3.  Controversial low estimates of advertising’s effect are coming to light. The same network study came up with one of those low estimates: 3.5%. Paul Tarr of MSW (McCollum Spielman) elicited this piece of information in answer to a question during the Q&A session. Similar concerns about the prevalence of low estimates were heard in the ROI forum. Michael Hess of PDI described what his firm was doing to check the 3%-6% estimates he says have often emerged from scanner-based analyses. He wants to see if they fall as far short of advertising’s "true" contribution as many are beginning to believe. We may finally be seeing some of the numbers that led to package goods’ long trend away from advertising toward price promotion – and lower profits.

4.  Those estimates are low because "Advertiser triers" make more repeat purchases. That was the conclusion of Jeff Hunter of General Mills. They had participated in beta testing of PDI’s 500,000 household panel that measures what scanners don’t - what happens after a person first tries a product. He found those who tried a product because of advertising made a lot more repeat purchases than those who tried it because of trade promotions. The result? The short-term effect of immediate trials - the effect that can be measured by scanners - added 4% to sales. But when they added in all the repeat purchases made by those households they came up with a more substantial total long-term contribution of 10%. Rosie Ware, Millward Brown CEO, made a similar point during this year’s "Great Debate." The question was: Is advertising good for short-term or long-term effects? In support of the latter she cited a number of estimates of the percent of sales due to advertising. IRI, recapping its scanner data, says it’s only 4%. The Marketing Science Institute said it’s 12%. She said that was an example of the difference between the short-term effects and the long-term effects. She cited additional studies where the ratio was as high as 9 to 1.

5.  Good Advertising has both short-term and long-term effects. This year’s "Great Debate" became something less than that when both sides expressed unanimous agreement with the evidence showing there are no "sleeper effects". No one had seen any advertising that finally shows some measurable impact over the long-term, without having shown any impact in the short-term. Advertising has both short and long term effects, or it has no effect. The closest thing to controversy came from Meg Blair, head of rsc. She took the opposite side saying you need to focus more on the short-term effects, and reviewed a lengthy bibliography of studies that have shown ad recall is a poor measure of ad effectiveness. (Among her opponent’s most popular products are Millward Brown’s telephone ad tracking surveys – based on recall.) Somehow, she missed the classic "How Advertising Works" that Wharton’s Len Lodish did for IRI; the one where he not only slammed recall, but also made a point of saying he could not replicate the results Meg Blair has been citing to show the validity of her ARS persuasion measure.

6. Tracking studies conducted online received a lot of attention. The emphasis speakers placed on the accurate measurement of advertising’s contribution to sales, and advertising’s ROI, was reflected in the number that described the tracking studies that made those measurements possible. Judi Jones of Levis described tracking 17,000 17 to 29 year olds. Mike Duffy of Kraft described the consumer tracking they have long conducted and their more recent use of causal modeling. NFO and Seagram described tracking by re-interviewing the same respondents. Perhaps the greatest amount of attention was given to the newest type of tracking conducted online as described by Lieberman, Burke, Harris and our firm.

9.  Advertising’s effect can be tracked better, faster and cheaper online. This was the title of our presentation. Online tracking is better because it lets you show pictures. Tracking by phone is less accurate because when people say they recall your advertising they may be thinking of your old advertising, or even your competitor’s advertising. When you can show something and ask "Did you see this?" accuracy improves dramatically. It is faster because the entire sample is contacted immediately. Half the eventual replies come in within 24 hours. Almost all are in 5 days. It is cheaper because you have eliminated virtually all of the traditional costs of interviewers and their facilities. We proved it works by comparing our regular Super Bowl results with results from three different types of online interviewing. They all worked. In the first we had Bernett Research phone people to see if they were willing to participate and reachable online. If not, we got their street address and mailed them a questionnaire. Next, we used the new InterSurvey panel, which recruits from both the online and offline segments of the public. Then it equips all of them with a Web TV to use in taking their surveys. Finally, we tested an "opt-in" sample of football fans from Survey Sampling’s E-Mail Lite database of 7 million e-mail addresses. In each test we found the commercials that performed the best and the worst were virtually the same. And, at least for these Super Bowl commercials, we found they ranked essentially the same whether we factored in results from the offline segment of the population – or ignored them.

10.  Larry Mock fills in some missing history. Some of the best insights from any conference can come from chance conversations. This time it was with Larry Mock who is leaving as head of research at Procter & Gamble. It was about the evidence P&G had that Day-After-Recall was a valid method of measuring advertising effectiveness. For over a quarter of a century it was the industry standard. But it virtually disappeared after performing poorly in the ARF’s Validity Study in the early 90’s. As an old agency researcher I can testify the only validity claims I ever heard from Burke about their Day-After-Recall procedure was that it was developed by P&G, the world’s largest advertiser (at the time), and it was done just the way P&G wanted it done. Ted Dunn and his committee took eight years to complete the ARF Validity Study because they started by asking advertisers to pool the validity evidence they had in their files. P&G declined, so the ARF had to start over and conduct its own study. When it showed the old industry standard was not one of the better predictors there was quite an uproar. 40 major advertisers contributed to IRI’s original "How Advertising Works", in part to check the ARF results. Again P&G declined to participate. Conventional wisdom at the time was that P&G had no evidence DAR was valid. What Larry Mock said was that the findings of the two validity studies were no surprise to P&G. As early as 1982 P&G decided to switch away from DAR to a persuasion based system. They had known for some time DAR was a flawed procedure, and according to Larry Mock, didn’t see any reason to let the rest of the industry know "their" technique didn’t work. But Larry did not lend any support to the P&G watchers who felt there was never any evidence of validity. He said there was. He had seen it. It just turned out to be wrong.

(These are only highlights of things that caught my interest. For more complete details call the ARF at (212) 751-5656 and order a copy of the proceedings.)

Don Bruzzone, April 2000