There is probably no subject that has received more coverage in Admap over the past five years than low-attention processing (LAP), to the extent that the concept is starting to feel like a brand in its own right. Like a lot of successful new brands, LAP exploited an unfulfilled need in its target audience--to explain why not all good ads perform well in tracking studies or pre-tests--and then successfully muscled itself onto the platform in all subsequent debates on advertising effectiveness. Robert Heath must be given credit for having engaged the advertising research world so effectively with his LAP model, even to the extent of staging a regrinding exercise mid-campaign from the original low-involvement processing model.
A quick review of articles published in Admap and elsewhere reveals widespread acceptance of the LAP theory. In her article 'Out with the new, in with the old,' Wendy Gordon cites 70 references to 'brain science' on the WARC database.1 Most of these papers support, in one way or another, the basic tenets of LAP and its role in understanding how advertising works.
Little Hard Evidence?
Nevertheless, there seems to be a paucity of hard evidence concerning the extent to which advertising processed with low attention influences brand performance versus advertising processed with high attention. Heath's evidence is based on experimental case studies and anecdotes of campaigns that seem to have successfully exploited LAP. David Penn provided additional `in-market' evidence with an analysis of 40-plus ads from tracking studies, from which he concluded that ads did not have to be consciously recalled to have an effect.2
There is a further question mark concerning how advertisers and agencies should respond to the challenges posed by the acceptance of LAP: from the agency world, there seem to be diverging opinions.
In December 2005, Jon Howard-Spink, planning director at Mustoes, suggested that although high recall might be preferable, 'with a unique product and a compelling consumer proposition you do not need memorable advertising,' citing the example of an 'eminently forgettable' commercial for Amoy Straight to Wok Noodles, which was tremendously successful despite failing to engage viewers at a high level.3 Perhaps, then, it is okay for advertising agencies to develop communications based on the exploitation of the potential offered by LAP?
In contrast, February 2006 saw Jim Carroll, UK chairman of BBH, continue to champion the idea of `engagement' at the Admap pre-testing conference, a follow-on from his December 2004 article in Research magazine, where he said `the ability to engage, inspire and entertain is now at the heart of strategy.' Consumers need to be reached by advertising in more places more often and need to participate and be immersed in brand communications.
This vision of the future demands that communications create high involvement among consumers and, although cases like Amoy are fortunate in their exploitation of LAP, they do not provide the best model on which to develop new campaigns.
So what does Ipsos ASI have to offer to the debate? Not least, a robust database that contains appropriate measures enabling us to do two things.
- We can demonstrate that low attention processing does indeed exist and influence brand preference.
- We can show that advertising processed with high attention has a profoundly more significant impact and should be the aspiration of all forms of communication.
As a world leader in advertising pre-testing, Ipsos ASI has accumulated a database of metrics covering the performance of tens of thousands of commercials from around the world. One of the distinctive elements of our Next*TV test is that respondents are not initially aware that they are participating in advertising research, but believe they are evaluating a new 30-minute TV program; a commercial break containing the test ad (and others) is embedded. It is as close to a natural viewing environment as research is likely to come. In day-after interviewing, we confirm that viewers have watched the show before asking both recall and recognition of the test ad. This allows us to segment respondents into the groups that Heath recommends for seeking evidence of LAP:
- Those who demonstrate proven recall of the ad (for example, they can describe it in an identifiable manner)
- Those who cannot spontaneously recall the ad, but are able to recognize it from a description
- Those who neither recall nor recognize the ad, but whom we know have been exposed because they have taken part in the test
Those in the first group might be said to have processed the ad with high attention, while those in the second group can be considered to have processed the ad with low attention. Further, we have the added benefit of a third group of respondents:
Would we find any evidence of advertising effectiveness when ads are processed with this kind of ultra low attention?
In seeking evidence of advertising effect, we are focusing on a shift in brand choice pre- and post-exposure to the advertising. When being recruited, test respondents are asked to choose products across a number of categories (including the test category) to make up a hamper that they will win if selected in a prize draw. These choices are made for a second time after watching the program. This allows us to measure the shifts in brand choice pre-exposure to post-exposure at the respondent level.
The net shift to the advertised brand reflects an actual impact of the advertising, regardless of whether it has been processed with high, low, or no level of attention. To be clear, respondents may switch to the advertised brand as a conscious or rational decision based on information processed cognitively from the ad, but are free to do so for any reason--or for no conscious reason at all. The learning about LAP starts when we assess shifts in brand choice against different levels of ad recall.
Our analysis is based on responses from 97,083 respondents; 512 ads covering 65 categories, and 47 different client companies. All test ads were for established U.S. brands in FMCG categories. The key findings can be summarized as follows.
- Advertising has a significant effect on brand choice, regardless of whether it is processed with high or low levels of attention.
- Ads processed with ultra low attention also have a significant impact on brand choice.
- Ads that succeed in being processed with high attention are over two-and-a-half times more impactful than ads processed with low attention, and six times more impactful than ads processed with ultra low attention (see figure 1).
Conventional wisdom holds that LAP should be more important for established products, given that new products depend more directly on the (presumably cognitive) processing of new product information. This appears to be true. When we expand our analysis to include a further 316 ads (and 65,000 respondents) for new products, we find that the same general relationship holds for shifts in brand choice, but the difference becomes even more marked in favor of high attention processing (9.8% shift) versus LAP (3.2%) and ultra low attention (1.4%).
