• by Michael Kubin , December 5, 2001

December is National Forecast Month. This year is almost over, say goodbye to the past twelve months. Next year hasn’t begun yet, so make some guesses as to what’s going to happen.

Think about it: Most economic forecasts turn out reasonably right as long as they’re within a reasonable range. That’s because in a year’s time most things don’t change all that much. For example: a year ago, how many economists predicted the Fed would cut interest rates ten times? None that I know of, because ten interest rate cuts in one year is way beyond anything we’ve ever seen before. But how many forecast that the unemployment rate would rise by about half a point? Quite a few, because it’s well within the range of probable outcomes.

So we can all become forecasters with reasonably good track records just by forecasting stuff that’s likely to happen. It’ll be cold in winter, hot in summer, that sort of thing. Besides, no one ever really checks on forecasters’ accuracy. And on the odd occasion that a forecaster is ever held to account for a klinker, there’s always a mountain of plausible excuses for the error: Earthquakes. Untimely deaths. Vinny Testaverde broke his leg. (September 11 will doubtless be the mother of all excuses for last year’s wayward guesses.)

This week’s media spending forecasts were of particular interest; Bob Coen of McCann-Erickson and John Perriss of Zenith Media are two that come to mind. In forecasts that received prominent display in major media, the two generally agreed that media spending will slowly begin to recover sometime around the middle of 2002.

So I wonder: how did they arrive at that conclusion? Reputable pundits that they are, I assume they did some research and asked around. Mostly questions like, “What are you going to spend next year?” That would be a reasonable question to ask, if your job is to forecast media spending.

And just who are they polling? I mean, would a BBDO client give away any spending information to a guy calling from McCann? (Imagine this phone conversation: “Yes, this is Joe Smith from Pepsi. You’re calling from McCann-Erickson, Coke’s agency? And you want to know how much I’m going to be spending next year? Sure, I’ll be more than glad to tell you. Pepsi has firm plans to spend forty billion dollars in media next year. Maybe more.”) Similarly, would a Media Edge client give away accurate intelligence to a Zenith researcher?

So, I conclude reasonably that these guys are polling their own clients and later projecting the information on the general market.

But I have other doubts. Imagine you’re a senior marketing guy at some company, and you get a call asking you to divulge your next year’s ad spending. What do you say? You’ve been reading the papers, which seem to point to a stronger economy around midyear. That’s six months away, a safe amount of time. So there are two ways you can answer the question:

1. “In about six months my competitors are going to pick up their advertising, so we’ll be increasing our spending too.” (But the problem with that answer is that it means you’re spending more money just to maintain your share of voice. Makes you sound dumb.)

2. “We’ve suffered a really bad year, so we’ve had no money to spend on advertising. But next year we’re going to be making money again, so we’ll be able to afford to advertise. Probably beginning midyear.” (Sounds plausible, but it implies that an increase in sales precedes an increase in advertising. Isn’t it supposed to be the other way around?)

Actually, it doesn’t really matter what you answer. Recoveries are always about six months away.

-- Michael Kubin is co-CEO of Evaliant, one of the web's leading sources for online ad data.