The Role of Memory in How Brands Grow

The way we approach a problem reveals something about what we believe.  So, as a market researcher, if I ask a question about whether my respondent intends to buy a pair of shoes then I have a belief in the personal agency of humans.  As a behavioural scientist if I set up an experiment to examine the contextual influences on shoe buying then in all likelihood I have a belief that context is important in human decision making and am less likely to consider personal agency is as important.

Sometimes these assumptions are clear and easily recognised but at other times they are less apparent.  All disciplines include philosophic assumptions that don’t stop being influential just because they have been forgotten.  As philosopher Mary Midgeley puts it, they lie under the floorboards of all disciplines.

So, just what are the assumptions that underpin the way in which brands are often researched and understood and specifically the principles outlined by the Ehrenburg-Bass Institute for Marketing with Byron Sharp’s analysis of ‘How Brand’s Grow’.

This hugely successful book has been widely adopted by brands many of which have fundamentally reshaped their marketing strategy around Sharp’s conclusions.  His work is based on an impressive analysis of data relating to consumer purchasing behaviours.  We shall first summarise the key principles before going on to explore the implicit assumptions that may underpin the analysis.

How brands grow – key principles

Sharp writes about the mental availability of a brand which is a function of the brand’s salience.  He considers a brand’s mental availability as the probability of a consumer noticing, recognizing, and thinking of your brand in a buying situation. Importantly, this is different from brand awareness, which is simply the link to the name of the product category and depends on a single, specific cue. Mental availability extends beyond brand awareness: it depends on the quality and quantity of a consumer’s mental structures.

Theories shape analysis

Sitting underneath this understanding of the way in which brand associations are formed must clearly sit a theory of memory.  If we are acting on associations that have been created in the past, then memory must have a role.

The challenge is that there are a great many theories of memory, although none appear to be explicitly referenced in Sharp’s analysis.  But just because a theory of memory has not been made, it is very clear that nevertheless, if we prise up the floorboards we will see that assumptions have been made that will reveal the theory that has informed the analysis (consciously or otherwise).

This is important because although we prefer to think of our analysis as value-free (where the data ‘speaks for itself’) these, often implicitly held theories, inevitably drive our analysis.  As every market researcher and academic researcher knows, data analysis is a messy subjective business.

Memory as information processing

It does not seem to be too much of a leap to consider that the implicit assumptions of memory underpinning Sharp’s analysis seem very consistent with Anderson and Bower’s (1973a) Associative Network Theory (ANT) of memory.  This framework proposes that memory is structurally organised as a network of nodes linked together whereby:

  • Nodes represent concepts
  • Links represent relationships

This class of memory theories (and there are many) represents memory as information processing.  We are operating in a way that encodes certain cues and associations albeit some are stronger than others.  These are then accessed when certain cues are present in the environment that taps into this ‘memory bank’.

This theory is widely used to understand how marketing works.  Keller’s (1993; 2003) Customer-Based Brand Equity (CBBE), for example, suggests that brands are stored in memory as focal concepts and the range of perceptions relating to the brand are memorized as concepts linked to the brand in a network of brand-related information referred to as brand (attribute) associations.

Romaniuk (2013) uses spontaneous associations as a measure of a brand’s propensity to be retrieved from long-term memory, or brand salience.  The count of associations is the brand salience metric and represents of memory structures.  Retrieval propensities are considered to depend on the size and the strength of a brand’s memory network and the relative size and strength of competing brands.

So, it is fairly clear that Sharp is (implicitly) referencing an information processing model of memory.  Whilst this is not unreasonable it is nevertheless important to determine the extent to which alternative theories of memory may also offer an explanation of how brands work.

The trouble with Information Processing models

There are some troubling aspects of these models which deserve closer inspection. First, they are not always very good at explaining some aspects of memory.  So, for example, the well-known phenomenon of false memories is hard to explain using this framework (Brainerd, Yang, Reyna, Howe, & Mills 2008).   In another instance, these frameworks struggle to explain how associations transfer to novel situations (e.g., Reyna & Brainerd, 1992).  Of course, some post hoc explanation using associations be constructed to account for these but by that measure, all theories can account for all results.

The broader problem that must surely trouble marketers is that associations which are at the heart of information processing models can create inherently meaningless relationships.  This is because these models suggest a very passive model of memory – not only in terms of storage but also recall.  That is not to say that it is necessarily wrong but can we really explain all brand-related behaviour in terms of simple associations that people have?  Or do we bring something a little more complex to the way we think about brands?

Constructivist models of memory

Although information processing theories of memory continue to dominate much of the academic literature, we are starting to see the emergence of a new class of memory theories which can broadly be characterised as ‘constructivist’.  These are more dynamic models of memory where our experiences are actively engaged both in the creation but also the recall of the memory.  These suggest that memory is more than ‘memorization’.

In these instances, experiences are stored as a function of a number of different attributes (Leboe-McGowan and Whittlesea, 2013) namely:

  • Stimulus (focus on processing during encoding)
  • Task (purpose one is engaged in while processing the stimulus)
  • Context (environment in which the processing occurs)

It is a recognition of the importance of context in memory creation that suggests a much more dynamic model of memory creation – with a much greater opportunity for a rich and varied type of memory.

