Three-quarters of the time a station has a ratings event that’s dramatically up – or dramatically down – they will tell you they have no idea why.
No idea at all.
And then, like magic, the ratings moderate to a place they were before the spike or the valley.
Real or not, we say, we have to live with this. Because our advertisers know only what the latest numbers tell them.
But is this a proper reading of our ratings? And isn’t it our responsibility to educate our advertisers on how to use these ratings properly?
In statistics – and in ratings – there’s a phenomenon called “regression to the mean.” And what it means in plain English is this: If your ratings are unusually and surprisingly high all of a sudden, you can bet that they will decline next book. And if your ratings are suddenly and inexplicably in the toilet, you’ll be back to normal in short order.
“Regression to the mean” means you should not be dazzled or discouraged by the latest swing up or down. On the contrary, the proper way to view your ratings is over time. So do this:
1. Plot several books worth of ratings in Excel 2. Remove the highest number (no matter how much you hate to) and remove the lowest (this you won’t mind). In statistics, these are called “outliers” – especially if they’re unusually high or low. 3. Draw a straight trend line between the numbers that remain (Excel can do this for you). 4. Is that trend flat? Then your ratings are, on balance, flat. Does it angle up? Then your ratings are growing. Does it angle down? Then your ratings are falling.
We can blame Arbitron for these swings up and down if we want, but the reality is that all audience estimates – all research samples – behave the same way.
Before you celebrate (or panic), make sure your reaction is justified. And make sure to have your advertisers do the same.
Bình luận