Fast Maturing Chinese Consumers
Over the past five years, patterns of consumer brand buying and attitudes towards brands of
televisions and mobile phones show that Chinese consumers are becoming sophisticated and
savvy. They buy similar products and brands, face similar marketing (if not at quite the same
level) and have brand loyalty and switching patterns similar to those of consumers in more
developed markets, and though brand loyalty levels are low, they are rising. The Duplication
of Purchase Law holds, as does double jeopardy.
The Chinese marketplace will probably remain dynamic for some time, both in terms
of growth and in terms of market share shifts among competitors. Marketers of brands with
broad appeal based on category drivers will likely grow the most whereas highly targeted
brands, or those trying to build on a base of loyal customers are likely to decline.
Keywords: Brand Loyalty, Duplication of Purchase, Dynamics, Consumer
Reader in Marketing, London South Bank University
During the past five years China‟s economy grew by more than half and its markets became the
worlds biggest for many things, including mobile phones and televisions. Millions of its people
enjoyed increased prosperity that enabled them to consider an ever-increasing array of products,
brands and commercial messages. Yet while much has been written about rising consumer spending,
little research has addressed whether Chinese consumers behave like those in more developed
markets, or in a particularly Chinese manner.
It might be tempting to assume that in China, with its unifying culture, consumer behavior will
take on a culturally determined character. This is not to ignore China‟s many distinct cultures,
ethnicities, languages, regional differences, etc. but to stress the common elements, shared social
values, collectivist orientation and strong group norms that may translate into consistent “Chinese”
buying patterns, or a distinctive model of consumption.
On the other hand the globalizing economy means Chinese consumers may choose between
products and brands that are very much like those in France, Japan or the United States. Global firms
seek economies of scale in marketing and global brands tend to be marketed consistently round the
world. Chinese consumers might therefore react with purchasing behavior similar to that found in
2. Research Aims
The main aim of this research was to describe brand buying in China, to see whether Chinese
consumers of television sets and mobile phones were as brand loyal, and how their behavior evolved
over time. The study ran over five years to 2008, gathering data on brand purchases and brand choice
criteria. The main research problem breaks down as follows:
(1) To describe the television and mobile phone markets in China, including growth, market
participants, competition, and market structure.
(2) To describe consumer brand buying behavior for televisions and mobile phones, particularly
brand loyalty and switching and how behaviour evolved over five years.
(3) To outline stated consumer preferences and buying motivations for TVs and mobile phones
and how they evolved during the study period.
(4) To lay out marketing implications of underlying consumer buying behavior for television
and mobile phone manufacturers.
The main finding was that Chinese consumers rapidly developed brand buying habits much like those
in other parts of the world. Their brand loyalty was lower, but rising. There were other mostly small
differences, discussed below, but by and large, Chinese consumer behavior assumed a character very
familiar and similar to that found in Western markets.
3. Dynamic market conditions for TVs and mobile phones
Televisions and mobile phones top the list of consumer ‟must have‟ items in China, far above
refrigerators or washing machines (Zheng and Cui, 2007). TVs and mobile phones are categories in
which China is both the largest producer and consumer nation. Mobile phones are bought particularly
by younger people and professionals who are partly motivated by fashion and by desires to be socially
connected, to by-pass the official bureaucracy of acquiring landlines, and obtain some privacy for
communication outside of home (Gallup Organization, 2006). TVs on the other hand provide
information and entertainment for its own sake and are on prominent display in the best room in the
house as emblems of modernity, wealth and status (Chen and Penhirin, 2004).
Ownership of televisions and mobiles is therefore a big issue--both are status symbols in the
way that cars and big houses are for Americans (Lu and Peng, 2000). As the market expanded, prices
fell, even for newer technologies. This was true throughout the country but in the coastal zones (Cui,
1999), home to 400 million people, lead to household penetration for both goods of nearly 100% with
most households having several of each. This multiple ownership and repeated purchasing allow us to
examine how consumers exercise brand preferences.
Chart 1 shows market share
Chart 1. Big local brands gain, then lose share
dynamics in the Chinese TV
25 market. Sales rose from 10m
sets in 2000 to 30m in 2007,
TCL Despite this growth many
smaller companies (Others)
15 dropped out. These regional
companies had made up a
quarter of the market, but by
Global 2008 were reduced to about
Others 4% by fierce competition
from national firms such as
TCL, Changhong, Konka and
pre 2000 2004 2005 2006 2007 2008
Hisense-- big firms that rode a
wave of prosperity by selling
competitively priced cathode ray tube (CRT) televisions.
