Monday, July 21, 2008

Ratings and gas prices: Two demonstrations of the same phenomenon

In my continuing attempt to clarify why weekly fluctuations in US ratings are meaningless, I tried to lay out this example at SON:

Weekly fluctuations in ratings are random...or at least due to forces unrelated to the shows.

A good parallel right now is gas prices. They have moved in a steadily upward direction (the opposite of soaps), and their current (national, US) mean is about $4.00 (give or take).

But then a wide variety of forces (not all random, but local/specific/one-time) cause those prices to fluctuate. Sometimes, they are $3.81, sometimes they are $4.25. It really depends on lots of factors.

Now, let's say you need a good idea of gas prices for PLANNING PURPOSES. For example, how much should you budget for gas? You can't do anything with that fluctuation information, because that just introduces uncertainty.

So, what you try to do is figure out the current value of the moving average (say $4) and plan for that. If the price fluctuates between $3.81 and $4.25, you can cope, because your current-average is a decent estimate. Now, if tomorrow the price goes to $4.50, your estimate is way bad. Even then, if it stays at $4.50 for one day, and then goes back to $4.18...well, it's a shock but not a meaningful one. But, if it goes to $4.50 and then starts to stay between $4.40 and $ know your moving average has shifted...and it is time to revise your budget projections.

So, in this case, there are two values that are informative: The average (the usual price of gas) and the maximum (the most you have to be prepared to spend).

How does this translate to soaps? Well, first, the week-to-week fluctuations are "random"...or at least due to local and basically non-systematic factors that we can't use to explain future ratings. Second, if you're a network or an advertiser, what you want to know is (a) the new value of the moving average, and (cool.gif the new value of the MINIMUM...the smallest number of viewers you can count on to see an ad.

So, if I were an advertiser making plans, I would first ask "What have the ratings been for this quarter?" and "What have the maxima and minima been?" Those would give me a sense of how many eyeballs I'm delivering. As an advertiser, I would NEVER care about the weekly or daily ratings...they are too unreliable. (Except...ratings seem to be higher on Tuesday and Wednesday--week after week--so I might pay attention to that in terms of "best days to advertise").

That is why I pay particular attention to SON’s Toups' flags "new low" and especially "ties low". Because when a show ties its' low for several weeks, that tells me the average...and certainly the "envelope of fluctuation" has trended down again.

In the last few weeks, two examples are germane: Y&R seems to be holding at that 3.5 average. This week it might have been up .1, last week down .1...but basically the new HH for that show is 3.5 (and, per the Jack Peyton axiom, it will NEVER come back).

The other germane example is GL. Now, GL is interesting because (a) it hit a new low; (b) it maintained that low for two weeks running, and (c) it kept that low EVEN WHEN every other show bounced up.

So, now my eye is on GL...unless they show fluctuations back up in the next few weeks, I'll start to assume that GL has hit its new lower plateau on the moving average.

I'm sorry, folks, when I go on like this. It is an obsession and I seem unable to control my fingers.

No comments: