Key Measures of Advertising Effectiveness
Much of our work is grounded in academic theory. The following text is taken from my Masters thesis Exploring the impact of clutter on advertising effectiveness across broadcast media which I completed at the Ehrenberg-Bass Institute for Marketing Science at the University of South Australia.
Memory measures (recall and recognition)
Memory measures (such as recall or recognition tests) are still the most common method used in testing advertising effectiveness (O'Guinn, Allen et al. 2000). Most clutter studies have included a variety of recall measures, and some have also included recognition. There is, however, some variation in how the measures are believed to access memory. Recall scores are said to reflect the ability of an audience to register the category, brand and the message being advertised (Wells 2000), whereas recognition measures are said to represent a subjective estimate of advertising exposure by the respondent (Heath 2004).
Recall is the most widely used measure of memory for print and broadcast messages, although it has often been suggested that it is not the most appropriate metric (Singh and Rothschild 1983). Heath (2005) states that the use of recall (particularly brand-name-prompted awareness and detailed recall) has encouraged a simplistic view of advertising effectiveness to develop. This is because the true impact of advertising cannot be measured by simply asking people what they remember, as this relates only to conscious responses and does not account for low attention (unconscious) processing (Heath and Feldwick 2007). As such, recognition has gained much support in more recent times, as it should be capable of tapping into both conscious and implicit (unconscious) memory (Penn 2002). Indeed, in many buying situations, recognition is probably a more important metric, as consumers are required to choose/recognise from an array of alternatives (e.g. on the supermarket shelf). Even so, recall is still considered a key metric because it is easy to measure and simple to understand (Penn 2005). It also makes intuitive sense that advertisements that can be remembered must surely be preferable over those that cannot, and many buying situations still require consumers to retrieve information without prompting.
There are two main forms of recall – ‘unprompted’ (sometimes referred as ‘unaided’) recall and ‘prompted’ (or ‘aided’) recall. When recall tests are performed, respondents are exposed to a stimulus (i.e. an ad) and then a cue to retrieve it (i.e. the verbal question). This task relies on respondents to reconstruct the stimulus (Bettman 1979). When only a temporal cue is used (e.g. “do you recall seeing any advertisements in the program you just watched?”), the test is known as unprompted recall. If information other than a temporal cue is provided (e.g. a brand or category), the test is referred to as prompted recall. Different cues achieve differing levels of ad retrieval (Singh and Rothschild 1983), that is, having less-detailed cues (such as with unprompted recall) require greater mental processing by the respondent, and therefore result in lower retrieval overall.
Recognition tests require that sufficient information be provided in order to differentiate or discriminate between choices (Bettman 1979). While recall tests tend to have minimal cues, recognition tests usually involve respondents being exposed to the original material (or a version of it) (Singh and Rothschild 1983). However, recognition is more easily achieved than recall (i.e. recognition scores are typically higher than recall scores), and it has been criticised as being a less sensitive measure (Krugman 2000).
It has often been proposed that recall and recognition should be used for different purposes. For example, it is thought that recall is a more important measure of learning for products that elicit high involvement (Singh and Rothschild 1983), whereas recognition is thought to be a more appropriate measure of learning for low involvement situations (Singh and Rothschild 1983). Similarly, Krugman (2000) suggested that recall is a more complex task and is more of a left-brain activity, whereas recognition is less complex and more of a right-brain activity. Not surprisingly, research to date has also shown that recall and recognition measures are poorly correlated (du Plessis 1994), providing further empirical evidence that the measures capture different memory constructs.
It is also thought that recognition is much more enduring than recall and is less dependent on attention (Penn 2005). As the average attention paid towards TV advertising is between one third and one half of the attention paid to newspaper advertising (Heath and Feldwick 2007), it makes sense that TV should perhaps be measured using recognition (Singh and Rothschild 1983). To date, however, recognition measures have largely been used with tests of print advertising, while recall measures have typically been used to assess the effectiveness in television.
In recent advertising research, it has not been uncommon for both recall and recognition to be used to measure memories of advertising. In a test of implicit learning, Penn (2005) found little support for the claim that recognition was a better predictor of brand disposition than claimed ad awareness (recall). The research concluded that advertisers should accept that learning may be implicit as well as explicit, and that branded recognition should be used in addition to recall, to fully understand the relationship between communication and its effect on memory.
While a lot of research on advertising effectiveness concentrates on memory measures, there are other measures or aspects that advertisers should be interested in. For example, Pham and Johor (1997) suggest that while it is important that consumers remember and value the message content, it is also crucial that consumers attribute that message to its intended source (e.g. the brand). Unfortunately, this aspect of advertising effectiveness is commonly overlooked (Pham and Johar 1997; Brengman, Geuens et al. 2001).