Our mental processing of the memory also differs from the information processing model as the means with which we recall memories are also much more active in terms of:

  • Production (the response that arises from the connection that occurs between the current environment and what is stored in memory)
  • Evaluation (reflective process that occurs about the type and quality of the production process)

Hence, in this class of theory, memory is much more than memorization – we are much more active participants and as such the contexts in which memories are generated and recalled has much greater significance.

These constructivist models suggest that the consistency between the way that memory is created and the recall environment typically linked to our ability to recall relevant information (Reber, Schwarz and Winkielman, 2004; Whittlesea and Price, 2001).  In other words, we are active participants in the way that we both create and recall memories – as such our context, meanings and interpretations have a much greater role to play.

What does this mean for Byron Sharp?

Any explanation involves a trade-off between adequacy and parsimony; that is, theories are judged to be more successful if they explain larger numbers of findings with the same assumptions or explain the same findings with smaller numbers of assumptions, or both.  Because of course, any theory can, post-hoc, explain any evidence.   The gold standard, of course, is to evaluate theories based on their ability to predict outcomes – but that is hard to achieve in the context of brand performance.

So, is an information processing model one that can sufficiently explain mental availability more important than how we think about brands?   Or are constructive theories of memory more effective?

Sharp may well argue that all mental availability is equal in importance.  This would suggest that there is no meaning attached other than recall – and basic brand associations.  So, it is mental availability that is critical here, not the meaning that is attached to it.  On this basis, the wider attitudes that people hold are not relevant and indeed this is supported by the Ehrenburg-Bass data which indicates very little if any, relationship between attitudes and consumer behaviour.

A constructivist model of memory would suggest that associations are important but perhaps are only one part of the picture.  There is more active engagement in the process of determining the degree to which different associations are relevant.   If the context in which associations are made is consistent with when they are retrieved than an information processing model appears to work well.    Hence, for stable categories this works well – it has good explanatory value.  However, if the context changes then we may start to see it not performing quite so efficiently.  As category (or even brand) context, meanings and interpretations change, then the situation for recall is different to the situation in which these brand associations (memories) were encoded.

For example, if category dynamics change then we can see the effect of this.  So, the context of soda drinks has changed and there is much more focus on sugar and unhealthy lifestyles – rather than sharing moments and living life to the full.  So, the associations that we created in the past are less longer relevant or helpful when making brand choices now.

A constructivist theory of memory is potentially better able to explain this change and provide an understanding of the way in which this directly impacts the sale of soda drinks.  An information processing model of memory works as long as the contexts remain static – in which case, a brand strategy based purely on salience will succeed.

It is starting to look as if constructivist approaches to memory could be helpful to help understand how brands grow in changing environments.

In summary

Sharp is creating a theory of consumer psychology based on his observations of the empirical data.  He considers that ‘mental availability’ or saliency is what drives brand choice (along with physical availability’).   The underlying theory of memory is unspoken but this implies an information processing model.

However, it is equally possible to consider constructivist theories of memory which place a much greater importance on the continuity of the context in which assoiciations were made and recalled.  These theories suggest that memories are more dynamic than we may previously have considered.

So why is this is important?  Well because theoretical models have significant implications for brand strategy.  An information processing model suggests that the most important activity for a brand is to generate associations – which implies mass TV advertising.  A convergence model does not directly contradict this approach but also suggests that within dynamic and changing environments then there is also a need to understand and reflect the changing cultural market context and constantly rethinking positioning and advertising.

By Colin Strong

My thanks to Dr Tamara Ansons for her input to this article


Anderson, John R. and Gordon H. Bower (1973a), Human Associative Memory. Washington: Winston & Sons.

Bartlett, Frederic C. (1932). Remembering. Cambridge: Cambridge University Press.

Craik, Fergus I. M. and Robert S. Lockhart (1972), Levels of Processing: A Framework for Memory Research, Journal of Verbal Learning and Verbal Behavior, 11(6), 671-684.

Keller, Kevin L. (1993), Conceptualizing, Measuring, and Managing Customer-Based Brand Equity, Journal of Marketing, 57 (1), 1-22.

Keller, Kevin L. (2003), Strategic Brand Management: Building, Measuring, and Managing Brand Equity, Upper Saddle River, N.J.: Prentice Hall.

Leboe-McGowan, Jason P. and Bruce W. Whittlesea (2013), Through the SCAPE Looking-Glass: Sources of Performance and Sources of Attribution, Handbook of Cognitive Psychology, Daniel Reisberg ed. New York: Oxford University Press.

Midgley, Mary (2006), Science and poetry. Routledge.

Morris, Donald C., John D. Bransford, and Jeffery J. Franks, (1977), Levels of Processing Versus Transfer Appropriate Processing, Journal of Verbal Learning and Verbal Behavior, 16(5), 519-533.

Romaniuk, Jennifer (2013), Modeling Mental Market Share, Journal of Business Research, 66(2), 188-195.

Sharp, Byron (2010), How Brands Grow: What Marketers Don’t Know, Oxford University Press

Whittlesea, Bruce W. A. (1997), Production, Evaluation and Preservation of Experiences: Constructive Processing in Remembering and Performance Tasks, The Psychology of Learning and Motivation, Vol. 37, Douglas L. Medin ed., New York: Academic Press, 211-264.