After 2005 though, as wealth rose and prices fell further (popular 32inch LCD TVs sold for RMB
5000 in mid 2008, 16% less than in 2007), (Euromonitor 2008A) buyers increasingly chose global
brands such as Sony, Samsung, and Global (Toshiba, LG, Panasonic, Sharp and Philips), perceived to
have superior quality and technology—especially in flat screens, High Definition (HD), Liquid Crystal
Display (LCD), Plasma Display Projection (PDP) and Laser projection. As a result, bigger Chinese
brands lost share to global competitors that were pushing newer formats to consumers moving up or
into new technology.
In total, during the past 5 years the market consolidated amongst the big brands while smaller
ones exite-- a prime example of a market shakeout. In addition, the large local brands that built huge
volumes with traditional CRT models found themselves challenged by global competitors who
capitalized on their strength in the high end of the market to increase market share through sales of
new formats and technologies.
The situation in mobile phones was different in that global firms have always dominated.
Chart 2 shows that Nokia, Motorola and Samsung went from controlling about 60% of the market in
2004 to almost 70% in 2008. Meanwhile the largest Chinese firm, Ningbo Bird, fell from 11 to 8%. As
in TVs, the industry consolidated and global brands gained share.
Another similarity to TVs was that global mobile phone brands were perceived to have better
quality and technology. Indeed the top three brands in China are the top brands in most markets round
the world. In China this high-end domination meant that these brands captured about 70% of unit sales
but over 80% of revenues (Euromonitor, 2008B). The Chinese brands on the other hand, competed
mainly with me-too technology and on price.
Overall, the world‟s fastest
Chart 2. Global giants dominate Mobile
growing economy had the
worlds fastest growing 35
mobile phone sector. 550
million mobile subscribers 30 Nokia
(china.org.cn) generated Motorola
phone sales of over RMB
144 billion in 2007. This 20
rapid growth was fuelled by LG
rising incomes and falling 15
price (-15% from 2006‟s 10 Amoi
level to RMB 1,400 (about Konka
USD 200) in 2007) driving 5 Other
sales volume up 16.7% but
revenue only 11.1% pre 2004 2004 2006 2008
In summary, The Chinese television and mobile phone categories grew faster than those in most other
parts of the world over the past five years. This growth was accompanied by rapid expansion of
supporting infrastructure, distribution and customer support systems. These categories were also
dynamic in terms of large changes in market share between the main competitors, and in the case of
TVs the exit of a number of smaller firms.
Buying behavior was examined at the brand level, tackling preference and loyalty first through
behavioral data covering brand purchasing, then through attitudinal data relating to brand choice. Each
respondent reported the brand of TV or mobile they bought on their last two purchases. Only
respondents who had made at least two purchases were included. The data were gathered in eight
tranches over five consecutive years (2004-2008) for televisions, and three alternate years for mobile
phones (2004, 2006, 2008). Each year‟s independent survey included 400 respondents for televisions
and 550 respondents for mobile phones.
Respondents were asked to rank the criteria they used to choose their brands including: Price,
Brand name, Quality, Service, and Design, and if they intended to buy their current brand on their next
purchase. 68% of respondents in 2004 had bought at least two TVs with an average of six years
between purchases; by 2008 80% had bought at least two with an interval of 4 years. Similarly, the
proportion of respondents who had bought at least two mobile phones rose from 76% in 2004 to 94%
in 2008 with average purchase interval declining from 18 to 10 months. Brand shares from the data in
both categories were verified against published share data (e.g. Euromonitor, Chinatimes) and were
found to agree +/-5 percentage points.
Individual consumer‟s purchase combinations were organized into switching tables (cross-
tabulating previous brand purchased against current brand) that showed the percentage of buyers who
bought the same brand on both purchase occasions (loyalty) and the percentage who bought different
brands (switching). This two-purchase technique reliably generates brand performance measures
(BPMs) such as duplication of purchase (Roy & Lahiri, 2004), brand buying loyalty (Bennett,
Ehrenberg & Goodhardt, 2000), and market structure, or partitioning (Dall‟Olmo Riley 2000,
Ehrenberg, Uncles and Goodhardt, 2004).
Early results in 2004 and 2005 revealed very low brand loyalty and high switching (see
Uncles, Kwok & Huang 2005). However, by 2008 loyalty rates rose to 35% from 16% for televisions,
and to 29% from 23% for mobile phones (Table 1). Because most of the activity was switching, this
will be discussed first. Loyalty will be discussed later in section 8.
