How Brands Grow

by Byron Sharp

  Chapter 7:

  Passionate Customer Commitment

  Byron Sharp

  Are buyers puppets?

  Nike is routinely held up as a super brand, but its sales and style are similar to Adidas. The sportswear market has a plethora of similar competing brands (Reebok, Puma, New Balance and Coq Sportif)40; one explanation for this brand proliferation is that marketers have used advertising and packaging to fool buyers into adopting irrational preferences and loyalties. This explanation has been put forward by both opponents and advocates of advertising and branding. There are plenty of psychological studies that support this argument, but they also suggest advertising and packaging have a very weak psychological effect (even under laboratory conditions).

  Two widely reported (and misreported) psychological studies are typical. Each purports to show that branding can massively distort preferences. In one study (McClure et al. 2004) 16 of the subjects undertook blind taste tests and about half chose Pepsi and half chose Coke. Their choice in the blind tests correlated poorly with the preference they stated prior to the test: about the same number of those who had stated they preferred Coke chose Pepsi in the blind trial, and about the same number of subjects who preferred Pepsi chose Coke.

  Other subjects were each given two samples of the same cola. One sample was identified as either Coke or Pepsi, while the brand of the other sample was not revealed – subjects were told that the second unidentified cola could be either Coke or Pepsi (though in reality it was always the same cola as the first sample).41 When the labelled cola was Coke, subjects largely said it tasted best. The same did not occur when Pepsi was the identified cola. This may have been due to the groups consisting of more people who usually tend to drink Coca-Cola (prior preferences were not checked, or at least not reported).

  Then, 67 subjects were given brain scans while tasting tiny computer administered squirts of cola dispensed via a plastic tube. An fMRI scan involves lying on a hard flat board inside an extremely loud (and, for some, claustrophobic, chamber). It is not a pleasant experience, and certainly different from most normal consumption situations! The main finding was that when subjects were told that Coke was the cola they were sipping more brain activity was recorded in the hippocampus and in other brain areas thought to be concerned with cultural knowledge and memories. This did not occur quite so much for Pepsi.

  What does this convoluted experiment tell us? First, in artificial tasting situations people are far more likely to trust their eyes than their sense of taste, this phenomenon has been observed across decades of taste testing research. They also may react as if it were a test – that their performance is being judged, not just the colas. So informing them that one cup contains their favourite cola (or non-favourite cola) largely overrides their taste buds – which is especially easy when the other cup contains the same cola (i.e. there is no difference between the two cups). Second, the brain scans probably reflect something that psychologists have known for decades: familiarity breeds liking. Usage also breeds familiarity and brand knowledge; this in turn breeds liking.

  This is very much what a second widely reported study shows. In this research (Robinson et al. 2007), 63 children aged between three and five years old were given two identical food items to taste. One item was in plain wrapping and the other was in McDonald's branded wrapping. The food items were McDonald's hamburgers, chicken McNuggets, McDonald's French fries, milk, and two baby carrots. With each exposure of a food item, for example, McDonald's French fries, the child was asked, “Can you tell me which of these foods is from McDonald's?” If the child did not answer or answered incorrectly the administrator pointed to the McDonald's branded option and said, “This food is from McDonald's”; this happened about a third of the time. Each child was then instructed to take a bite of each of the two options and the administrator then asked, “Tell me if they taste the same or point to the food [drink] that tastes best to you”. The children chose the branded option (as tasting best) 182 times. They chose the unbranded option 122 times or said they both tasted the same (or that they did not know).

  The more regularly a child ate at McDonald's (as reported by their parents), the more likely they were to choose the branded item. Also, children with more television sets in their home were more likely to select the branded item. But other moderator variables – having a television in the bedroom, having McDonald's toys in the house, and hours of television viewing – had no detectable influence.

  Decades of testing preferences and quality cues has taught us that when a single cue is introduced into a study respondents react strongly to it. This is probably because subjects feel they are supposed to react to the cue.

  Because it is the only cue, it becomes artificially salient. It is surprising that these sorts of artificial studies are still being done. Similar weak preferential results might probably be gained by simply wrapping one of the items in coloured paper and the other in plain.

  Therefore, we shouldn't read too much into these sorts of studies, which very often have a political agenda behind them – they invariably find results in the direction the researchers expected. What is worth noting though is how mild the effects are, even in these artificial constructions. The main finding again seems to be that prior usage of the brand makes a person slightly more favourable towards it. This is one of the psychological factors underpinning brand loyalty.

  Brand loyalty – a natural part of buying behaviour

  Brand loyalty is part of every market, even in so called commodity categories. Rather than thinking of loyalty as a market imperfection, it is more appropriately considered to be a sensible buyer strategy, one of many developed by human beings in order to balance risk and avoid wasting the precious commodity of time.

