The benefits of archetypal prototyping when profiling market segments Chris Sherley, Charles Sturt University Mark Morrison, Charles Sturt University Roderick Duncan * , Charles Sturt University Abstract Prototyping is a market profiling technique used to develop more detailed descriptions of market segments than is possible with traditional segmentation analyses. This method describes real people that embody the salient features of a market segment. Morris and Schmolze (2006) demonstrated an approach that, unlike previous studies, used predominately quantitative procedures to identify prototypes. We extend Morris and Schmolze’s (2006) approach by utilising more sophisticated quantitative tools and archetypal interviewing techniques to build information rich segment profiles for Australian attitudes to climate change policy. Our improvements allow an improved selection of prototypical respondents, and the results highlight the importance of using both qualitative and quantitative analysis * Corresponding author: Dr. Roderick Duncan, Email: [email protected], Postal Address: School of Accounting and Finance, Faculty of Business, Panorama Avenue, Bathurst NSW 2795 Australia. 1
41
Embed
The benefits of archetypal prototyping when profiling …csusap.csu.edu.au/~rduncan/Academic/Archetypal... · Web viewThe respondent is asked to indicate any images, thoughts or words
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The benefits of archetypal prototyping when profiling market segments
Chris Sherley, Charles Sturt University
Mark Morrison, Charles Sturt University
Roderick Duncan*, Charles Sturt University
AbstractPrototyping is a market profiling technique used to develop more detailed descriptions of
market segments than is possible with traditional segmentation analyses. This method
describes real people that embody the salient features of a market segment. Morris and
Schmolze (2006) demonstrated an approach that, unlike previous studies, used predominately
quantitative procedures to identify prototypes. We extend Morris and Schmolze’s (2006)
approach by utilising more sophisticated quantitative tools and archetypal interviewing
techniques to build information rich segment profiles for Australian attitudes to climate change
policy. Our improvements allow an improved selection of prototypical respondents, and the
results highlight the importance of using both qualitative and quantitative analysis procedures.
For practicing market and advertising researchers, this new technique offers a fast and cost
effective way of researching those consumer traits such as archetypes that are often only found
* Corresponding author: Dr. Roderick Duncan, Email: [email protected], Postal Address: School of Accounting and Finance, Faculty of Business, Panorama Avenue, Bathurst NSW 2795 Australia.
We used a structure similar to the archetypal focus group suggested by Rapaille (2007). We
begin by asking the respondent about themselves, their hobbies and their family. This step is
very relaxed and occurs much like an unstructured conversation. The goal is to understand the
behavioral and psychographic profile of the respondent, as well as build rapport.
In the second step of the interview, a projective technique is applied (Donoghue, 2000) to begin
discovering why the respondent feels and acts the way they do regarding the topic. Rapaille
12
(2007) suggested that this may help to provide access to respondent archetypes. From these
archetypes we may begin to identify what marketing strategies may engage these prototypes.
Using this projective technique, the moderator begins by telling the respondent he or she is an
alien from another planet (Rapaille, 2007). They then ask the respondent to describe the
research topic (climate change) in as much detail as possible. This exercise serves to get the
respondent thinking about the topic without being directed toward specific issues by the
researcher. The light hearted nature of the exercise also allows the researcher to continue
building a positive relationship with the respondent.
In the third section, another series of projective exercises involving free association are used. A
number of relevant words such as “global warming” and “environment” are stated. The
respondent is asked to indicate any images, thoughts or words that come to mind when each
word is stated. The point of this exercise is to uncover underlying archetypal schemas that may
exist. For example, when asked about climate change, all of the respondents spoke about
clouds. This type of response indicates that marketers should be cautious of what images they
use when discussing climate change. Using images that are not clouds or the sky may challenge
Cautious and Disengaged consumers’ interpretation of the topic.
Next the respondent is shown a number of pictures with a bland, but topical call to action on
them, such as: “You can help stop climate change”. The pictures are all archetypal; for
example, the archetypal image of a polluting power plant, the clean image of a wind farm, the
farmer standing over his crop, the image of children walking into the future through nature,
and so on. The respondent is then asked to describe the pictures and indicate the emotions
they feel when they see the image. This exercise provides an indication of what marketing
imagery is likely to stimulate consumer action. This section is completed by running the same
exercise with six phrases that also make a call to action. The respondent is again asked to
communicate their feelings for each and what their behavioral response may be.
In the last part of the interview the respondent is asked their thoughts on a number of issues,
such as the role of science in the climate change debate. This section is located last so as to not
influence the respondent prior to the archetypal exercises occurring. These questions are
13
aimed at clarifying the respondent’s conscious thoughts on climate change issues. These
questions have no archetypal implications. However, they are helpful when attempting to build
contextually relevant consumer profiles.
