The question "how does advertising work?" seems to be raised more and more often in connection with research into advertising. There is increasing awareness that a number of advertising research methods in common use imply very different assumptions about the way advertising works. Rarely - if ever _ is evidence presented that the assumptions made are correct. Also, discussions of difficulties of communication which are experienced between creative and research people often seem to point to different assumptions about the functions of advertising made by the two sides concerned.
Most of the approaches were instituted when a simple conversion theory of advertising effectiveness was generally accepted. We now, however, have a more sophisticated view of consumer behaviour, and are beginning to accept that advertising may have as big a role to play in reinforcing or intensifying existing behaviour as it does in creating "new" modes of behaviour. Does not all this imply that our chances of understanding how advertising works will be increased if we set ourselves the task of trying to understand the mechanics of consumer purchasing behaviour and work backwards from there, rather than concentrating on the advertising per se?
I want to express my concern that we should avoid at all costs the error of supposing that advertising only works one way and that the best way to communicate with people and motivate them to action is constant between people and between situations. As some of the papers have indicated we have become more sophisticated In trying to define our real target market. Complex segmentation studies by attitude groups arc carried out so that wc know whether our new product is likely to appeal to experimental or traditional housewives, whether we are trying to sell our dog food to people who treat their dogs as a substitute or as a member of the family and so on. We have learnt to be specific about whether we are trying to get users to buy more or non-users to try our brand, and to study the differences between users and non-users to see which strategy is likely to be more successful. Beyond this Dr. Joyce has pointed out in his excellent paper that some consumers are worth more than others to the advertiser because they consume more of the product and if converted will consume more of his brand.
There is little sense in testing advertising efforts without having a theory about what advertising is supposed to do. It is equally without much sense to pretest advertisements (or elements that would go into advertisements like themes, artwork, slogans etc.) without assumptions about what post- testing of the same advertisements is supposed to show. And finally: theories that cannot be tested for proof or disproof are not very useful. It should also be realised that in most fields of scientific endeavour the theories have followed rather than preceded every day-life and observations of what went on in actual practice.
The paper has been developed as a generalisation of the conditions where certain methods, advice and cautions are in order on the choice and interpretation of research. Two main ideas are brought out: the importance of the product to a person and the previous level of advertising. Before these ideas are developed further it is worthwhile to close up some blind alleys.
Mr. Stapel has found a positive correlation between recall and behaviour. Now facts are facts and facts are holy. Let's further assume that his interpretation of what he found is correct. Now other researchers have done work on the relationship between recall and behaviour. And these researchers have assembled quite a bit of evidence, that recall and behaviour show no correlation at all or sometimes even a negative correlation.
I want to stress the fact that the ultimate objective of an advertiser is not sales but profit. Some advertisers will be delighted and we may even earn a lot of money as profit-measurers. But then, an advertising campaign should be dismissed as ineffective, only because, for example, the firms buying department happens to have been so stupid as to purchase raw materials at a far too high price. And on the other hand, we should praise an advertising campaign as very effective, not for the large sales it produces but all the way to the large profits.
I am delighted to have this opportunity to share opinions with you in a discussion of how advertising works. It is a subject of much discussion and little agreement. A useful starting point in considering how advertising works is to reflect briefly on why so little progress has been made where so much desire exists. The answer lies, I think, with two obstacles. First, many of us in research simply are not approaching this subject in a way which is likely to produce anything very useful in the way of theory - that is, theory grounded in empirical findings. The major problem is that we have difficulty in proving that advertising has worked in producing the only consumer response usually worth knowing - which is sales. To talk about how advertising works without establishing that it has worked in selling the product can at best lead us to knowledge of how advertising works by luck, and at worst lead us to prolonged argument and judicious speculation.
Is it possible today for a firm which is thinking of devoting a considerable sum to the financing of an advertising campaign to obtain a fairly accurate forecast of tho results of the campaign either from its publicity department or its advertising agent? To answer this question, it is necessary first of ail to define very clearly the type of results that may be expected from an advertising campaign. Company directors and managers generally reply that by results they mean an increase in sales during the weeks immediately following the end of the campaign.
When trying to link the results that our group arrived at to the theme of this Seminar, I want to call your attention to the following. I am very content to find that we, while trying to establish the design for the post-test, almost unanimously arrived at a method which explicitly or implicitly is in essence the DAGMAR approach to advertising. Because we decided to measure the various stages of communication, which - I stress - includes the activity of selling.
I think that this seminar has provided just that, a stimulus for further thinking and research, I think that many of us, if they have not learned a lot, certainly have been awakened, to a number of aspects of the question which preoccupied us all during our seminar.
I would like to comment on Mr. Lockwood's paper. I appreciate his attempt to base his paper upon published empirical evidence. It is not our purpose here to go through a detailed analysis of every statement made by Mr. Lockwood but I would like to single out a number of them which are indicative of his finking.