Media placement versus advertising execution

4 downloads 26300 Views 188KB Size Report
Mar 3, 2010 - sidebar ads do not consider consumer 'engagement' with the hosting site. ... buy space in magazines and TV programmes; they do not buy individual .... domain. No set of indicators would be exhaustive of this domain but this.
International Journal of Market Research Vol. 52 Issue 2

Media placement versus advertising execution Edward C. Malthouse and Bobby J. Calder Northwestern University

We make three contributions towards understanding how engagement with the surrounding editorial context affects reactions to ads. First, while previous studies have shown that respondent-level engagement affects ads, we argue that vehicle-level engagement is more relevant to placement decisions, and show that magazine-level engagement affects actions taken from seeing an ad. Second, we compare the relative importance of engagement to the execution factors size, position and colour, and show that engagement is of comparable importance. Third, evaluations are done with more realistic procedures than previous studies and with real ads.

Introduction The effectiveness of an advertisement depends on the product being advertised and characteristics of the execution, such as the quality of the ad itself, the size of the ad and location within the medium (e.g. back cover, inner front cover). An additional factor, which has received substantial attention in the advertising community, is reader engagement with the media context itself. Many studies have shown that when consumers are highly ‘engaged’ with a media vehicle they can be more responsive to advertising (e.g. Aaker & Brown 1972; Feltham & Arnold 1994; Coulter 1998; Gallagher et al. 2001; DePelsmacker et al. 2002; Nicovich 2005; Bronner & Neijens 2006; Cunningham et al. 2006; Wang 2006). Others have called for additional research on how the surrounding context affects reactions to ads (Galpin & Gullen 2000; Baltas 2003, p. 512). While this conclusion is not surprising, its implications for media planning are potentially profound because the price of most advertising is determined by audience size and execution factors such as position and Received (in revised form): ?? ?????? ????

© 2010 The Market Research Society

1

DOI: 10.2501/S147078530920????

Malthouse.indd 1

03/03/2010 09:21:01

Media placement versus advertising execution

size, without considering engagement in any formal way. For example, the price of print advertising is determined by circulation, the location of the ad within the publication and characteristics of the ad such as the number of colours. Likewise, the algorithms used to place banner and sidebar ads do not consider consumer ‘engagement’ with the hosting site. If engagement with the vehicle affects reactions to ads then it should be considered when deciding which vehicles to use. At the same time, advertisers are searching for ways to overcome the problems of ad clutter and avoidance (Cho & Cheon 2004), and media organisations are going bankrupt as advertising revenue streams disappear and circulation declines. Focusing on the media context is a potential solution to both problems. For advertisers, who have (at least some) control over where their ads appear, we know that context can affect reactions to ads and should therefore be considered when selecting media vehicles. Media organisations with exceptionally engaging content argue that they should be able to charge a premium price for their advertising inventory. Both advertisers and media companies are searching for medianeural metrics for the purpose of common-currency comparisons – for example, a website with a print vehicle (Winer 2009). One reason why media planners have not explicitly accounted for media context in their models is that they have not had systematic and widely available measures of context, but this is changing. Bronner and Neijens (2006) have developed scales to measure the consumer’s experience with different media and shown that these measurements predict reactions to ads. Calder and Malthouse (CM) have independently developed similar experience scales and shown similar predictive validity (e.g. Malthouse et al. 2007; Calder & Malthouse 2008; Calder et al. 2009). They have also conceptualised media engagement as the collective experiences that a reader has with the editorial content, and shown how to measure it as a second-order construct from first-order experiences. The Bronner–Neijens and Calder–Malthouse studies have advanced our understanding of media context and ability to measure it, but many additional questions must be answered before experiences and engagement can be routinely incorporated in media-placement decisions. This research makes three contributions towards this goal. First, previous research has focused on the effect of the respondent’s engagement with the vehicle on ads, but media are usually purchased at the vehicle level. Advertisers buy space in magazines and TV programmes; they do not buy individual readers or viewers. This research tests whether the average level of engagement with a vehicle affects reactions to ads.

2

Malthouse.indd 2

03/03/2010 09:21:01

International Journal of Market Research Vol. 52 Issue 2

The second contribution is to compare the potency of engagement (media placement) with that of execution factors currently used to price advertising, in predicting advertising effectiveness. This raises the possibility of ‘trading off’, for example, ad size or location with selecting more engaging vehicles. In doing this, we attempt to monetise the value of engagement. The third contribution concerns the robustness of the engagement effect. Does the relationship between engagement and advertising effectiveness hold up over a broad cross-section of real ads from different product categories? And, along the same lines, this research employs a more stringent, behavioural measure of advertising effectiveness and a more realistic testing procedure than previous studies.