Table 1. Repeat buying rates increase
% buying the same brand 2004 2005 2006 2007 2008
Television Sets 16 29 34 33 35
Mobile Phones 23 23 29
5. Switching between brands is in line with market share
On making a second purchase in a category a consumer either buys the same brand again and is brand
loyal or switches to a new brand. This division between loyalty and switching is predictable at an
aggregate level using the Duplication of Purchase Law (DoP) (Ehrenberg & Uncles 2000). The DoP is
an algebraic simplification of part of a larger model, the Dirichlet, a theoretical model for brand choice
that has repeatedly shown that simple parameters such as penetration and purchase frequency can
accurately predict brand performance measures in established FMCG markets (Ehrenberg, Uncles &
Goodhardt, 2004), product variants (Singh, Ehrenberg and Goodhardt, 2004), fast food (Bennett
2004), electrical goods (Hsu, 2002), subscription markets (Sharp, Wright & Goodhardt, 2002), cars
(Ehrenberg & Bound, 1999, Terech 2004), and developing markets, including China (Uncles, Kwok &
In examining multiple sets of data from many product categories, Ehrenberg and Goodhardt
(1970) found that the proportion of people who buy both brand X and Y (buyers shared by the brands)
is proportional to the number of customers who buy each brand—the bigger the brand, the more it
shares customers with other brands. For example, McDonalds is huge and nearly all fast food buyers
have bought it, but Speedy Noodle is small and many fast food buyers will not even have heard of it.
The proportion of Speedy Noodle customers who have also bought McDonalds will therefore be high,
conversely; the portion of McDonald‟s customers who have also bought Speedy Noodle will be very
small. The connection between sharing and brand size is supported by many others, e.g. Frank Bass
(1974) and Kalwani & Morrison (1977). The duplication law formula is written as:
Bxy = Dbx where
Bxy = the proportion of buyers who buy both hypothetical brands X and Y (i.e. the penetration of X
bx = the proportion of the market who buy brand X (the penetration of X) and
Dbx = the average level of switching for the category (D) multiplied by the penetration of brand X
The proportion of brand Y customers who will also be customers of brand X equals the average
amount of switching in the category (the D coefficient) multiplied by the penetration of brand X. Two
methods exist for the detection of the DoP law. The first involves observing the average level of
duplication between brands (see Colombo, Ehrenberg and Sabavala 2000) when the brands are
ordered by size in a purchase duplication as in table 2. (Subsequent tables are also ordered by share).
Global (non-Chinese) and Others (Chinese) are amalgams of smaller brands of less than 3% share.
The table presents switching between brands of televisions in China in 2007 (the same basic pattern
occurred in all years). Note that the diagonal showing repeat buying has been removed here for clarity.
Table 2 can be read in two ways; reading across by row, it shows that of those who previously
bought a TCL television, 15% bought a Changhong TV on their most recent purchase, 14% Sony,
11% Samsung and so on. Note that duplication generally declines across each row in line with the size
of the brands
Reading down the columns shows TCL was bought by 18% of Changhong‟s previous
customers, 6% of Sony‟s, etc. The figures within columns are fairly consistent because the competitive
attraction of a brand can be understood by its duplication with other brands. On average TCL attracts
13% of every other brand‟s customers, Changhong 12% on down to Other at 2%. The proportion of
customers each brand attracts, or average duplication, declines with share. This shows that the DoP
law holds here.
Table 2. Duplication Values for each pair of brands, TVs (2007)
Previous TV Current TV
% Chang- Sam- Sky-
TV Brand buying TCL hong Sony sung Hisense Konka Global worth Other
TCL 16 15 14 11 13 7 5 7 2
Changhong 15 18 11 11 16 7 7 9 2
Sony 13 6 3 17 7 10 7 3 3
Samsung 13 7 10 23 12 0 8 0 0
Hisense 12 15 18 9 6 12 3 9 6
Konka 10 17 14 11 10 10 8 6 2
Global 9 8 4 19 16 4 0 8 0
Skyworth 8 19 15 9 9 13 6 6 6
Other 4 13 17 13 4 13 4 0 9
Ave Duplication 13 12 14 11 10 6 5 5 2
There are also some deviations, e.g. average duplication for TCL is 13% but it attracts 19% of
Skyworth‟s customers. Likewise Sony attracts an average 14% of other brands‟ customers but 23% of
Samsung‟s. But even with these deviations the MAD (Mean absolute deviation) for the entire category
is about 4, giving a good overall fit of the DoP law (Ehrenberg 1972). The exceptions to the general
pattern suggest the market is partitioned as will be explained below.
The second method used with the DoP involves applying a formula based on the overall level
of sharing in the market and the size (share) of the different brands to generate expected levels of
sharing. Observed sharing can then be compared to theoretical estimates by using the DoP constant D.
The main pattern was that actual switching was close to projected levels, with slight deviations from
predictions, reinforcing the impression of structural change seen above in the penetration figures.