  In 1964 Professor Tucker of Texas University conducted an interesting experiment into the development of brand loyalty. His experiment revealed that buyers naturally tend towards brand loyalty – loyalty is a buying strategy that most of us adopt much of the time. In Tucker's experiment, each day, for 12 days, loaves of bread from the same commercial oven were offered on a tray for choice to 42 women. The loaves were wrapped identically, and each day one was labelled L, another M, P and H. The position of the brands on the tray was rotated. In spite of the loaves being identical, many of the respondents quickly adopted loyal behaviours, with each woman favouring a particular brand. Some also favoured positions on the tray (e.g. tending to choose the brand that was on the left of the tray). This landmark study concluded that it is clear that brand loyalty is based on what may seem to be trivial distinctions.

  However, Tucker's study was limited and somewhat artificial. It could be argued that buyers are used to more substantial differences in real-world brands and so acted as if there might be in this study. Also, tastes vary day by day, and maybe the bread varied too – if one day a loaf didn't taste so good to a subject perhaps she avoided that brand on the following days. Even so, the study did show that consumers have a predilection to adopt loyalty behaviours, for whatever reason. This is further shown by the fact that brand loyalty is a characteristic of every market.

  In Chapter 3 we saw that repeat-purchasing levels for car brands are quite high. That is, around 50% repeat levels (with some variation between countries); almost half the time new car buyers buy the same brand as they bought last time. This level of repeat-buying seems incredible considering that in modern market economies there are typically 60 or so brands to choose from. Also, usually a number of years pass between purchases, models change, as do buyer's life circumstances and needs. The odds of someone choosing the same car brand again appears to be very slim. If buying were random, there would be less than 1 chance in 50 of a person purchasing the same brand again. But buying is not random; there is clear behavioural preference and loyalty.

  We observe this sort of loyalty in all categories, including service categories and corporate buying. Buyers restrict their purchases to a personal repertoire of brands. People buy far fewer differ
ent brands than they could. Ten purchases seldom results in ten different brands being bought. Buyers keep returning to their favourites42, and this set of favourites can be very small. The 50% repeat-buying rate for car brands occurs because buyers consider only around two brands on average; one of which is usually the brand they bought last time (i.e. the brand they drive now) (Lambert-Pandraud 2005). About a fifth of buyers consider only a single brand, which is almost always the one they bought last (Lapersonne 1995). Yet this is what consumer behaviour textbooks refer to as a 'highly involving purchase decision'!

  Here's another surprising example. The choice of which television channel to watch is a category where loyalty is unexpected. Most of us think we watch television for the programs, and, aside from the genre-specific channels (e.g. the weather channel), we have only a vague idea of what channel offers our favourite programs. Also, there are no real switching costs: we can defect from one channel by pushing a button. In spite of all this, television channel choice shows the same pattern as other categories – people substantially restrict their repertoires and show loyalty.

  Figure 7.1 was compiled by Ehrenberg-Bass Institute researcher Dr Virginia Beal; it is based on Nielsen data on US households. The figure shows the number of television channels each household has access to in their subscription and how many they actually watch. The key point is how quickly the line flattens; when a household has access to 40 channels they watch about a dozen or so each week, households with access to 80 channels watch a few more than a dozen each week, and households with access to 200 channels still only watch a few more than a dozen each week.

  Figure 7.1: Number of television channels viewed per week in US households

  Nielsen, 2002

  In 2014 Nielsen released data showing how this pattern and the average number of channels viewed has held steady in spite of households gaining access to more channels. What this shows is that viewers (sensibly) keep returning to their favourite channels. There is obvious loyalty. Most viewers watch enough television to watch many more channels than they do, but they prefer to restrict their personal repertoire. They learn the channel numbers of their favourites and go there first. The chart looks almost identical if we examine brand buying in any category, for example, if it instead showed the number of purchases made by an individual and how many different brands they had bought (see Figure 7.2). Loyalty is everywhere, from soap to soap operas.

  Figure 7.2: Breakfast cereal purchasing, Australia; data courtesy of Nielsen.

  Divided loyalty

  Loyalty is everywhere but it’s seldom exclusive – buyers purchase more than one brand, and the more purchases an individual makes the more brands he or she buys. Polygamous or divided loyalty is quite the norm. So no brand should expect its buyers to be 100% loyal.

  Table 7.1 lists the proportion of a brand's buyers who were exclusively loyal to that brand (in a 12 month period). The average for this selection of brands is 13% of their buyer base. The double jeopardy law holds again for this loyalty metric: smaller brands have a smaller proportion of their customer base who are 100% loyal. All these brands are among the major brands in their respective categories – the metric is even lower for very small brands.

  Notice that categories that are purchased less frequently have higher levels of 100% loyalty. Analgesics are bought only five times a year on average, and because this is a skewed average (see Chapter 4), more than half of all households bought from the category only once. These once-only buyers, by definition, have to appear among the 100% loyal buyers – if you only buy once then you can buy only one brand. As time goes by these buyers make more purchases and the proportion that are 100% loyal drops dramatically. This can be seen in categories like breakfast cereals and yoghurts, which are bought more frequently and so show much lower levels of 100% brand loyalty.