4. Results
4.1 Quantitative
Stages 1 and 2
As previously noted, six clusters were found by Maibach et al. (2011) and then in our replication
study. From these six clusters, the two moderate segments Cautious and Disengaged were
selected for prototyping. In order to identify the core members of each segment, the
probability of group membership scores for the Cautious and Disengaged segments were
identified using multiple discriminant analysis. For the Cautious segment, respondents that
scored less than 98.7% of group probability were removed. This left a total of 167 core
members from the original 502. For the Disengaged segment, members that scored less than
98.8% probability of group membership were removed. This left a total of 129 core members
from the original 385. These people provided the basis for the archetypal pool analysis.
Stages 3 and 4
Seven socio-demographic variables were used in the LCA to define the archetypal pools. These
were: gender, education, age, household income, religion, political orientation, employment
status.
The statistics for the LCA suggested that the optimal number of clusters was larger than three
for both the Cautious and Disengaged segments. However, using more than three prototypes
per segment is not practical, and the archetypal pools for each of the segments did not present
notably different profiles when a larger cluster solution was used.
The Cautious archetypal pools were labeled: C1, C2 and C3; while the Disengaged segments
were labeled D1, D2 and D3. Each pool had a unique set of socio-demographic variables that
were not highlighted in the original segmentation. For example, the first Cautious pool C1
(n=78) was 83% women, had an average age of 39, most of the pool had completed a trade
14
certificate or TAFE course, approximately 90% of the pool did not work or worked part time,
and most had at least one child living at home. Finally, the average household income of this
segment was approximately $65,000 - $77,999 per year ($1,250 - $1,499 per week).
This profile differed quite notably from the next two pools C2 (n=59) and C3 (n=30).
Approximately 98% of respondents in C2 had full time work. The average age of the pool was
40 and the majority had completed a university degree. Gender was split with 56% of members
being male. C3 differed again having approximately 93% retirees. The average age of the pool
was 63 years old, 70% of the pool was male and the majority had not completed any education
beyond high school.
These socio-demographic differences may mean there are a number of alternative motivations
for the same attitudes (Morris and Schmolze, 2006). For advertisers and marketers, this could
mean that promotional campaigns targeted at specific segments may not be appropriately
aligned. Identifying prototypes and utilising in-depth qualitative interviews may help identify
more relevant promotional activities that would have a greater likelihood of appealing to the
various archetypal pools within a broad segment.
To isolate the prototypes, we conducted a stepwise multiple discriminant analysis. The analysis
produced mixed results. For the Cautious cluster, three distinct and tightly bound clusters were
identified. Figure 1 shows a Scatterplot of the three Cautious archetypal pools. As can be seen,
there was a small number of outliers. These members affect the placement of the group
centroids so they were removed. This step helps provide an accurate indication of the group
centroid and, in turn, the identification of the most appropriate prototypes. In contrast to the
Cautious segment, the Disengaged analysis (figure 2) produced pools that had less central
tendency. As with the Cautious analysis, outliers were removed.
A Scatterplot of the function scores for each archetypal pool was used to make the final
prototype selection. Members that were closest to the centroid were identified as the possible
prototypes. Figure 3 shows an example of this step using the C1 archetypal pool.
15
An issue with discriminant analysis at this step relates to missing data. Although the LCA was
able to run without converting missing scores to means, the discriminant analysis cannot do
this. In the Disengaged pool; D2, the most central member had a number of missing scores.
Hence this person was not interviewed.
4.2 Qualitative results
In this section, we report the results of the C1 prototype that was identified using the approach
outlined in this paper. For each of the prototypes interviewed, including the one discussed
below, the results indicated that the prototype was representative of their archetypal pool.
Furthermore, the follow up interviews revealed many relevant characteristics that were not
evident solely from the quantitative data.
Table 2 shows the quantitative profile of the C1 archetypal pool. Although this gives us some
indication of who is in the sub-segment, it is impersonal and gives no real indication of who
these people may actually be.
4.3 Archetypal profile
In order to present the final profile, we develop a narrative from the results of the quantitative
questionnaire and the in-depth interview describing the prototype. This humanizes the sub-
segments resulting in a rich profile of the people in each segment. Although this format is
similar to the more traditional Pen Portrait (Ratneshwar and Shocker, 1988), there are a
number of improvements that need to be noted:
1. We are able to avoid approximations and over generalisations. For example, if we were
describing the C1 archetypal pool we’d say that the majority of C1 think climate change
is real and it is caused by man. However, this is not entirely accurate. When we
interviewed the C1 prototype the results indicated that while she believed climate
change is caused by mankind, she also believes that the climate will always change
because of a natural cycle.
