Tuesday, August 9, 2011
Before we begin...
As we head into lead-ins, the last of the three big steps of this project, I've finally started tinkering with a rough draft of a "final" True Strength rating. (Yay!) I'm not sure it's a particularly good practice to look at the results and then go back and make changes to get the results I want, but it is pretty clear that many of the True Strengths toward the end of the season are way too high. I'm assuming that can be attributed to a problem in the viewing level adjustments for Nielsen's new methodology. So I've gone back and switched that adjustment to something based more on actual viewing changes (though with a smaller sample size than I'd like) than on ratings changes. I think my average declines were inflated by the big drops from Saturday shows like Cops and America's Most Wanted. A more detailed explanation is now back on the original Methodology Adjustment post.
As with competition (and viewing levels, too, though I didn't really formalize it at the beginning of that one), there are two baselines that we have to establish before we can really turn this issue into a formula:
1) How much does a change in lead-in change a show's rating?
2) What is an average "expected" lead-in?
There are a couple other really big things we have to untangle on the lead-in front:
1) Size of show. One thing that I don't think gets mentioned enough when people talk about lead-in retention is that the impact of the lead-in depends in part on how big the lead-out show is. The more time the show lasts, the more people who are hanging on due to the lead-in flee. A half-hour show's average rating is gonna be more boosted by a huge lead-in than an hour show, because there's half as much time for the show to drop toward its "true strength."
2) The local programming lead-in. We can't just remove the lead-in effect from shows that start at the beginning of primetime (usually 8/7c), even though we don't know what the national rating is of all the local 7:30 programs combined. So we have to come up with an approximation of the "true strength" of the local programming lead-in by comparing shows that air both at 8:00 and also later in the evening.
The first baseline and the first "other issue" sort of have to be explored in tandem, because different length shows change by different amounts, so that will be our next post.