programs in the timeline.
2. For each data point, find what the program’s average +/-% versus the overall +/-% average would be if that data point was factored out.
3. For each data point, use the corresponding +/-% result found in the previous operation and multiply it by the segment average, then add it to the segment average.
After you determined “Expected Viewership”, then what? Then I subtracted the Expected Viewership from the Actual Viewership, for each quarter-hour. I calculated this both in terms of number of viewers and as a percentage. The result is an attempt to answer whether a given quarter-hour over-performed or under-performed versus expectations.
Why didn’t you just measure viewership lost or gained from one segment to the next? Such a study would be skewed by the quarter-hour bias discussed above, and would not provide the fairest assessment of whether a quarter-hour’s viewership met, exceeded or fell below expectations. Determining an expectation that tries to factor in quarter-hour bias seems like the best alternative with the data available.
There are other factors I would have liked to consider, like commercial breaks, but doing so would require minute-by-minute viewership data that is not available.
What other weaknesses does this study have? We used descriptions from the Wrestling Observer Newsletter to label what happened during each quarter-hour. These descriptions are not exhaustive. Although they name the key people involved, they do not name every person who appeared onscreen during the quarter-hour, nor do they give credit for how long they appeared within the 15-minute period. Again, a minute-by-minute analysis of each program would allow for a more comprehensive analysis.
You can look over all the math for yourself here: https://docs.google.com/spreadsheets/d/1FAowangTzZ5AHs72tsH530An169of2pw_GuqDdw5B6w/edit?usp=sharing
Since the P2+ data is more complete, we will focus on that metric and less on P18-49 for which we have incomplete data.
The tables below show the top and bottom 16 quarter-hours that over- and under-performed expectations, respectively, for each NXT and AEW Dynamite.
See the spreadsheet in tabs “AEW quarters ranked” and “NXT quarters ranked” for full view.
Of the names in our labels, which are associated with the most positive differences (in Actual Viewership vs. Expected Viewership) on average?
The results are below. Again, this is based on quarter-hour analysis only and doesn’t account for how long each person was onscreen for during a given 15-minute segment, nor does it account for who they appeared along with. There are also problems with collective names. For example, when the Inner Circle or Undisputed Era appears in a label, the individual members names often do not; the stable is credited, the individuals often are not.
The below is a screenshot from the spreadsheet used for this study. Short names or last names were used to search because sometimes in labels a person is referenced only using that shorter version of their name.
Instances are filtered to a minimum of 5 label appearances. This rules out some who have high percentages but who only appeared a few times, as those associated with positive effects consistently over the timeline.
It’s remarkable that women dominate the top-ranks for NXT by this analysis. Likewise, the individual male main eventers for NXT are absent from the top end of the list, although Adam Cole’s Undisputed Era makes the list.
AEW is topped by Chris Jericho’s Inner Circle stable, followed by MJF and Cody who are involved in a feud.