Grateful as I am to the researchers who have helped me to produce effective advertising, I have nine bones to pick with them: 1 They take three months when I only have three weeks. When Eisenhower was President, the White House called Dr. Gallup one evening at six o’clock. Eisenhower wanted to know the state of public opinion on an important issue of foreign policy. The report had to be on the President’s desk at eight o’clock the next morning. Gallup sent for six of his henchmen and dictated three questions. Then each of the henchmen telephoned six interviewers in different parts of the country, and they interviewed ten people each. By midnight they had called in their results. Gallup tabulated them, wrote his report and dictated it to a White House stenographer. The report was on Eisenhower’s desk two hours before it was due. Nor is this merely an example of presidential clout. When Robert Kennedy lost the Oregon primary in 1968, his campaign manager had a research report on his desk eighteen hours after the polls closed, analysing the reasons for his defeat. When I first went to run the Audience Research Institute for Dr. Gallup, it took our statisticians two months to deliver their reports. I bullied them into telescoping the work into two days, thereby making the reports of much greater value to the Hollywood executives who were our clients. So why does it take agency researchers three months to answer a few simple questions? They are natural slowpokes, and too frightened of making mistakes. 2 They cannot agree among themselves on methodology. It recently took the Research Directors of the 21 biggest agencies two years to reach agreement on the principles which should govern copy-testing. Now they have started to debate methodology. Five years? 3 It is in research departments that you find the eggheads of the agency business. Too many of them are more interested in sociology and economics than advertising. They concentrate their attention on subjects which are only peripherally related to advertising. 4 They have little or no system for retrieving research which has