Table 3. Duplication Values for each pair of brands, Mobile Phones (2006)
Previous Mobile Current Mobile
% Moto- Sam- Sony/
TV Brand buying Nokia rola sung Bird Erics LG TCL Other Konka Amoi
Nokia 25 15 21 7 12 8 9 2 3 7
Motorola 19 32 24 16 9 5 7 1 9 2
Samsung 18 17 15 9 8 7 0 0 3 0
Bird 10 24 4 31 22 12 11 4 0 3
Sony/Ericsson 9 15 5 14 2 0 2 2 3 4
LG 6 19 5 26 6 3 6 5 6 7
TCL 6 8 13 19 15 0 4 0 8 0
Other 6 19 7 17 0 16 9 0 3 3
Konka 5 31 14 9 6 8 5 6 3 3
Amoi 3 11 17 9 2 13 13 3 0 0
Ave Duplication 20 11 19 7 9 7 5 2 3 3
The same was true for Mobile phones as shown in Table 3. Overall average switching was 9%, with
larger brands to the left having higher switching than smaller ones. Nokia and Samsung gained share
and had notably higher switching while Bird and smaller brands which lost share were lower.
6. Application of theory to Chinese television market data
The DoP law states that a brand shares its customers with other brands in its market in line with the
market share (penetration) of those other brands. This relationship of switching to penetration is
expressed in the duplication coefficient D, essentially a measure of how much buyers overlap
(duplicate) purchases of different brands. D reflects the likelihood of switching from a previous brand
P to a newly-bought brand N:
D= Ave. % of owners of P who switch from P to N
Ave. % who buy N at all
The overall average switching from one brand to another was 8.6% in Table 2, while average market
penetration was 11.1%. Dividing average switching by average penetration, 8.6/11.1 = .77, which is
D, the duplication or switching coefficient for TVs in 2007.
D is a constant for the average sharing of buyers between brands in the category. To generate
theoretical duplications for individual brands, D is multiplied by each brand‟s penetration as shown in
the bottom line of table 4. For example, for TCL, D x penetration is .77 x 16% = 12%, which is close
to the observed value. If D x penetration entirely explains the sharing of customers, then the market is
seen to be unsegmented (i.e. the size of the brands is the only determinant of sharing) (Ehrenberg and
Uncles, 2000). There was a very good fit between actual and theoretical duplication (r = .92), giving
the second confirmation of the DoP model.
Table 4. Theoretical Duplications (2007)
Chang- Sam- Sky-
TV Brand TCL hong Sony sung Hisense Konka Global worth Other Ave.
% buying 16 15 13 13 12 10 9 8 4 11.1
13 12 14 11 10 6 5 5 2 8.6
12 12 10 10 9 8 7 6 3 8.6
Table 5 shows average duplication rates for TV brands over five years. Note that average switching in
the final column declines each year. This is because the repeat purchase rate rose from year to year,
driving down switching and therefore duplication. Within the table the consistent main pattern is that
larger brands on the left had much larger duplications than smaller brands to the right: TCL usually
had four or five times the duplication of Others. So television buyers were four or five times more
likely to have TCL as one of their two purchases. The only variable considered here is brand size—
positioning, price, etc. are ignored because each brand‟s switching is clearly in line with its size.
Table 5, Average duplication or switching for TV brands over five years
Brands TCL Changhong Sony Samsung Hisense Konka Global Skyworth Others Ave
2004 19 15 5 8 7 19 11 11 6 11
2005 18 14 5 12 6 17 9 12 5 10
2006 16 15 9 8 5 13 7 8 5 9
2007 13 12 13 11 10 6 5 5 2 9
2008 10 11 12 11 11 6 6 6 2 8
Another pattern to note is the declining switching levels for TCL, Changhong and Konka. These
brands all lost market share, largely because they attracted too few buyers from other brands. This
showed up as falling switching levels. Sony, Samsung and Hisense on the other hand grew and this
can be seen in their increased switching levels (see chart 1).
In five years the average D-value for all brands fell from .87 in 2004, to .78 in 2005, to .76 in
2006, .77 in 2007 and .70 in 2008. (D is a measure of switching, thus as loyalty rises, D falls). By and
large for all years switching decreased in line with D times each brand‟s penetration, from high for
TCL on the left to low for Others on the right. (The correlation was r = 0.97 for 2004, r = 0.96 for
2005, r = .99 for 2006, r = .92 for 2007 and r = .92 for 2008).
The same patterns occurred in mobile phones (Table 6), where switching was fairly constant
over the years, declining slowly as the big brands consolidated their hold on the market. Note that fast-
rising Nokia and Samsung consistently attracted more switchers than other brands. Indeed that was
how they grew. In addition Duplication levels fell from D=.85 in 2004 to .81 in 2006 and .76 in 2008
and accurately predicted switching levels (r = .88 in 2004, .93 in 2006 and .92 in 2008).
