Thursday, November 20, 2008

August-September-October = The Most Unpredictable ARB Trend Of The Year

I have been blogging about the country format's up's and down's as well as Arbitron's sample and methodology issues for many years, so no doubt this is not the first time I have noted this seasonal blip on the ratings radar, but each year about the time Winter "Station Information Packages" go into the mail things often get the strangest of the year. Here's why during this particular phase it's wisest to wait at least 30, if not 60, days before either contemplating despair or celebration over the current Arbitrend.

#1. August is often country's weakest month of the year, (it certainly was this past summer at least and then the format bounced back in September), perhaps due to family vacations and kids out of school, distorting response rates, midday and at work listening estimates.

#2. The annual population estimates update from the federal Office Of Management And Budget's yearly census stats start to change all the weighting factors on October 1. This not only affects local governments because U.S. government revenue-sharing is based on these estimated changes in local populations extrapolated back to the 2000 census data, updated each year at this time. It also means you need to reassess your targeting. If your narrow target declined in its proportion of the metro population, you may find that your share of audience could also go down as well.

#3. And, then there's the elections and coming Christmas music on AC in a year when emotions and family celebrations will be more meaningful and important than ever.

So, if you did well in Fall's Phase One Trends, congrats! If you didn't, I'm sorry. But, please hold off on any changes based on your trend until you check out the proportionality of the sample, any changes in populations important to your success and, most importantly, you see it three more times. That's what it takes to call any up or down a "real trend."

No comments: