1
Arguments against brand positioning
Maxwell Winchester & Dr. Byron Sharp
2
Marketing theory or marketing rubbish?
• Most marketing theories developed on single cross-sectional studies
• Rarely replicated– When replicated usually results in contrary
findings » e.g. Hubbard & Armstrong (1994)
• Research & Development Initiative into Marketing (school of empirical generalisationalists)– Study marketing phenomena across MSOD– In different countries– Across different market structures
3
Brand positioning
• First appeared in the Advertising Age– Reis and Trout (1972)
• Now in every marketing textbook– Seen as a fundamental aim of marketing– Yet not scientifically tested
• Position brand in consumers’ minds– Make it the preferred brand for your brand’s
target market
4
The arguments…
1. Brand image varies with usage2. Attributes that are prototypical are
prototypical for every brand3. Attitudes are fickle4. Brand image remains stable over time5. Consumers have repertoires of brands6. There is no brand segmentation
5
1) Brand image varies with usage
• Evaluative brand attributes vary with usage– e.g. “reliable”, “a bank I can trust”, “good
value for money”
• Users respond to an attribute more often than non-users
6
1) Overall scores
Bank 1 Bank 2 Bank 3 Bank 4
Up to date with Technology
58 50 52 49
Offers friendly service 25 20 18 20
Responsive 15 12 9 10
Low fees and charges 6 4 3 4
USAGE 25% 15% 15% 14%
7
2) User & non-user response level
USERS Bank 1 Bank 2 Bank 3 Bank 4
Up to date with Technology
65 59 67 63
Offers friendly service 49 43 33 48
Responsive 31 36 20 25
Low fees and charges 7 7 8 7
NON-USERS Bank 1 Bank 2 Bank 3 Bank 4
Up to date with Technology
56 48 50 48
Offers friendly service 19 16 15 15
Responsive 11 8 7 8
Low fees and charges 3 4 3 3
8
2) Response level and usage?
• Big brands score higher than smaller brands and users respond to attribute more often than non-users– Usage drives brand image/brand attitudes?
• This pattern has held up:– Different countries– Different market structures– Different industries
• If positioning theory held, wouldn’t we expect – Smaller ‘niche’ brands to show significantly higher
response levels on specific attributes (e.g. Volvo - safety)
– This response level would drive usage
9
2) Attributes are protototypical consistently
• Prototypicality comes from taxonomy– How we categorise things– In this case - brands
• Attribute that is scored highly for one brand is so for others– E.g. “Up to date with technology”
10
2) Attributes scores & prototypicality
IMAGE SCORES Bank 1 Bank 2 Bank 3 Bank 4
Up to date with Technology
58 50 52 49
Offers friendly service 25 20 18 20
Responsive 15 12 9 10
Low fees and charges 6 4 3 4
RANK Bank 1 Bank 2 Bank 3 Bank 4
Up to date with Technology
1 1 1 1
Offers friendly service 2 2 2 2
Responsive 3 3 3 3
Low fees and charges 4 4 4 4
11
2) Prototypicality arguments…
• This pattern has held up:– Different countries– Different market structures– Different industries
• If positioning theory held, would we not expect brands to score highly on different attributes?– i.e. the ones they were positioned on
12
3) Attitudes are fickle/unstable
• Only about half of the people who gave a particular attitudinal response on one occasion do so on the second interview
13
4) Brand Attitudes are fickle
Insurer 1
Insurer 2
Insurer 3
Insurer 4
Mean
Provides complete cover 81 58 32 26 49
Easy to understand policies
83 40 45 29 49
Fair on paying claims 74 42 38 43 49
Competitive on price 74 44 46 47 53
• Table of % of respondents who responded on 1st interview who also responded on 2nd interview
14
3) Attitudes are fickle/unstable
• Individuals' responses are as-if random• But this variability cancels out at
aggregate level - this is why so few researchers know about the individual variability
• If positioning theory held, we would expect– Consistent responses to the attributes
brands were positioned on» By the same respondents
15
4) Brand image remains stable over time
• While we have seen at an individual level, image responses are fickle
• At an aggregate level, over time, brand image remains stable in stable markets– Whether it 3 weeks or one year between
interviews results tend to be relatively the same
• These results are from a longitudinal study in the insurance market– Interviews were 3 months apart
16
4) Brand image @ t1 and t2
INTERVIEW 1 Insurer 1 Insurer 2
Insurer 3 Insurer 4
Provides complete cover
65 34 11 3
Easy to understand policies
59 19 7 3
Fair on paying claims 58 18 6 2
Competitive on price 53 26 17 5
INTERVIEW 2 Insurer 1 Insurer 2
Insurer 3 Insurer 4
Provides complete cover
75 36 14 4
Easy to understand policies
65 18 9 2
Fair on paying claims 62 19 6 2
Competitive on price 59 25 20 4
17
4) Brand image remains stable over time
• We see little change in the aggregate results– So brand image does not change much– Except with changes in market share
• If positioning theory held, we would expect– Dramatic changes in brand perceptions as
different competitors re-positioned their brands in the marketplace
18
5) Consumers have repertoires of brands
• Proponents of positioning theory believe:– If you position your brand well, people will
prefer your brand over all of the others
• But!– Consumers have brand repertoires– They are generally not loyal to one brand in
repertoire markets
• “Your buyers are buyers of other brands who occasionally buy you”– Professor Andrew Ehrenberg
19
6) There is no brand segmentation
• Are Ford buyers different from GM buyers?
• A fundamental argument provided by proponents of the positioning theory– Different brands are bought by different types
of people
• Study in Research & Development Initiative into Marketing– Ehrenberg & Kennedy– 42 industries, 200+ segmentation variables– Only minor differences found
20
Credit Card 1 -3 3 3Credit Card 2 -3 3 3Credit Card 3 0 0 0
Credit Card n 2 -2 2
Av. MAD 2 2 2
MADAv.
. . . .. .. ..
. . . .. .. ..
6) There is no brand segmentation
21
• If positioning theory held…– We would expect to see large demographic,
behavioural and psychographic differences between brands
» This assumes we do not hold brand repertoires» This assumes we can target different competing
brands at different segments
• We are not saying that…– You cannot segment markets
» Cat food is generally bought by cat owners!
6) There is no brand segmentation
22
So where does this leave us ?
• Evidence is not conclusive, BUT– You’ve seen a sufficient challenge to the
tradition of brand positioning
• Assumptions about the existence of 'ideal' or 'killer' attributes or image positions may be unfounded
• Users of different brands think pretty much the same thing about their brands– Just because you tell consumers something,
doesn’t mean they’ll act on it!
23
An interesting study…
• Romaniuk & Sharp (2000) found that:– Image perceptions are linked to future
buying behaviour in a systematic and predictable manner
– Mentioning a brand for any attribute means you are slightly more likely to keep buying it
– Mentioning one particular brand attribute does not lead to purchase
24
Where to from here?
• We’re not saying that your brand can not be distinct from other brands
– But we do compete in a competitive market
• More scientific studies required– Such as those conducted as part of the Research and
Development Initiative into Marketing
• The marketing task does not seem to be about repositioning to some desirable spot
– but rather is very much about taking into account what people already think of you
– Building salience for your brand may be the answer
• Perhaps brand positioning is marketing rubbish rather than marketing theory?
– Replicate & extend– Longitudinal studies