Table 6, Average duplication or switching for Mobile phone brands
Brands Nokia Moto Samsung Bird Sony/E LG TCL Others Konka Amoi Ave
2004 23 14 21 10 8 7 4 3 3 2 10
2006 20 11 19 7 9 7 5 2 3 3 9
2008 19 9 21 5 6 6 4 3 2 4 8
Thus on average, switching levels for brands were closely predicted by the duplication of purchase
law. This is a good indication that the Chinese markets for televisions and mobile phones have
familiar structures based on well-established patterns of brand buying behavior.
7. Partitioning between competitive brands
Duplication was examined on a brand-by-brand basis, making it possible to note sub-patterns of
market partitioning seen in brand performance measures (BPMs). Both categories had slight
partitioning between big global brands and Chinese brands, probably reflecting price differences
(Ehrenberg, Uncles and Goodhardt, 2004).
Table 7. Duplication Coefficients D for TV brands, 2008
Previous Brand Current Brand
% Chang- Sam- Sky-
TV Brand buying TCL hong Sony sung Hisense Konka Global worth Other
TCL 16 1.0 .7 1.1 1.1 .5 .6 .6 .2
Changhong 15 .7 .5 1.1 1.3 .5 .7 .9 .3
Sony 13 .2 .2 1.7 .6 .8 .8 .4 .6
Samsung 13 .3 .9 2.1 1.2 0 1.0 0 0
Hisense 12 .9 1.0 .7 .6 .9 .3 .9 0
Konka 10 .6 .9 .6 1.0 .8 .9 .6 .3
Global 9 .5 .3 1.2 1.6 .3 0 .8 0
Skyworth 8 1.0 .6 .5 .9 1.1 .5 .7 .9
Other 4 .8 1.2 1.0 .4 1.1 .3 0 .9
Average D .6 .8 .9 1.1 1.0 .4 .6 .6 .3
Overall average duplications in observed switching in Table 2 and 3 (and in other years not shown)
were in line with those predicted by each brand‟s penetration. This is essentially the standard DoP
Law pattern noted above. Against this, the purchase duplications between pairs of brands may reflect
clusters or submarkets within a marketplace; clusters here were derived from the differing likelihoods
of switching between pairs of brands, (assuming a high degree of substitutability), as expressed by
their D-values (Table 7).
Some deviations from average are picked out in bold, e.g., in the second column TCL and
Skyworth have a D value of 1.0 that is out of line with the rest of the column. This deviation was
noted in Table 2 where duplications were expressed as percentages. In the middle of the table other
higher Ds show brands that share more customers than average but their meaning does not become
clear until they are grouped by clusters as shown in Table 8.
Table 8, Duplication-Coefficients show partitioning and change (2008)
Small Chinese Big Chinese brands Global
Small Chinese brands .6 .9 .7
Big Chinese brands .6 1.0 .7
Global brands .3 .5 1.4
At the bottom right the Global cluster had a duplication of 1.4, twice the average, hinting that some
TV buyers prefer foreign brands at the high end of the market. Similarly big Chinese brands were
more likely to share customers with each other than with small Chinese or global brands. This was
seen in all years of the study.
Partitioning is rare and is usually based on a functional difference between product variants
such as regular or lead-free petrol (Ehrenberg, Uncles & Goodhardt, 2004). Here it may reflect the
higher price of Global brands. Another feature of Table 8 is that duplication of Big Chinese Brands in
the middle column is above average with small Chinese (.9) showing a gain of customers from them,
but below average for Global brands. Also, the low figures in the first column and higher figures in the
third show that small Chinese brands attract the least switching (losing share, see chart 1), while
global brands grow by attracting switchers.
As seen earlier in Table 6 mobile phones brand duplications were above average between the
big three brands (D = .9 in 2004, 1.1 in 2006 and 1.2 in 2008) showing they shared customers more
than with the rest of the market where D = .6 for all three data sets, showing a loss of customers to the
bigger brands consistent with previous findings.
8. Brand Loyalty for TV brands in China is rising, and becoming more regular
Table 9 shows the percentage of TV buyers who bought the same brand on their two most recent
purchases. These repeat buying rates are low: 40 to 50% is normal for other consumer durables,
(Ehrenberg & Bound 1999, Terech 2004). In addition, they are more variable compared to previous
work, (Bennett 2007) and this is partly due to the small data sets that induce some statistical variation,
especially for smaller brands, and partly due to market dynamics mostly changes in market share
(Kato and Honjo, 2006).
