Video advertising continues to be a mainstay of the marketing arsenal. The increased targeting capabilities of digital television broadcast networks, consumer adoption of broadband and the accompanying changes in consumption is driving greater interest in digital video advertising by consumers, broadcasters and advertisers. Online digital video advertising is growing rapidly driven by greater capacity to target audiences, drive engagement and measure impact. Advances in customer profiling technologies and music classification technologies provide the basis for personalised background music in advertising. Extant research is based on the traditional notion of an advertisement having static non-personalised music, a ‘one size fits all’ approach. This paper extends existing research to examine the impact of personalised background advertising in video advertisements on cognitive, affective and conative outcomes. The findings of this preliminary study suggest that the personalisation of background music can result significantly higher results for advertisement recall, attitudes towards the advertisement and emotional effects, and also purchase intention. The results also suggest there is no impact on perceived fit or music congruence where the background music is selected using music classification technologies.
1. A STUDY OF THE POTENTIAL ROLE OF
MUSIC CLASSIFICATION TECHNOLOGIES
IN VIDEO ADVERTISING
Presented by:
Mick Lynham HND, BA (Hons.), MSc. MSc. MMII
Dublin City University
2. Qualifications: Higher National
Diploma in Music Management &
Production / BA (Hons.) Degree in
Media / MSc. Marketing / MSc. Digital
Marketing
Profession: International Marketing
Officer with Trinity College Dublin /
Lecturer in Marketing, Business
Strategy & Supply Chain
Management
Mick Lynham HND, BA (Hons.), MSc, MSc. MMII
4. The new multi-screen world is changing consumer behaviour
■ The majority of media consumption is screen-based
■ Consumers move between multiple devices to achieve their goals
■ Television no longer consumes our full attention
Source: Google, 2012
6. Personalisation is a common feature of the online landscape
E-commerce Personal Shopper
Advertising
7. Digital technologies and personalisation are transforming
the consumption and engagement of broadcast media
Television Radio
Music
8. Despite advances in digital formats and delivery, video
based advertising has remained relatively static in form and
has negative impacts on consumer engagement
Audiences are largely treated as homogenous
‘One size fits all’
9. WHAT HAPPENS IF WE
CHANGE ONE VARIABLE
IN ADVERTISING?
WHAT HAPPENS IF WE
APPLIED THE SAME
TECHNOLOGICAL APPROACH
AS SPOTIFY TO BACKGROUND
MUSIC IN ADVERTISING?
10. Background Music and Marketing Outcomes
■ Music serves a variety of functions in
advertising including entertainment,
structure and continuity, memorability,
lyrical language, targeting and authority
establishment (Huron, 1989).
■ It attracts consumer interest,
communicates information and acts as a
memory mechanism (Hecker, 1984;
Park and Young, 1986; Heaton and Paris
2006).
■ Brands seek to connect advertisements
with music to facilitate the recruitment
of favourable brand attitudes,
awareness and positively influence
purchase behaviours among consumers
(Oakes and North, 2013).
11. Music classification is a method of classifying musical
content through the presentation of labels (Conklin,
2013)
■ Analyses a music file based on a set of
criteria which may include external data
■ Content-based methods use manually-
tagged metadata
■ title, artist, genre, duration etc
■ costly and time-consuming
■ Acoustic-based methods used machine
extracted metadata
■ e.g structure of music, rhythm, melody,
timbre, and the acoustics
■ non-trivial from a computer science
perspective
■ Method of utilising a large dataset of
consumers’ preferences to recommend
other potential tangibles which users
may wish to consume (Militaru and
Zaharia, 2011)
■ users’ past consumption behaviours
■ recommendations to others who have
similar behavioural patterns
■ often augmented with other data
Content Filtering Collaborative Filtering
12. DOES PERSONALISED BACKGROUND MUSIC IN
VIDEO ADVERTISING GENERATED USING MUSIC
CLASSIFICATION TECHNOLOGIES INCREASE THE
EFFECTIVENESS OF THE VIDEO ADVERTISING?
RESEARCH STUDY
13. Hypotheses
■ H1: Background music recall of video advertisements featuring personalised background
music is higher when compared to background music recall with video advertisements
featuring non-personalised background music.
■ H2: Advertisement recall is higher with video advertisements featuring personalised
background music compared to advertisement recall with video advertisements featuring
non-personalised background music.