2. The prototype is given the opportunity to explain why they think and act as they do.
The original C1 profile showed that the respondents were apathetic towards climate
16
change actions. However, when given a chance to elaborate on this, the prototype
explained that she strongly supported any pro-environmental activities. The problem
was that she felt many government policies directly threatened the well being of her
family by targeting their weekly budget. Therefore, she was unable to support many
policies such as a carbon tax.
3. The researcher is given the opportunity to describe what marketing communications are
most likely to engage the sub-cluster. During the third section of the in-depth interview
the prototype showed strong support for communications that focus on the impacts
climate change may have on farmers.
A simple pen portrait is unable to provide this level of detail. Therefore, the researcher will
often have no real direction with which to begin developing further advertising research or
activities. Table 3 shows the narrative developed by the researchers for the C1 prototype to
whom we allocated the pseudonym, Angela. All of the information in the narrative was
taken from either Angela’s questionnaire results, or from the Archetypal interview. The
bolded text indicates information that was uncovered in the interview. This was also true
for the other prototypes, and gives the reader some indication of the value of the follow up
research.
5. Conclusion
Effective and reliable market profiling techniques are an important part of market research.
Successful marketing strategies can often rely on having an in depth understanding of who
constitutes a target market (Wansink, 2000). The archetypal prototyping technique outlined in
this paper offers a more nuanced consumer profiling process than traditional techniques that
may overlook important underlying trends within a market. For practitioners, this new
procedure may be beneficial for three reasons. Firstly, having a better understanding of the
underlying motivations that drive target market behaviour may allow more appropriate
marketing strategies to be developed. Secondly, for mixed methods research involving
quantitative then qualitative research, archetypal prototyping may uncover important
preliminary findings that can be utilized to help direct follow up focus groups or in-depth
17
interviews. As the number of archetypal interviews required is minimal, this can be a cost
effective way of ensuring that extensive follow-up qualitative research is focused on those
areas that are most crucial to understanding consumer behaviour or attitudes. The final benefit
for practitioners using archetypal prototyping is that this procedure can provide a more human
description of market segments than is possible using purely quantitative research. On the
other hand, it is has more statistical reliability than is perhaps the case with some qualitative
research strategies, this can be important when presenting research to clients that prefer
research with a strong statistical basis.
The motivation for this research was to perform a methodological improvement of Morris and
Schmolze’s (2006) consumer understanding framework. The results of this paper present a
number of interesting findings.
5.1 Conclusions for researchers
Firstly, we found that the archetypal prototyping technique provided a richer description of the
Cautious and Disengaged segments than what was identified in the replication of Maibach et
al.’s (2011) study. While the original segmentation analysis identified a series of consistent
attitudes, our results showed evidence that there may be a number of different motivations for
these attitudes occurring. For example, the C1 archetypal pool and its prototype, Angela,
showed that this group of people were cautious of climate change action because they felt it
may threaten their family’s fiscal situation. However, this differed from the results of the C2
and C3 prototypes. The C2 sub-cluster were opinionated professionals that considered
themselves to be well informed on the climate change debate. The results of the
corresponding interview indicated that the C2 prototype felt he was being ignored by the
government. This is why he is cautious of climate change action. Different again, the C3
archetypal prototype was a less informed retiree. His caution of climate change was driven by
underlying apathy. Although he supports climate change action, he doesn’t think he is the man
for the job.
The second finding was that the prototypes were representative of the archetypal pools. As
stated in the introduction, traditional market profiling techniques can often lead to issues of
18
over-generalising and subjective interpretations of data (Morris and Schmolze, 2006; Wansink,
2000). However, by identifying real individuals that embody the salient features of a market
segment, it is possible to reduce these limitations and increase the level of information we have
for a segment. Although Morris and Schmolze (2006) suggested in their original study that the
archetypal prototypes would in fact be representative, they provided no evidence to support
this claim. However, this study shows the technique is valid.
The third finding is that follow-up archetypal interviews with the prototypes are a useful tool
for uncovering important data. While a good preliminary profile of a prototype may be
developed using their quantitative responses, we found that the qualitative interviews gave the
researchers a much better understanding of who the prototype was and why they act and feel
the way they do.
Perhaps the most beneficial part of the qualitative interview was the ability to test preliminary
marketing communications. By exposing the respondent to archetypal stimuli we were able to
test what words, images and climate change messages were most likely to engage the
respondent (Randazzo, 2006). As the prototype is representative of their archetypal pool we
can then translate these results into what will be effective for the majority of the sub-segment.