Literature review and hypotheses Media experiences and engagement There are many independent streams of research examining consumers’ experiences with media. While there is substantial overlap between the experiences posited by the different streams, unfortunately they are not entirely consistent. Certain experiences exist in some frameworks but not others. Among the experiences that consistently exist in multiple frameworks, there are often subtle differences in the way in which they are conceptualised. In some cases, multiple experiences under one framework are subsumed by a single experience of another. The uses and gratifications (U&G) approach provides a functionalist explanation of why people use media and has been an active area of research since the 1940s (e.g. see Ruggiero 2000 for a recent survey). The U&G literature is vast; McQuail (1983, pp. 82–3) gives a concise summary that is often cited: •



Information – finding out about relevant events and conditions in immediate surroundings, society and the world; seeking advice on practical matters or opinion and decision choices; satisfying curiosity and general interest; learning, self-education; gaining a sense of security through knowledge. Personal Identity – finding reinforcement for personal values; finding models of behaviour; identifying with valued others (in the media); gaining insight into one’s self.

3

Malthouse.indd 3

03/03/2010 09:21:01

Media placement versus advertising execution





Integration and Social Interaction – gaining insight into the circumstances of others: social empathy; identifying with others and gaining a sense of belonging; finding a basis for conversation and social interaction; having a substitute for real-life companionship; helping to carry out social roles; enabling one to connect with family, friends and society. Entertainment – escaping, or being diverted, from problems; relaxing; getting intrinsic cultural or aesthetic enjoyment; filling time; emotional release; sexual arousal.

U&G approaches have been used in marketing. For example, Bronner and Neijens (2006) measure eight experiences that are consistent with the U&G approach: practical use, social, identification, pastime, transformation, stimulation, information and negative emotion. The Calder–Malthouse experiences are also consistent with U&G. Both teams show how to measure different aspects of the U&G framework. Nambisan and Baron (2007) applied a variation of the U&G constructs to explain virtual customer environments with four experiences: cognitive, social integrative, personal integrative and hedonic. Childers et al. (2001) discuss utilitarian and hedonic (a type of ‘entertainment’ in the U&G approach) experiences as explanations of online shopping behaviour. The same approach is also followed by Fiore, Kim and Lee (2005) and Cotte et al. (2006). Flow is another construct that has received substantial attention (see, e.g., Hoffman & Novak 2009), and is consistent with the U&G approach of understanding the consumer experience with media. Clearly there are many different dimensions of consumer experience with media, and different media vehicles create different experiences for their readers and viewers. Calder et al. (2009) propose that different experiences are manifestations of the second-order construct they call engagement. There are many different ways of being engaged with a vehicle. For example, some media are engaging because they inform their consumers and give good advice. Other media help their consumers relax and escape from the pressures of daily life. Some media could do both. Engagement is a higher-level measurement of consumers’ relationship with the surrounding media context than individual experience measures. Both are useful – experience measures provide a greater level of specificity, while engagement provides an overall measure. It is unnecessary for purposes of this article to sort out differences in the ways that various frameworks have conceptualised experiences because, for the purpose of measuring engagement, all we need is a set

4

Malthouse.indd 4

03/03/2010 09:21:02

International Journal of Market Research Vol. 52 Issue 2

of experiences that can serve as indicators of the engagement construct domain. No set of indicators would be exhaustive of this domain but this is not required from a measurement point of view.1 Our approach is to develop scales measuring a representative set of experiences that parallel those noted in the literature. We shall then factor-analyse the experience measures and test whether they could plausibly be manifestations of a second-order engagement construct.

Engagement and advertising effectiveness Bronner and Neijens (2006) compare the experiences of different types of media with the experiences of advertising content. Malthouse et al. (2007) and Calder et al. (2009) show that various experience measures are associated with copy-testing measures of magazine and banner ads. They find, for instance, that the consumer’s experience of usefulness with a site is related to the ads on that site being experienced as useful. There are several theoretical explanations for why engagement should affect reactions to advertising, including categorisation theory (Cohen & Basu 1987) and affect transfer (e.g. Broniarczyk & Alba 1991, p. 215). Dahlén (2005) does a literature review of media context effects and summarises three possible theoretical rationales for why context should affect reactions to ads. The first is the mood congruency–accessibility hypothesis: ‘The ad context makes a certain mood or affect more accessible and relieves the processing of stimuli with similar moods or affects’ (p. 90). The second is the congruity principle: ‘the medium and the advertised brand converge and become more similar in consumers’ minds’ (p. 90). The third is that the context serves as a cognitive prime that ‘activates a semantic network of related material that guides attention and determines the interpretation of the ad’ (p. 90). It should be noted that these explanations are not alternative explanations but, rather, all of them are plausible mechanisms for how media context can affect advertising. They lead us to formally hypothesise:

1   The question arises of whether to treat experiences and engagement as formative or reflective. We follow Jarvis et al.’s (2003) criteria for making the decision. We treat both as reflective (a Type I second-order factor specification, in the language of Jarvis et al.). In the case of experiences, the items are manifestations of some experience, are interchangeable and should covary. The items we have used represent a sample from the respective construct domains, e.g. there are many ways that a person can have a utilitarian experience and different items could represent the construct domain equally well. Thus, experiences are reflective according to the Jarvis et al. framework. We also think of engagement as a reflective construct because we view experiences as manifestations of engagement (reflective) rather than as ‘defining characteristics’ (formative).

5

Malthouse.indd 5

03/03/2010 09:21:02

Media placement versus advertising execution

H1: The average level of engagement with a media vehicle increases advertising effectiveness.

Ad effectiveness

The distinction between this hypothesis and those tested previously is that the relationship is studied at the vehicle level rather than respondent level. Both levels are important; media buyers typically buy a vehicle rather than individuals and so we should test the relationship at the vehicle level. It is possible for individual-level engagement to be associated with advertising effectiveness, but not vehicle-level engagement. While informative, it could be the case, however, that this overall level of engagement might not be associated with greater ad effectiveness. This could be true even though at the level of individual readers there is an association between engagement and advertising effectiveness. Figure 1 illustrates how this could be the case. The scatter plot for each vehicle is shown by the ellipses. For any given advertising vehicle, the readers or viewers who are more engaged have higher average ad effectiveness. Notice, however, that there is no relationship between the level of engagement with the vehicle and ad effectiveness. The present research, therefore, addresses the question of whether the overall average level of engagement with a magazine affects reactions to ads.

cle

i Veh

cle

5

i Veh cle

i Veh le hic

cle

i Veh cle

i Veh

1 cle

i Veh

3

Ve

4

7

cle

i Veh

8 cle

i Veh

6 9

2

Engagement

Figure 1 Possible relationship between engagement and ad effectiveness, where, within any vehicle, individual-level engagement has a positive relationship with ad effectiveness, but there is no relationship between vehicle-level engagement and effectiveness

6

Malthouse.indd 6

03/03/2010 09:21:05

International Journal of Market Research Vol. 52 Issue 2

Methodology and results Survey This research examines 72 magazines across 15 genres and over 5000 ad executions using the VISTA Print Effectiveness Rating Service offered by the Affinity copy-testing service, which has become an industry standard in the United States. The 72 magazines are the largest in the US. For each of the 5000 ad executions, we know: (1) the characteristics that usually determine their price, including size (third page, half page, full page, two-page spread, multi-page unit), position (run of book, back cover, inner back cover, inner front cover), and colour (four colours or fewer than four); and (2) a measure of advertising effectiveness for a sample of readers. Affinity surveys readers by first screening potential respondents for issue-specific readership by showing the front cover of the magazine. Next, qualifying respondents are shown an ad surrounded by its original context and asked if they recall seeing it. Those with affirmative recall are asked to check which actions they took from a short list. This process is repeated for approximately 20 ads. Respondents are then asked about their engagement, readership and demographics. This is a more realistic procedure than those in previous studies because the ad is surrounded by the original editorial content. Each advertisement and editorial feature is viewed by a minimum of 100 respondents, with 75% of respondents rating exactly 20 advertisements and two editorial features (the mean number of ads rated was 19.6 and the standard deviation was 1.6). There are 25,705 unique respondents in this research for the readership measurement, and each of them focused on only one of the 72 magazines. The 25,075 individual respondents yielded a total of 503,443 ad effectiveness observations.

Scales The scales used to measure four experiences (Table 1) were developed in Malthouse et al. (2007). Constraints on the length of the survey and respondent fatigue limited us to four experiences, which were selected to represent a range of magazine experiences from the U&G framework. Advertising effectiveness is measured by showing respondents an ad surrounded by the original editorial content, and then asking if they recall seeing it in the particular issue. Those who did not recall seeing the ad were shown the next one. Respondents who did recall seeing the ad were asked a series of questions about brand awareness and the following seven

7

Malthouse.indd 7

03/03/2010 09:21:05

Media placement versus advertising execution

Table 1  Questions and standardised factor loadings from second-order model. Utilitarian