Overall, purchase-to-purchase repeat rates more than doubled from 16% to 35% and became
more uniform (standard deviation fell from 11.1 to 7.2 over the five years). Bigger brands tended to
have more stable brand loyalty rates (Dekimpe et al, 1997) and repeat buying was a bit higher for
foreign brands than most Chinese brands despite them being smaller; a similar effect is observed for
luxury car makes (Ehrenberg & Bound 1999) as a result of price-based market partitioning. Over the
years these changes indicate that the market is both becoming more stable, and more dominated by the
Table 9, Percentage of TV buyers who bought the same brand on 2 purchase occasions
Brands TCL Changhong Sony Samsung Hisense Konka Global Skyworth Others Ave
% Repeat Rate
2004 29 9 35 18 0 9 21 6 15 16
2005 43 24 35 21 23 24 36 25 15 29
2006 39 30 43 35 24 30 40 29 18 34
2007 32 31 41 41 31 31 42 26 21 33
2008 31 32 46 41 31 32 41 27 24 35
The same held true for mobiles where repeat buying rose from 23% to 29% as shown in Table 10.
Again, while slightly lower than initially expected, these figures compare to 25% repeat buying for
mobile phones in Taiwan (Hsu 2002). And as in TVs, the figures for brands are becoming a bit more
uniform, with standard deviation falling from 9.7 in 2004 to 9.3 in 2008.
Table 10, Percentage of Mobile phone buyers remaining loyal
Brands Nokia Moto Samsung Bird Sony/E LG TCL Others Konka Amoi Ave
2004 26 34 25 17 14 9 8 16 3 9 23
2006 26 28 33 17 9 13 9 23 7 7 23
2008 36 28 34 17 14 13 9 17 15 16 29
Both mobile phones and TVs showed a double jeopardy (DJ) pattern, (Ehrenberg and Uncles, 2004)
which says that smaller brands are punished twice—they have fewer people who buy them and those
people buy them less loyally. This was seen in Tables 9 and 10 where bigger brands have higher
repeat buying rates. DJ also suggests that bigger brands attract more switching (more customers) as
was seen earlier in the switching section. In part this is a statistical selection effect in that bigger
brands have more opportunity to be bought than smaller ones.
9. Attitudes to Brands closely match behavior
Having observed behavior it was then possible to examine the attitudes that drove it (see also Kwok,
Uncles and Huang, 2006). Respondents ranked from 1 to 5 the criteria they used in purchasing (Price,
Brand name, Quality, Service, and Design) with 1 being highest priority and 5 lowest (“other” was
rarely used). For all surveys, Quality was rated highest. Table 11 shows the rankings of purchase
criteria for current purchases of televisions in 2004 and 2008. In 2004 Price was the third ranked
criterion, only slightly less important rated than brand.
By 2008 buyers had reordered their criteria. Quality and Brand are still top but Design
displaced Price as third most important. Brand and Design also raised in score from 3.1 to 2.8 and 3.7
to 2.9 respectively. This suggests that consumers became more brand loyal (explored below) and more
knowledgeable, i.e. evaluation criteria shifted as buyers gained experience.
Respondents were asked whether they intended to buy the same brand on their next purchase
as they had on their most recent purchase and about half said “no.” The remainder split evenly
between “yes” and “maybe.” This result was remarkably consistent across all data sets. The extent to
which stated intention is matched by actual purchase behavior has been extensively examined (e.g.
Morwitz 1997, Baldinger & Rubinson 1996, Morrison & Schmittlein 1988, and many others) and the
general conclusion is that intention is only an approximate predictor of behavior. This is especially
true in fast-changing technology or fashion-driven categories.
Table 11, Rank order of selection criteria for current TV purchase, 2004, 2008
Criterion Ave Score Criterion Ave Score
Quality 1.7 Quality 1.9
Brand 3.1 Brand 2.8
Price 3.2 Design 2.9
Design 3.7 Price 3.7
Service 3.8 Service 3.9
Intent to repurchase is an indicator of satisfaction. The low figure here may reflect dissatisfaction or
the knowledge that a brand is less widely available (brands leaving the market) or that newer brands
were more desirable, for whatever reason. It also probably shows a market in transition—buyers learn
to anticipate the arrival of new brands, technologies and retailers. The fact that low repurchase
intentions were matched by low repurchase behavior gives a measure of internal consistency to the
Mobile phone buyers also altered their choice criteria (see Table 12). Again, quality was most
important, followed by design, with price and brand equally important. Price may be less important
than in TVs because mobiles cost less. Low price may also be synonymous with low quality—„Black‟
phones—remanufactured older models or counterfeits are the cheapest and are notoriously unreliable,
proving that you get what you pay for.
