BATAVIA, Ohio (AdAge.com) -- In the short term, marketers can get away with cutting media spending without much real harm. But that term is as short as a quarter, and the harm, once it begins, can last long after the media switch gets turned back on.

Those are recent findings of ThinkVine, a Cincinnati analytics firm that does predictive media modeling for marketers such as PepsiCo, MillerCoors and Colgate-Palmolive Co. ThinkVine CEO Damon Ragusa said lately he's been getting a lot of inquiries about the potential impact of going dark altogether for a quarter or more.

Related stories: Unilever, Walmart, P&G Buck the Short-Term Trend As Recession Forces Cuts, Leaders Build for Future With Marketing, R&D Laundry Business Embraces Risk

An analysis the firm did for one unnamed brand looked at the impact of turning off media entirely for a year, then turning it back on the next year at prior levels. For about 16 weeks, sales volume was about the same. By the end of year one, however, sales volume was about 20% lower without media than with it.

Turning media back on in year two reversed the sales decline as the brand began growing again at the same rate it would have otherwise. But it never closed the gap in sales results compared with what it would have achieved had it maintained media spending both years (see chart above).

Different brands respond differently to media cuts, Mr. Ragusa said, and some brands with dominant positions or in less advertising-responsive categories may get away with cutting budgets unscathed. But for many, possibly most, getting back sales and share lost from cutting budgets can be a lengthy and expensive process. "There is a downside risk," he said. "The cost of getting back what you lose is often greater than the savings."

A brand that is on a downward trajectory anyway because of the economy is much more at risk from a temporary withdrawal of media support than one with flat or rising sales growth, he added.

ThinkVine uses correlation analysis* and sales data. But it bases its predictions on the makeup of each brand's consumers, including detailed data on the media and spending habits of various consumer segments. It then crunches as many as a trillion data points for each forecast.

The results aren't perfect but tend to come within 2.4% of the real-world sales data in ThinkVine's validation work, Mr. Ragusa said.

~ ~ ~
CORRECTION: An earlier version of the story incorrectly said ThinkVine uses regression analysis.