Boosting ad impact by optimizing media placement

Pedro Almeida | April 9, 2020
';

The impact of the context in which TV ads are displayed is a major area in academic advertising research. Variables such as programming, format, and familiarity with the content, have been analysed in experimental studies, which normally use self-report measures (e.g. ad likability, recall) as outcome variables.

One important meta-analysis (Kwon, King, Nyilasy, and Reid, 2018) has shown that involvement with the context, congruency between context and advertisement, and attitude towards the program are key factors in promoting ad recall. This suggests that positive or enjoyable viewing experiences, along with congruent and relevant ads, have a significant impact on ad recall.

In our work with broadcasters and brands, we have repeatedly shown major effects of emotional contexts in promoting positive brand outcomes. A repeated observation is that there is a strong carry-over effect from the content to the ads: when people are excited with the content, they are more reactive to the ads they watch during the commercial break.

Incidentally, this means that breaks positioned immediately after engaging moments and ads that are closer to the initial positions benefit from the excitement caused by the content, and are seen as more engaging themselves.

These results have significant implications for media planning and raise big questions about the way TV advertising is traded. One implication is that advertising should be more effective is served in multiple smaller breaks (each ad is closer to the content) than in longer breaks.

To test this, our research department has designed an experimental study, where we manipulated some features such as length, position on the content, and the number of breaks. For context, in some markets, it is common to have only one long break (~15 min) per program, while in others, such as the UK, advertisement breaks are shorter (~5 min) and more frequent.

In this experiment, we recorded an entertainment show (Masked Singer), and created two versions: one version had a single, 15-minute ad break, while the other had 3 five-minute breaks (equally distributed on the content).

We served this content to the members of a panel who had not seen this programme before (each group watched one version of the content). Participants watched the content as their electrodermal activity was collected in the comfort of their homes. At the end of the exposure, participants filled a spontaneous ad recall survey.

Our main results show that ads displayed in the second condition (shorter and more frequent blocks) evoked higher skin conductance responses than ads inserted in longer breaks (p < .001; figure 1)

Figure 1. Impact score (measured from skin conductance responses) for ads delivered in one long (C1) vs. multiple short (C2) breaks

For a clearer illustration, figure 2 shows a comparison of engagement timelines for the two conditions while a specific ad frame is displayed. We can clearly see that in Condition 2 the EDA level is higher when compared to the first experimental condition.

Figure 2. Electrodermal activity elicited for the same ad in the 2 conditions

Concerning ad recall, participants were asked to name all the brands they remembered being advertised during the programme they were watching. The results displayed showed that condition 2 translated into a significantly higher percentage of recall (p = .035). In addition, some brands were only evoked in the second condition.

Figure 3. (within brand) Standardized recall rates for ads delivered in C1 (one long break) vs. C2 (multiple short breaks).

These results have strong implications for the way ad delivery is planned and measured. They also clearly show that creating engaging content is key for superior ad performance and that ads placed in premium contexts are lifted way beyond the share number of eyeballs they are exposed to.
Want to know more about these results? Contact us for queries.

share on

About the author
Pedro Almeida
Pedro Almeida


You may also like…