■ H3: Video advertisements featuring personalised background music are perceived to have
less brand congruence than video advertisements featuring non-personalised background
music.
■ H4: Prior familiarity with personalised background music featured in video advertisements
results in more positive emotional feelings when compared to video advertisements
featuring non-personalised background music.
■ H5: Preference for video advertisements featuring personalised background music will be
greater than video advertisements featuring non-personalised background music.
■ H6: Intention to purchase is greater with video advertisements featuring personalised
background music compared to intention to purchase levels with video advertisements
featuring non-personalised background music.
15. Sample and Measures
■ The sample consisted of 61 participants using snowball sampling
■ Respondents were between 18 and 54 years old (M = 33.5, SD=43.1), 56 per cent were
female and 44 per cent were male.
Measures
■ Prior familiarity with the brand, advertisement video and background music was measured by
asking participants to rate their prior familiarity using a 7-pt Likert scale (Becker-Olson and
Karen 2003).
■ Recall - Participants were asked to correctly select from a list provided the following: brand,
name of collection featured in the advertisement, advertised product type, advertisement
position within the plot, the name of the background music featured in the original and
personalised advertisement (Park and Young, 1986). Distractors items included in all cases.
■ Recognition - participants had to identify which video they were exposed to during Stage 2,
and which video they were exposed to during Stage 3.
■ Attitudes towards advertisement in general - utilised bi-polar adjectives over the four
questionnaires to enable the gathering of data based on participants’ complete responses
(Rifon et al. 2004).
■ Emotions - a 7-pt Likert Scale based on Happy/Sad items was utilised to assess participants’
emotions before and after stimuli exposure, adapted from Sloboda and O’Neill (2001).
■ Affective and conative behaviour - measured following Cobb-Walgreen et al (1995).
16. Results
1. Background music for the Video 2 (Personalised) was recalled by 68.9% while only 50.8% correctly recalled the
background music for Video 1 (non-personalised).
– H1 was supported.
2. Video 2 (Personalised) was correctly identified by 65.6% while only 50.8% identified Video 1.
– H2 was supported.
3. There was no significant difference between Video 1 (Non-personalised) and Video 2 (Personalised) in relation to music
fit (congruence) and participant’s perception of the brand (p >.05)
– H3 was not supported.
4. Prior familiarity of the background music featured in Video 1 (Non-personalised) and Video 2 (Personalised) and
feelings/emotions had no statistically significant relationship, as showed by the Spearman’s rank correlation.
– H4 was not supported.
5. A Wilcoxon signed rank test was conducted to compare attitudes towards each video advertisements. The results from
the analysis showed a significant difference in the level of favourable attitudes towards Video 2 (Personalised) (Z=-4.679,
P <.001) when compared to attitudes expressed towards Video 1 (Non-personalised). Video 2 was most preferred among
participants (75.4 per cent) when compared to Video 1 (24.6 per cent).
– H5 was supported.
6. Results of a Wilcoxon signed rank test suggested that there was a significant difference the intention to purchase (Z=-
6.539, P<.001) with video 2 (Median = 5) having a higher level of intention to purchase than Video 1 (Non-personalised)
(Median = 3).
– H6 was supported.
17. Discussion & Conclusion
The findings of this research study:
■ Personalisation of background music can result in significantly higher results for
advertisement recall and attitudes towards the advertisement but also purchase
intention within consumer engagement.
■ No impact on perceived fit with brand values where the background music is
selected using music classification technologies.
■ Matching of music metadata enables fit with a brands values
■ By achieving high degrees of match between the original background music
from a technical perspective (e.g. duration, BPM, mood etc.) a higher degree of
fit is achieved.
■ From this platform, personalisation takes over and delivers the greater effects
based on familiarity, preference etc. and helps to increase consumer
engagement.
18. Discussion & Conclusion
Future research
■ Study the role of involvement in impact of personalised advertising
■ Role of other media and mode
Practice implications
■ Shift from mass consumer broadcast advertising to personalised advertising
not merely within a channel but within the advertising creative along with
enabling an increase in consumer engagement.
■ Change in the role of the brand and advertiser in background music
selection.
■ Change in the economic model for background music licensing? Familiarity
need not necessarily be as significant a consideration as in the past…
■ Another change in control and influence within the advertising ecosystem
towards technology enablers?