Research from the United Kingdom and United States suggest that branding is a problem for more than half of all advertising. For example, it has been suggested that 52% of advertisements in the UK could not be branded (i.e. were ‘nameless’) (Brook 2002). Similarly, in a test of 22,000 TV commercials by a New York research firm, it was found that 35% of television advertisements were unable to be branded, and 10% were incorrectly identified as being advertising for a competing brand (Rossiter and Bellman 2005).
Misattribution to a competing brand is not only ineffective, it can also be damaging. Energizer have long been using the toy bunny to promote its battery products, yet up to 40 percent of consumers were found to incorrectly believe that the Energizer commercial was promoting Duracell (n.b. a competing brand who also used a toy bunny to promote their product) (Liesse 1990). Kennedy, Sharp and Rungie (2000) investigated occurrences of incorrect branding, whilst also identifying campaigns that were well liked by consumers. The research identified that such ‘dysfunctional’ advertisements (i.e. ads that were liked, but incorrectly attributed to a competitor) were quite prevalent, and encouraged the increased use of correct branding measurement in research aimed at uncovering the effectiveness of advertising.
Evaluative reactions towards the advertising have been shown to be linked to advertising effectiveness.
One of the most popular measures in recent research is advertising likeability.
Measures of likeability have been present in the literature since the 1930s, being one of the first variables measured to investigate reactions to advertising (Peterman 1940; Schwerin 1986).
What likeability is (and captures) has been interpreted in a number of ways:
the overall impression of the commercial (Haley 1994);
the net attractiveness of the ad in a personal magnetic sense (Cramphorn 1992);
attitude to the ad (Biel 1990; Thorson 1991; McKechnie and Leather 1998);
emotions or feelings (du Plessis 1998);
enjoyment (Hollis 1995); and
entertainment (Miller 1992).
What academics might call affect, practitioners tend to call liking (Batra 1991). However, most would agree that likeability is the viewers’ global response or reaction to an individual advertisement. That is, it captures respondents’ evaluative reaction to a brand’s advertising – how the respondent felt about it, whether they had an overall liking or disliking for the advertisement.
Contrary to popular belief, likeability is not necessarily derived from only entertainment or humour. It can be created or positively influenced by entertainment, but also by empathy and/or being relevant news. It can also be eroded or negatively influenced by familiarity (boredom), confusion and/or alienation (du Plessis 1994). Understanding the dimensions of likeability can assist in ad development, so that such elements may be incorporated (or avoided) to create more likeable advertising.
Most marketers are interested in predicting a sales change that may or may not be the result of
advertising efforts. This is particularly so for issues (or product categories) that are considered to be low involvement. The Advertising Research Foundation Copy Research Validity Project, the first extensive study of advertising pre-test measures, found that mean likeability scores were the best predictor of ad sales effectiveness compared to other commonly used advertising pre-test measures (Haley and Baldinger 2000), which broadened the support for this measure.
There are two key theories that have been postulated as to why likeability might lead to increased sales:
The first proposes that likeability of the advertisement leads to likeability of the brand (or brand preference), which in turn leads to sales. This attitudinal approach to consumer behaviour implies a strong user-engagement in which a positive affect is transferred from the advertisement to the brand (Biel 1990).
An alternative theory suggests that likeability is an immediate emotive reaction that negates ‘screening out’ of advertising, resulting in a higher level of attention being given to the advertisement (and thus its messages, brand image, etc.). This increased attention helps to increase brand salience, providing the opportunity for more extensive processing of the advertisement (i.e. greater brand awareness, stronger brand associations), and thus increased likelihood of brand retrieval from memory in the next purchase situation, which leads to increased sales (Kennedy and Romaniuk 1997).
Importantly, likeable advertising tends to be watched, whereas disliked advertising tends to be avoided.
That is, likeability acts as a ‘gatekeeper’ for further processing of the advertisement (Leather, McKechnie et al. 1994). This would explain why likeable ads achieve greater recall, as compared to disliked ads (du Plessis 1994). It may also account for the positive aggregate relationship found for likeability and correct branding (i.e. people are more likely to link the correct brand to liked advertisements) (Kennedy, Sharp et al. 2000).
Further Reading :
Batra, R. (1991). How Ad-Evoked Emotions Influence Processing of Information in the Ad. Tears, Cheers, and Fears: The Role of Emotions in Advertising, Fuqua School of Business, Duke University, Marketing Science Institute.