Our final finding was the benefit the new methodology gave to the validity of this technique.
Morris and Schmolze’s (2006) original study was an innovative way of selecting prototypes from
a data set. However, the steps used by Morris and Schmolze were, at times, limited. We found
that LCA and multiple discriminant analyses used in this study had fewer of these limitations
than with the previously used K-means and single discriminant procedures. We also found that
the final two discriminant analyses used by Morris and Schmolze (2006) gave a misleading
indication of who may be the most appropriate prototype. Instead, we found that identifying
the most centrally located members using their functional scores gave a much better idea of
who would ideally represent the archetypal pool.
5.2 Other issues for further research
Nonetheless, there are a few limitations of this technique that should be noted. Wansink
(2000) suggested that prototyping may be too in-depth. This can lead to communication
19
strategies that are too specific. However, it is possible to overcome this by running a thematic
analysis over all three prototypes from a broad segment. Reoccurring themes may be treated
as applicable to the whole cluster. For example, the Cautious prototypes all responded well to
communications about Australian farmers and farmland.
Another issue is the impact of missing data on the final solution. Although the LCA can model
missing data, this is not the case with discriminant analysis. Missing scores need to be
converted to mean scores, meaning respondents with a large amount of data missing will be
moved to the centre of the cluster. For this reason, we suggest closely studying the results of
any potential prototype to ensure that they do not have an unacceptable number of missing
scores.
Despite these issues, archetypal prototyping is an innovative technique that has the potential to
produce a better understanding of target segments in a population. The results of this paper
indicate that, with the improvements to Morris and Schmolze’s (2006) original methodology,
this technique allows researchers to objectively and accurately identify respondents to be
prototypes. By interviewing these people, one is able to build comprehensive archetypal
profiles of broader market segments. From this, advertisers and marketers may develop
effective communication strategies.
20
Stages Stage name Details1 Initial segmentation analysis
Run the initial cluster analysis using attitudinal items
Select segments for prototyping
2 Identify core members
Run a multiple discriminant analysis on all clusters.
Remove the members from the selected segments with the lowest probability of group membership (approximately two-thirds of the segment)
3 Archetypal pool analysis
Run a cluster analysis on the remaining members using socio-demographic items
Select the number of clusters or 'archetypal pools'
4 Select the prototypes
Run a stepwise discriminant analysis on the archetypal pools
Save functional scores
Select members that are closest to the mean functional scores
5 Develop archetypal profiles
Conduct qualitative interviews with the selected respondents
Tables and Figures
Table 1: The five steps of the archetypal prototype analysis procedure
21
22
Figure 1: Scatterplot of the of the discriminant function scores for the Cautious archetypal pools. The results show three distinct clusters supporting the results of the second LCA.
23
Figure 2: Scatterplot of the of the discriminant function scores for the Disengaged Archetypal pools.
24
Figure 3: Scatterplot of the final scores for archetypal pool C1.
- The cross hair shows the cluster centroid. The circled members are the possible prototypes for the pool.
25
Table 2: Mean scores of the C1 archetypal pool.
26
C1 (means)
Is climate change real? (1 = yes, 2 = not sure, 3 = no)
1.35
Is it man made or natural? (1 = yes, 2 = other, 3 = natural)
1.04
Age 39.5Income
$46,800.00
EducationTrade or TAFE qualification
Gender (1 = male) 0.17
Employment status 76% - Stay at home mothers/ part time
work
Household description 61% - Couple with at least one child living at
home
Table 3: The archetypal prototype narrative
Angela Angela (35) is one of many married women in Australia that are cautious of climate change. She lives with her husband and two children in the outskirts of a capital city. Angela has completed studies at TAFE for her work in administration.
However, she now takes pride in being a stay at home mother. This puts extra strain on her family’s weekly budget meaning that they must be aware of any extra costs in their lives.
Angela believes in climate change, she also believes it should be a high priority issue for the Australian government, however, only if it will have a moderate impact on the economy.
She goes out of her way to reduce her personal impact on the environment with simple cost effective methods such as having her own vegetable garden and recycling. She believes that every person can make a difference by taking these small steps. Angela has noticed the area around her being affected by climate change and she is concerned about the impact this will have on her family’s health.
Despite these beliefs, Angela is a Liberal voter. She believes that direct action policies are the best way for Australia to reduce climate change. Angela is not interested in Julia Gillard or Bob Brown’s policies on the issue.
