If your cluster subscribes to the Nielsen Audio PPM service, you’ve no doubt been subjected to the acronym “DDI”. This column will explain what it is and further, why it may be the most useless metric in radio audience measurement, in other words, one that you should ignore.
DDI stands for Designated Delivery Index and is a way to obfuscate the problem of proportionality in Nielsen’s surveys. First, let me explain to you how it works.
In PPM, the data are weighted using 18 discrete age/sex cells (I’m ignoring the other weighting variables because DDI is typically used for age/sex demos, however, it can be applied to other variables as well). Working from youngest to oldest, you have Boys 6-11, Girls 6-11, Boys 12-17, Girls 12-17, Men 18-24, Women 18-24, and on up to Men and Women 65+.
Let’s say that five percent of your market’s population is Men 25-34. In that case, ideally, five percent of the average daily sample should be Men 25-34. It won’t always work out that way and that’s why weighting is used.
For this example, let’s assume that Nielsen was a little short on that cell and that 4.5 percent of the average daily sample for the month was Men 25-34. That means, on average, that those meters would be weighted up a bit (each meter would carry some extra weight to get the total representation to five percent). 4.5 is 90 percent of 5 percent (4.5 divided by 5), so the proportionality for that demo would be a 90. Good so far?
Next, let’s say that the Persons 6+ target sample for the market is 1,000 meters (average daily intab), and because Men 25-34 are five percent of the population, there should be 50 in the sample (.05 times 1000). However, Nielsen overshot (a good thing for the subscribers), and the average daily intab for Persons 6+ was 1,120.
Here’s where DDI comes into play. The DDI is based on the size of the actual sample, not the population percentage. In this case, the math looks like this: We know 4.5 percent of the sample is Men 25-34 and 4.5 percent of 1,120 is 50.4, but let’s round that and call it 50. Since we have 50 men in the average daily sample and we needed 50 based on the sample target, the DDI is 100 (50 divided by 50 times 100 to create an index).
Your Nielsen rep will tell you that the DDI is dead on, but does this have any effect on the results that you’re seeing? Not one bit. The proportionality was still 90 (and 90 is not bad) and that meant the Men 25-34 meters had to weighted up a bit on a typical day.
Why did DDI come to pass? Whatever the PPM P6+ sample target is in your market, Arbitron used to typically overshoot and that was by design. Nielsen generally does as well, but they try to get closer to the target because, well, it costs money to meter a PPM household. If the average daily sample is bigger than the target, then DDI makes the sample look better in key demos than does proportionality. It’s a ruse and if you don’t know any better, it’s misleading.
What should you do about it? Don’t do anything except ignore DDI. What matters is the proportionality of the sample, period. Remove the DDI acronym from your lexicon. If your Nielsen rep sends you data citing DDIs in your market, ask them to stop doing that and include only proportionality, which you can do yourself using the E-Book.
Keep a running track of the monthly results of each weighting variable for your market. As a reminder, the variables are:
Age/sex demos
Geography (if more than one geographic unit, typically a county)
Race and/or ethnicity (Black and Hispanic in most markets, but some have only one)
Language preference among Hispanics (if the Hispanic population is large enough)
Presence of children (children is defined as 0-17)
Working status for persons 18+ (working full-time or not working full-time)
You have enough things to do during your workday. Don’t waste your time on metrics that don’t matter.
Let’s meet again next week.