Loading

The magazine gets me to try new things I really like the tips in the magazine I use the magazine to learn how to make things It shows me how to do things the right way

0.7669 0.8056 0.6441 0.7506

Transportation I lose myself in the pleasure of reading it It’s an escape The magazine takes my mind off other things that are going on I like to picture things in my own mind while I am reading it I can picture myself at the scene of the events and places described

0.7868 0.8381 0.8268 0.7121 0.6551

Makes me smarter It addresses issues or topics of special concern to me It’s important that I remember later what I have read in this magazine It updates me on things I try to keep up with I look at the magazine as educational. I am gaining something

0.7934 0.7762 0.7941 0.7863

Sophisticated The magazine is very sophisticated It is very professional The articles really are in-depth They do a good job of covering things. They don’t miss things

0.7375 0.8100 0.8175 0.7969

Engagement Utilitarian Transportation Makes me smarter Sophisticated

0.8672 0.8146 0.8172 0.6340

actions they could have taken as a result of seeing the ad: being more favourable to the brand, gathering additional information, visiting the brand’s website, visiting a store, recommending the product to someone, purchasing the product, or taking another action. Instead of looking at those actions individually, we count the number ticked as our measure of actions taken.

Measuring respondent-level experiences and engagement We develop a scale to measure engagement in two steps following Calder et al. (2009). First, the measurement properties of the four experience scales are studied with a confirmatory factor analysis measurement model, allowing correlations between all first-order experience factors. The model fits acceptably well with GFI = 0.90, CFI = 0.92, NNFI = 0.90,

8

Malthouse.indd 8

03/03/2010 09:21:05

International Journal of Market Research Vol. 52 Issue 2

and RMSEA  =  0.0683. Table 1 gives the question wording, estimated standardised loadings, and reliabilities. Next, we estimate a second-order factor analysis model, where the covariances between the experience factors are hypothesised to be due to a second-order engagement construct. The fit statistics are good, with GFI = 0.89, CFI = 0.91, NNFI = 0.90 and RMSEA = 0.0775. Engagement will be estimated by the simple mean of the experience scores. The average value of respondent-level engagement is 3.52, the standard deviation is 0.50, the quartiles are 3.17, 3.55 and 3.88, and the range is 1 to 5.

Mean engagement by magazine The first contribution of the paper is to study how magazine-level engagement affects ads. To do this, we computed the average level of engagement for each of the 72 magazines. Mean engagement can be more actionable than respondent-level engagement because advertisers buy a magazine rather than individual readers. The average of the 72 values of magazine-level engagement is 3.52, the standard deviation is 0.15, the quartiles are 3.44, 3.54 and 3.62, and the range is 3.00 to 3.82. Thus, some are more engaging than others; we would like to know whether ads placed in those with higher means perform better.

Explaining ad effectiveness with engagement and execution variables We estimate a hierarchical linear model (HLM) predicting actions taken for those who recall seeing the ad using magazine- and respondent-level engagement, and the execution factors size, location and colour.2 More precisely, for respondent j of magazine i, yij is the value of actions taken and xijk is the value of engagement or the execution dummy k (k = 1 for magazine-level mean engagement, k = 2 for individual-level engagement, and k = 3, …, 11 are for execution dummies). Then 11

yij = α + β1 xij1 ∑ (βk + bik )xijk + eij k= 2

where a and bk are the intercept and slopes across all ads, bik are normal random variables with means 0 and standard deviations sbk quantifying 2   Some levels of the execution variables have very small sample sizes and these are dropped from this analysis. For example, only one of the ads was a ‘centre spread’, so ratings of this ad were dropped from this analysis.

9

Malthouse.indd 9

03/03/2010 09:21:05

Media placement versus advertising execution

specific effects for magazine i, and eij is a residual with mean 0 and standard deviation se.3 Table 2 reports parameter estimates from the HLM. The first contribution of the paper is to test whether engagement at the magazine-level is related to ad effectiveness. The magazine-level slope is 0.9972, which is significantly different from 0 and positive, indicating that ads appearing in magazines that are, on the average, more engaging have more actions taken. This implies that the effectiveness of ads can be improved by selecting more engaging media vehicles. The slope for respondent-level engagement (0.7851) is also positive and highly significant, indicating that, among the readers of a magazine who recall seeing the ad, those who are more engaged with the editorial content take more actions. The random effect for individual-level engagement (0.1005) is significant, indicating that the effect of respondent-level engagement on ads varies across publication; Table 2  Parameter estimates from HLM predicting actions taken (n = 258,882) Parameter

Fixed effects b

Std err (b)

Random effects P-value

s

Intercept

5.8977

0.6986