I worked for an ad measurement firm; we developed a causal ad measurement methodology based on a research tool used for observational data in clinical trials. Roughly 1/3 of the thousands of consumer packaged goods (toothpaste, canned soup, breakfast cereal, etc) studies we performed showed no effect of advertising. Another 1/3 didnt have enough of an impact to justify the ad spend (ROI < $1.00), and about 1/3 showed more revenue than ad spend.

Obviously, different industries are likely to show different distributions of results, but CPG companies are the biggest ad spenders, I believe.