Table 12, Rank order of selection criteria for current Mobile phones, 2004 and 2008
Criterion Ave Score Criterion Ave Score
Quality 1.8 Quality 1.8
Design 2.7 Design 1.9
Price 2.8 Brand 2.1
Brand 2.8 Price 3.3
Service 4.6 Service 4.1
In 2008 design and brand were much more highly valued, rising from 2.7 to 1.9 and 2.8 to 2.1
respectively. This may be because many buyers keep 2 or 3 phones, e.g. one for family, and one for
email, and the design criterion may reflect different functions. But it may also reflect fashion and
implies that buyers value how a phone looks or its „cool‟ factor. Note also that price slipped further in
This finding may also illustrate a generational shift in consumer purchasing in which younger
consumers roughly equivalent to generation X or Y in the US behave more individualistically. The
younger generation in China is media savvy, aware of Western products and brands and does not
necessarily buy into the collectivist values of their parents‟ generation. It also probably reflects the
family or child centered notion of Chinese consumerism (Zhao, 1997)
In any case, Chinese consumers increasingly seem to want more than just basic goods. This is
one reason why Nokia, which emphasizes fashion as well as function, has surged ahead of Motorola
and Sony/Ericsson. It is now being seriously challenged by Samsung which is also riding a fashion
wave and finding that the Chinese, like the rest of the world are willing to pay to be seen to have the
„right‟ or fashionable product or brand.
Brand buying in China showed the dominance of market share in determining brand performance
measures. While switching was the norm, loyalty was rising as the categories matured and became
more like western ones. Low repeat buying was not due to the unique character of Chinese buyers but
to the dynamic conditions of the marketplace. In the case of TVs, the withdrawal of many brands
made it impossible for many buyers to remain loyal even had they wanted to.
Chinese consumers are becoming more experienced and sophisticated as they make repeated
category purchases. Rising loyalty most likely reflects greater knowledge of products and brands
combined with wealth effects. Rising loyalty therefore suggests that the buying of durable goods in
China is becoming routine or perhaps even habitual. This is emphasized in the shift in ratings of
evaluation criteria towards design and brand name, and away from price. Whether Chinese consumers
continue to become more brand loyal as they gain experience remains to be seen, but the trend is in
It was also clear that though the TV and mobile phone categories are maturing, they are still
relatively dynamic. As a result, BPMs will probably remain relatively unstable and brand loyalty is
likely to remain low for some time.
The Chinese are becoming sophisticated consumers. They buy the same types of products and
brands and face the same marketing (if not at quite the same level) as consumers in more developed
markets. Chinese consumers buy in recognizable patterns, and have attitudes towards purchasing
brands that mirror their behavior. In sum, they are not so different from consumers in the rest of the
world. Moreover they become less different all the time. Consumerism is global, manifest in much the
same way in different parts of the world, separated more by timing than by culture.
11. Key implications
In markets categorized by switching rather than brand loyalty, marketers should aim to grow through
attracting new customers rather than through customer loyalty. In fact, most brands in China can
expect to retain only a third of their customers. Growth then will have to come partly from customers
new to the market, but mostly from the customers of other brands. To appeal to these buyers,
marketers should capitalize on broad-based market-driving attributes.
The Chinese TV and mobile phone markets had slight clustering of brand buyers. This means
each brand competes more or less with the entire category rather than with a sub-set of similar brands.
For example, the Chinese TV brand Hisense had above average duplication of .9 with its fellow brand
Konka, and below average .6 with Samsung. This means it can expect .9 x 10% (Konka‟s share) = 9%
of it‟s switching customers to come from Konka and .6 x 13% = 8% from Samsung. So even though
Hisense shares customers more with Konka and might be tempted to target those customers, in reality
it should expect to gain nearly as many customers from Samsung, even though it is a less similar
brand. This implies that Hisense (and other brands) should employ marketing strategies that appeal to
all category customers.
That said, growth is more likely to come from attracting the existing customers of bigger
brands simply because there are more of them. In other words, a general marketing strategy that
reaches bigger customer groups will be more successful than one that only reaches the customers of
smaller brands, even if the small brand has a high duplication level.
Marketers need also to examine the criteria that buyers use to select brands. In China price is
still important, but less so than it used to be. Rather, more sophisticated, wealthier buyers consider it
after they have considered other attributes such as design and brand name. Competing on price is
therefore likely to be a less fruitful strategy.
Baldinger, AL & Rubinson, J. (1996), “Brand Loyalty: the Link Between Attitude and Behaviour”
Journal of Advertising Research, 36(6), 22-3
Bass, FM. 1974. The theory of stochastic preference and brand switching. Journal of Marketing
Research Vol. 11, p 1-20
Bennett, DR. 2004. The Taiwanese are Just Like Australians in their Loyalty to Fast Food Outlets.