What evidence regarding the relative importance of LAP and HAP do we have from tracking? Within tracking studies, we can follow a similar segmentation of respondents into an HAP group (proven recall of advertising for the brand) and an LAP group (no proven recall, but recognize from a description or stills from the ad). Since ad trackers rarely use longitudinal interviewing with the same respondents over time, it is not possible to look for brand effects at the respondent level for pre- to post-exposure. Instead, we look to several brand metrics and compare how these measures differ between groups: HAP versus LAP versus no awareness. In this case, it is also important to analyze separately for brand users and non-users, since this is usually the biggest discriminating factor on key brand measures.
Tracking studies are more problematic to assess on an aggregate level. They are much less standardized, with many questions varying from study to study. Further, we can never be certain of the true exposure of any respondent to any one ad, thus recall and recognition measures may be based on multiple immeasurable numbers of exposures. Nevertheless, in almost every ad tracker we can identify a segment of LAP respondents. This ranges from near zero to almost 20% in extreme cases. Overall, in our tracking database of thousands of campaigns, we find that the LAP segment accounts for about 10% of respondents. This compares with the average level of HAP respondents at 24%, with some commercials showing 60+%.
So what do we find among these LAP respondents? In some cases, we find no brand effects, while in other cases we find statistically significant impact; sometimes we find a few image attributes are affected; sometimes future purchase intent is a little higher; sometimes unaided brand awareness mentions are better. Although the brand measures differ case by case, we conclude that we can find evidence of the effectiveness of LAP in tracking studies. However, we find the best gains among respondents who demonstrate HAP (those who have internalized conscious recall of the content of the advertising--proven recall). Thus, consistent with the pre-testing data, we find that when HAP is achieved it has a stronger and more discriminating effect on brand metrics than LAP alone.
Different Types of Ads
Before finishing, it is also interesting to note that the level and impact of LAP does not differ significantly according to the nature of the ad. We might expect that ads that depend more on cognitive communication of information would logically be more dependent on higher-order processing. Ads that rely less on information and more on image or affect are assumed less dependent on cognitive processing, and should show a greater effect of LAP. As there is no objective way to sort hundreds of ads according to type, we relied on the respondents themselves, who provide ad diagnostics within each pre-test. For example, from individual responses we can see if the ad is judged more or less informative, more or less emotional, more or less humorous, and so on. We reviewed 24 advertising attributes evaluated as standard in Next*TV diagnostics. These diagnostics are asked after our recall measures in each ad test, so they should not be considered as leading recall, but are ideal for explaining recall.
Our analysis illustrates the results for brand choice shift (percentage shift from pre to post) by the different recall groups among the ads that scored in the top tertile on each of the 24 diagnostics. In figure 2, we have provided just the extremes from the 24 measures: the more information-rich ads versus the more emotional/imagery driven ads, showing brand choice shift for both HAP and LAP groups.
This suggests that LAP is a slightly more potent force for information-driven ads than for the less informative, more fun, and emotionally driven ads. We recognize these respondent ratings are an imperfect tool to differentiate affective response from cognitive processing. Simple ratings for something like 'stirred your emotions' may not be very sensitive to underlying emotional response; for that matter, when viewers say an ad is 'informative' or 'told them something important,' that perception could be based on emotion as much as any rational evaluative process. As Penn has noted, brain research indicates that these processes are not mutually exclusive, and affective response is not always strongest for ads that consumers or practitioners might assume to be most 'emotional.'
These arguments notwithstanding, we see that across all of these groups of ads the persuasive effect of LAP remains smaller than the impact among viewers who demonstrate HAP. Thus in every circumstance, for ads of all types, it is always preferable to aim to create ads that achieve HAP.
There are three key concluding points from our database analyses:
- There is much evidence to support the contention that more HAP is better than less in terms of preference shift, a measure that has widely validated to in-market success. The relative importance of HAP in brand choice is clear from the data shown here and tallies with other validation of Next*TV metrics to in-market sales performance.
- Even though we can clearly see that LAP can have a significant impact on brand choice, there is nothing to suggest that having low HAP and high LAP is preferable. Achieving HAP seems always to be better.
- If the level of HAP is low, then certainly having high LAP is better than having no LAP. Now we are talking about qualifying a less than ideal piece of creative versus a truly poor ad. Owing to the high importance of `creativity' in explaining the variance between advertising success and failure (typically more than 75% of explained variance), there seems little reason to rely on LAP while ignoring HAP as the basis for designing and measuring the success of a new piece of creative.
In recent years, our understanding of how consumers process brand-related information has progressed, while our ability to reach them easily with our communications has degenerated. If advertisers are to create the type of buzz that successful brands display, they need to be increasingly creative, not just with the communication itself, but also with the media employed to engage consumer interest.
Consumer-generated media such as blogging and podcasting are increasing the importance of word-of-mouth as a brand communications channel, but it is a channel that only works if consumers have the words to mouth (that is, have something consciously accessible in their minds to discuss). One danger that has arisen from the growing interest in LAP is that advertisers and agencies might use it as an excuse to justify low-impact and ineffective advertising.
But the increasing difficulty of achieving active consumer engagement does not mean that advertisers should rely on trying to engage with consumers on a less conscious level instead. From the evidence presented here, we can see that the very best advertising (advertising that makes a difference for the brand), while it should connect with people on many different levels, wakes consumers from their somnolent state and gives them something worth talking about. After all, the prominence of LAP itself comes entirely as a result of its ability to create a very high level of involvement among its target audience. Where would it have been today if it had to rely on its own guiding principles?
This article has been reprinted with the permission of Admap. For more details go to www.warc.com/admap.
1. Gordon, Wendy. "Out With the New, In With the Old." IJMR, 48 (1), 2006. 2. Penn, David. "Could Brain Science Be the Peace Broker in the 'Recall Wars'?" Admap 464, September 2005. 3. Howard-Spink, Jon. "Does invisible mean ineffective?" Admap 467, December 2005.