  Table 7.1: 100% loyal buyers are the minority

  Product Annual category purchase rate (average) Market share (%) 100% loyal buyers among brand buyers (%)

  Analgesics 5.1

  Tesco 22 37

  Nurofen 8 28

  Boots 3 26

  Panadol 2 29

  Deodorants 5.6

  Lynx 17 21

  Dove 7 17

  Tesco 4 14

  Nivea 2 10

  Crisps (potato chips) 17.5

  Walkers 68 37

  KP 7 6

  Kettle Foods 1 10

  Breakfast cereals 21.5

  Kelloggs 29 7

  Cereal Partners 17 2

  Weet-Bix 9 2

  Yoghurts 29.7

  Muller 24 7

  Muller Light 14 4

  Ski 4 2

  Danone 3 2

  Average

  13 13

  Source: Kantar Worldpanel, UK, 2005.

  The standout in the table 7.1 is Walkers potato crisp brand: one-third of its buyers were exclusively loyal over the time period (one year). Yet this level of 100% loyalty is normal for a brand with such extraordinary market share.43 Walkers dominates the UK potato chip market, and has incredible availability; this makes it very easy for buyers to only buy Walkers, particularly if they don't make many purchases. The double jeopardy law again.

  Polygamous loyalty is seen in panels that record the buying of individuals – so it isn’t due simply to different members of the household having different loyalties (McDonald et al 2000, 2002). It is also seen in the industrial buying of individual companies (Uncles et al 1990, Pickford & Goodhardt 2000).

  Natural monopoly law

  Larger brands tend to have proportionately more light category buyers in their user bases. Light, occasional buyers favour the bigger brands. This isn’t that they particularly prefer larger brands, it is purely a statistical effect and is known as the 'Natural Monopoly Law' (i.e. larger brands have a monopoly on light buyers). For example, if you only buy one soft drink this year, then odds are it will be Coca-Cola. If you buy just one pack of crisps in the UK, then odds are it will be Walkers.

  Put another way, if we observe an English person eating Walkers crisps then we know little about their buying. They are simply part of the majority of the UK population who buy Walkers at some time or another. But if we observe someone eating a far smaller brand like Golden Wonder crisps then odds are that the person is higher than average buyer of salty snacks.

  The implication for marketing strategy is that to grow a brand needs to recruit more light category buyers. Targeting the heaviest buyers of the category (‘super consumers’) is a bit like saying “let’s be a small brand”.

  The natural monopoly law is illustrated with the data in Table 7.2. Heinz buyers purchase tomato sauce four times a year, whereas buyers of C&B sauce buy tomato sauce eight times a year, that is, twice as often44 (though C&B is only, on average, bought for 1.2 of these eight purchases).

  Table 7.2: Smaller brands are bought by heavier category buyers

  Tomato sauce (ketchup) brands Market share (%) Penetration (%) Frequency of buying this brand (units per annum) Frequency of buying any tomato sauce (units per annum)

  Heinz 53 50 2.9 4

  Daddies sauce 4 5 2 6

  C&B sauce 1 2 1.2 8

  Data source: Kantar Worldpanel

  Therefore, larger brands have more 100% loyal buyers partly because they have a lot of very light category buyers who haven't made many purchases during the analysis period.

  If we use a longer analysis period, the proportion of 100% loyal buyers always declines. Give people sufficient time to make multiple purchases and they will buy more brands.

  Loyalty certainly exists, but it is tempered by opportunity. People who buy from a category less often have less opportunity to be disloyal. Similarly, people who shop from stores that stock few brands also appear more loyal (e.g. people who live in smaller towns).

  Prosaic, not passionate, loyalty

  So this is the everyday story of loyalty that we gain from observing people's real buying behaviour over time.45 In the real world loyalty is everywhere, and
every brand has some, but loyalty is divided and strongly driven by opportunity. This is a more prosaic picture than most marketing textbooks and consultants paint.

  LOVEMARKS?

  Pity the poor marketing directors. It's not enough to merely build and maintain sales and deliver profits to stockholders; they are told they also have to get consumers to feel a deep, passionate commitment to their brand. It's no longer sufficient to have consumers who pay money to buy the brand; they have to also love the brand.

  This sounds like a very tough task, yet, perversely, many marketers embrace the idea, perhaps because merely selling things sounds crass. Many marketing texts talk about creating value, delivering customer satisfaction and building relationships. This makes the marketing profession appear so much more honourable.

  So it was a marketing masterstroke of Facebook to use the (ridiculously over-the-top) term ‘fan’ for any Facebook user who merely happened to once click the Like button on a post that came from the brand. Of course besotted marketers clamoured to build their Facebook ‘fan bases’.