Australasian Marketing Journal Vol 12, No. 3 p97-103
Bennett, DR. 2007. Meta analysis of repeat buying loyalty. Proceedings of the Academy of Marketing,
UK conference, Kingston University, UK, July 3-6
Chen Y & Penhirin, J. 2004. Marketing to China‟s Consumers. The McKinsey Quarterly, Aug
China.org.cn, http://www.china.org.cn/english/business/243075.htm 2008-02-18
Colombo, R, ASC Ehrenberg and D Sabavala. 2000. Diversity in Analyzing Brand Switching Tables:
The Car Challenge. Canadian Journal of Marketing Research, 19 pp23-26
Cui, G. 1999. Segmenting China‟s consumer market: a hybrid approach. Journal of International
Consumer Marketing Vol. 11 No.1 p55-76
Dall‟Olmo Riley, F. 2000. Patterns of Reported Behaviour in Survey Research. Academy of
Marketing Conference, Conference Proceedings, Derby, UK, July
Dekimpe, MG, Steenkamp, J-B, Mellens, M & Abeele, P. 1997. Decline and Variability in Brand
Loyalty. International Journal of Research in Marketing, Vol 14, Issue 5, p405-420
Ehrenberg, ASC. 1972. Repeat Buying, North Holland Publishing Company, London
Ehrenberg, ASC & Bound, J. 1999. Customer Retention and Switching in the Car Market. Research
Report 6, The R&D Initiative, London South Bank University
Ehrenberg ASC & Goodhardt, G. 1970. A Model of multi-brand buying. Journal of Marketing
Research. Vol. 7 p77-84
Ehrenberg, ASC & Uncles, MD. 2000. Understanding Dirichlet-type Markets. The R&D Initiative
Research Report 1, London South Bank University, London
Ehrenberg, ASC, Uncles, MD & Goodhardt, G. 2004. Understanding brand performance measures:
using Dirichlet benchmarks. Journal of Business Research, 57 (12), 1307-1325
Euromonitor International Estimates, Reports -- China, Televisions and projectors, 2008A
Euromonitor International Estimates, Reports -- China, Mobile Phones, 2008B
The Gallup Organization (2006) Chinese Consumer Survey, Reprinted through the Harvard Business
Hsu, K-C. 2002. Analysis of Loyalty and Switching in Taiwan. Master‟s Degree thesis, London South
Kalwani, MU & Morrison, DG. 1977. A Parsimonious Description of the Hendry System.
Management Science, 23, (5, January): 467-477
Kato, M & Honjo, Y (2006) “Market Share Instability and the Dynamics of Competition: a Panel Data
Analysis of Japanese Manufacturing Industries,” Review of Industrial Organization, March, vol. 28,
Issue 2, p165-182
Kwok, S, Uncles, M & Huang Y. 2006. Brand preferences and brand choices among urban Chinese
consumers: An investigation of country-of-origin effects. Asia Pacific Journal of Marketing and
Logistics, Vol. 18, Issue 3 p163-172
Lu, JR & Peng, A. 2000. Evolution of rural consumption pattern in China. Consumer Interests Annual,
Vol. 46, p222-225
Morrison, D. G. & Schmittlein, D.C. (1988) “Generalizing the NBD Model for Customer Purchases:
What are the Implications and is it Worth the Effort?” Journal of Business And Economic Statistics, 6
(April), p 145-159
Morwitz, V. (1997) “Why consumers don‟t always accurately predict their own future behavior.”
Marketing Letters Vol 8 Number 1, January
Roy, D & Lahiri, I. 2004. Some tests for suitability of brand switching model. European Journal of
Marketing Vol. 38 No. 5/6, p 524-536
Sharp, B, Wright, M, and Goodhardt, GJ. 2002. Purchase Loyalty is Polarised into either
Repertoire or Subscription Patterns. Australasian Marketing Journal 10 (3), 7-19.
Singh, J, Ehrenberg, ASC & Goodhardt, GJ. 2004. Loyalty to Product Variants-A Pilot Study. Journal
of Customer Behaviour, Vol. 3 No. 2 July, p 123-132
Terech, A. 2004. Three Essays on Consideration Sets. Ph.D. Dissertation, Anderson School of Business,
University of California, Los Angeles
Uncles, MD, Kwok, S & Huang, S. 2005. Modeling Retail Performance using Consumer Panel Data:
a Shanghai Case study. Working paper at the School of Marketing, University of New South Wales,
Zhao, B. (1997), Consumerism, Confucianism, Communism: Making Sense of China Today, The New
Left Review, I/222, March/April
Zheng, Z & Cui, C. (2007) Chinese Consumers‟ Comparison Behaviour: Exploring some Critical
Issues towards a Theoretical Framework, Proceedings of the Academy of Marketing (UK)
Conference, Kingston University, July