Bettman, J. R. (1979). "Memory factors in consumer choice." Journal of Marketing 43(2): 37-53.
Biel, A. L. (1990). "Love the Ad. Buy the Product?" Admap (September): 21-25.
Brengman, M., M. Geuens, et al. (2001). "The impact of consumer characteristics and campaign related factors on brand confusion in print advertising." Journal of Marketing Communications 7: 231 - 243.
Brook, S. (2002). Consumers left dumbstruck by local campaigns. The Australian.
Cramphorn, M. F. (1992). Launches Need Likeable Ads. B&T Magazine: 19.
Du Plessis, E. (1994). "Recognition Versus Recall." Journal of Advertising Research 34(3, May/June): 7591.
Du Plessis, E. (1994). "Understanding and Using Likeability." Journal of Advertising Research (September/October).
Du Plessis, E. (1998). "Memory and Likeability: Keys to Understanding Ad Effects." Admap(July/August): 42-46.
Haley, R. I. (1994). "A Rejoinder to "Conclusions From the ARF's Copy Research Validity Project"." Journal of Advertising Research 34(3, May/June): 33-34.
Haley, R. I. and A. L. Baldinger (2000). "The ARF Copy Research Validity Project." Journal of Advertising Research December/January: 114-135.
Heath, R. (2004). "Ah Yes, I remember it well!" Admap(450): 36 - 38.
Heath, R. and A. Nairn (2005). "Measuring affective advertising: implications of low attention processing on recall." Journal of Advertising Research 45(2): 269-281.
Heath, R. and P. Feldwick (2007). 50 years using the wrong model of TV advertising. Annual Conference of the Market Research Society, Brighton, Market Research Society.
Hollis, N. S. (1995). "Like It or Not, Liking Is Not Enough." Journal of Advertising Research 35(5, September/October): 7-16.
Kennedy, R. and J. Romaniuk (1997). How Does Ad Likeability (La) Work? ANZMEC 97, Melbourne, Department of Marketing , Monash University.
Kennedy, R., B. Sharp, et al. (2000). "How Ad Liking (LA) Relates to Branding & the Implications for Advertising Testing." Australasian Journal of Market Research 8(No. 2, July): 9-19.
Krugman, H. E. (2000). "Memory without recall, exposure without perception." Journal of Advertising Research 40(6): 49-54.
Leather, P., S. McKechnie, et al. (1994). "The Importance of Likeability as a Measure of Television Advertising Effectiveness." International Journal of Advertising 13: 265-280.
Liesse, J. (1990). "Bunny back to battle duracell." Advertising Age 61(4).
McKechnie, S. and P. Leather (1998). "Likeability as a Measure of Advertising Effectiveness: The Case of Financial Services." Journal of Marketing Communications 4(No. 2): 63-85.
Miller, C. (1992). New Study Downplays 'Likability' as Major Factor in Ad Success. Marketing News. 26: 2 & 6.
O'Guinn, T., C. T. Allen, et al. (2000). Advertising. Cincinnati, OH, Southwestern Publishing.
Penn, D. (2002). "LIP to HIP: responding to changing views." Admap 433(November): 32-33.
Penn, D. (2005). Brain science, that's interesting, but what do I do about it? Market Research Society Annual Conference, 2005.
Peterman, J. N. (1940). "The "Program Analyzer" - A New Technique in Studying Liked and Disliked Items in Radio Programs." Journal of Applied Psychology 24(No. 9): 728-741.
Pham, M. T. and G. V. Johar (1997). "Contingent processes of source identification." Journal of Consumer Research 24(December): 249-265.
Rossiter, J. R. and S. Bellman (2005). Marketing Communications: Theory and Applications. Frenchs Forest, Pearson Education.
Schwerin, H. S. (1986). Persuasion Testing: One Researcher's Experience. Copy Research: A Historical Retrospective. B. Lipstein. New York, Advertising Research Foundation: 63-78.
Singh, S. N. and M. L. Rothschild (1983). "Recognition as a Measure of Learning from Television Commercials." Journal of Marketing Research 20(August): 235-248.
Singh, S. N. and M. L. Rothschild (1983). "The effect of recall on recognition: An empirical investigation of consecutive learning measures." Advances in Consumer Research.
Thorson, E. (1991). Likeability: 10 Years of Academic Research. Copy Research The New Evidence - Eighth Annual ARF Copy Research Workshop, New York City, Advertising Research Foundation.
Wells, W. D. (2000)."Recognition, Recall and Rating Scales." Journal of Advertising Research(December/January).