She feels that these politicians have lied to her and they will do it again. If she is to support climate change policies it can only be from people that are trustworthy. People that are trustworthy are those politicians that are looking out for the good of her family. Angela is threatened by new taxes as she comes from a single income household. Any new taxes may threaten her family and as a mother this is unacceptable. Instead, Angela thinks that the government should support companies that have a low impact on the environment, while high polluters should be punished.
Angela is likely to use Newspapers and pop media to find her information on climate change. She is willing to listen to people like Ray Hadley and Alan Jones but doesn’t like to hear from celebrities such as Cate Blanchett. Angela does not often discuss climate change with her friends. However, she feels that a few of them follow her thoughts on the matter.
Along with her tendency towards popular media, Angela wants to hear from people she feels are trustworthy scientists, such as the experts that are interviewed on popular current affair programs.
Angela responds to marketing communications that show how we can help Australia and her children. For example, how can she help protect Australian farmers from drought and other extreme weather conditions? Alternatively, she responds to messages that encourage the protection of her children’s future. Angela also engages with positive messages that empower her such as: “You can help stop climate change” or “Help make a better life for everyone in the future”. These sorts of messages make her feel like she is a better person when she contributes.
In summary, Angela feels that climate change is a major issue that needs to be addressed soon. She is not sure that Australia will have any real impact, but, it is worth it to try and secure a future for her kids. She does not support the current government’s policies. Direct action and punishing high polluters is the way to go, not taxing her average family.
27
28
References Babin, B., & Babin, L. (2001). Seeking something different? A model of schema typicality, consumer
affect, purchase intentions and perceived shopping value. Journal of Business Research, 54, 89-96.
Cantor, N., & Mischel, W. (1979). Prototypicality and personality: effects on free recall and personality impressions. Journal of Research in Personality, 13(2), 187-205.
Donoghue, S. (2000). Projective techniques in consumer research. Journal of Family Ecology and Consumer Sciences, 28(1), 47-53.
Hair, J., William, C., Babin, B., & Anderson, R. (2010). Multivariate Data Analysis (7th ed.). New Jersey: Pearson.
Hofstede, G. (1997). Cultural differences in teaching and learning. International Journal of Intercultural Relations, 10(1), 301-320.
Hogg, M., A., & Reid, S., A. (2006). Social identity, self categorization, and the commuincation of group norms. International Communication Association, 16(1), 7-30.
Knox, J. (2004). From archetypes to reflective function. Journal of Analytical Psychology, 49, 1-19.
Loken, B., Barsalou, L., & Joiner, C. (2008). Categorization theory and research in consumer psychology. In C. Haugtvedt, P. Herr & F. Kardes (Eds.), Handbook of Consumer Psychology (pp. 133-163). New York: Taylor and Francis group.
Madgison, J., & Vermunt, J. (2002). Latent class models for clustering: A comparison with K-means. Canadian Journal of Marketing Research, 20(1), 37-44.
Maibach, E., W., Leiserowitz, A., Roser-Renouf, C., & Mertz, C., K. (2011). Identifying like-minded audiences for global warming public engagement campaigns: An audience segmentation analysis and tool development. PLoS ONE, 6(3), e17571. doi: 10.1371/journal.pone.0017571
Maso-Fleischman, R. (1997). Archetype research for advertising: A Spanish language example. Journal of Advertising Research, 37(5), 81-84.
McCutcheon, A., L. (1987). Latent Class Analysis. California: Sage.
Morris, L., & Schmolze, R. (2006). Consumer archetypes: A new approach to developing consumer understanding frameworks. Journal of Advertising Research, 46(3), 289-300.
Peattie, K., & Peattie, S. (2009). Social marketing: A pathway to consumption reduction. Journal of Business Research, 62(2), 260-268.
29
Randazzo, S. (2006). Subaru: the emotional myths behind the brands growth. Journal of Advertising Research, 46(1), 8-17.
Rapaille, G., C. (2007). The Culture Code: An Ingenious Way To understand Why People Around The World Live And Buy As They Do. New York: Random House Inc.
Ratneshwar, S., & Shocker, A. (1988). The application of prototypes and categorization theory in marketing: Some problems and alternative perspectives. Advances in Consumer Research, 15(2), 280-285.
Shalit, R. (1999). The Return of the Hidden Persuaders: Part 1. Salon Retrieved 1/12/2010, 2010
Wansink, B. (2000). New techniques to generate key marketing insights. Marketing Research, 12(2), 28-36.
Woodside, A., G., Stood, S., & Miller, K., E. (2008). When consumers and brands talk: Storytelling theory and research in psychology and marketing. Psychology and Marketing, 25(2), 97-145.