Harvard Business Review says our elite research is irrelevant: why ‘soft science’ is the key to regaining leadership in marketing knowledge

In this paper the authors present an argument for academic marketing to shift away from a purely positivistic approach to its ‘top journal’ research and towards a critical realist approach that better accounts for the ontological reality of the discipline. We argue that this debate matters because, at present, ‘top quality’ research is, wrongly in our view, regarded as that which is based on marketing as a science. However we are not arguing for a radical shift into purely interpretive or social constructionist perspectives. We debate some different areas within the discipline of marketing and conclude that some areas may still respond well to scientific approaches, while others may benefit from a relaxation into interpretive approaches.

convinced the government that, sooner or later, an alternative hegemony needs to be found. The high profile of the Harvard Business Review article may accelerate this process. With HBR's intervention the Emperor's clothes are well and truly off. It is becoming increasingly embarrassing to blithely carry on producing research that is routinely ignored by all its stakeholders except one -fellow academics.
In this paper we examine the principles that we and others (see for example Weber 2004) believe represent the pragmatic mid-ground view of much good quality management and marketing research today. Taking a recent paper from the Journal of Marketing, we offer a deconstruction of the typical problems that arise from an uncompromising 'hard science' methodology. (These problems are not as a result of incompetence by the authors; in our opinion they illustrate the methodological flaws common across this mode of enquiry). Along the way, we try and point out how an approach that combines predictive approaches with interpretive methods may be better suited to this kind of research problem. We also suggest a 'horses for courses' mentality for academic marketing research such that different types of problems should be tackled in different ways, and our call is for an end to the false superiority of one type of approach. We argue that a more open minded attitude to research also needs to sit alongside an attitude of more open sharing of research within the wider marketing community, both in considering what subjects are important to research and in interpreting results.
Our first argument is that 'hard science' is often a 'fish out of water' when applied to marketing. To illustrate this it is helpful to look at the scientific approach where it has been most successful, that is, in the physical sciences.

Sodium and water -the successful application of 'hard science'
Let's take the example of how an element, sodium, reacts with water. We can divide up our description into observable facts and their underlying theories. Look at Figure 1 above. Physical scientists have no trouble distinguishing facts and theories. It is an observable fact that if you throw sodium into hot water it explodes, creating hydrogen and sodium hydroxide. Chemists have a settled view of why this happens -sodium's position in the periodic table makes it a 'group 1' element, with 11 electrons, and Valence Theory says a maximum of 2 followed by 8 electrons can orbit before the next electron must move into a new orbit. Hence one sodium electron is isolated in its own orbit, and is highly vulnerable. This vulnerability makes sodium highly reactive. The same is predicted of other elements with one isolated electron. The Periodic Table predicts that Potassium and Caesium will also be highly reactive, and observations demonstrate that this is so. Now, Valence Theory may not be the literal truth. It's just a theory. But it explains the known facts -and, at least in chemistry, the core facts don't change. Non changing factual evidence provides a solid platform for building theories, which are adjusted by subsequent data to better explain and predict. Further, in science theories do not exist in isolation but as part of a 'web of belief' (Williams, 2000). So to be credible any theory must exist as part of a network of interwoven meta theory, which is constantly scrutinised through replication of experiments. The chemistry 'web of belief' led to the periodic table and Valence Theory, which together elegantly and beautifully explain chemical behaviour. Indeed when brand new elements are created they react according to the predictions of the Periodic Table. What are the characteristics of chemistry like this that allows a 'hard science' approach to be successful? Epistemologically, reductionism -the drilling down of knowledge into smaller elements -makes intuitive sense in this field. Researchers can isolate the experiments, agree the constructs, control all the variables, and deploy hypothetico deductive method with complete confidence. Ontologically, facts are facts, and they don't change, nor are they disputed. There is an underlying order and pattern in chemical reactions that appears absolute. The underlying theories are supported by data and are highly predictive. These theories contain detailed explanations to help us understand why things happen. The theories may not be literally 'true' -Kuhn exposed the risk in absolute belief in science -but they work in an everyday way to help progress.

Scientific marketing -horses for courses?
Now let's compare chemistry and marketing: In figure 2 we suggest that marketing science can sometimes work, and sometimes not work well, depending on the nature of what is being researched. If we break the discipline into some of its constituents we may see that some behave more 'scientifically' than others. Examples might include studies of macro social trends and their impact on aggregate shopping behaviour. An aging population might shop in different ways from a younger one. Here a 'hard science' approach may well pay dividends, provided it has due lassitude for the inevitable uncertainties about different contexts and imprecise variables built in. At an aggregate level geo-demographic data predicts purchase behaviour, albeit not strongly. Extensive empirical work by Ehrenberg and co-workers (see for example Charlton & Ehrenberg 1976 but there is a long list) demonstrates that previous purchase behaviour is a strong predictor of future purchase; existing market share predicts future market share and so on. Theoretically, investigating the profit impact of marketing strategies (the famous PIMS studies) across many firms also looks attractive science. Problems here may not arise because this is inherently un-scientific, but because the sheer complexity across products, markets, conditions and so on prevents control and comparability.
On the other hand a lot of the social (marketing) world may be more safely described as a social construct that is difficult or impossible to precisely externalise outside of human debate. Here we are thinking of complex, nebulous, difficult to define constructs such as people's feelings of nostalgia and how this may link to the phenomenon of car firms 'retro-marketing'. A number of questions arise when researching such a topic, of which whether we are dealing with external reality or social constructions is probably less important than others. To begin with, can a definition of nostalgia and for that matter 'retro marketing' be agreed upon? Next, can these entities be consistently measured, in different situations, and over time? Given the uncertainties and complexities in delving into the socio-psychology of nostalgia, is there much point in trying to express the concepts in a quantified model?
The problem with our literature on complex phenomena like nostalgia is that the inherent uncertainties of such social constructs simply aren't acknowledged, and hence aren't dealt with. The pretence of a precise, predictive model that everyone agrees on is then entered into. Uncertainties of definition are skipped over. The external validity issues of different contexts aren't properly addressed, which is unfortunate if the phenomena is situation specific (nostalgia emerges very differently across different nationalities).
This leads us to our first suggestion, shown in Figure 3, that research in marketing should be methodologically driven in a 'horses for courses' pragmatic manner.

Scientific marketing -a reality check
A more pragmatic approach to epistemology needs to be accompanied by finding ways to link our research into context and to get closer to an often complex reality, as the wider marketing community understands it. Let us consider market segmentation, a subject of ongoing interest and importance to the wider marketing community, particularly in the light of developments in media and technology over the last ten years. While it is true that segmentation's place in marketing is reasonably secure, academically not much recent work has been done on this in our elite journals. In fact there have been no significant papers in the Journal of Consumer Research (JCR) on segmentation since 1997; and only two in the Journal of Marketing since 2001 that focus upon segmentation. There have however, been plenty in the Journal of Marketing Research (JMR) concentrating on obscure debates about complex mathematical techniques predicated on the 'hard science' belief that segments are stable, useful and real entities that can be refined and improved by number crunching. However, these assumptions are open to question at anything below the grossest level of aggregation, as demonstrated by Wensley (1995) and although Wensley's conclusions have been contested by Saunders (1995) we would argue that there is sufficient uncertainty and debate concerning the entire edifice of segmentation as 'hard science' to cause us to question the point of the dozens of JMR papers over the last decade. Our contention here is that JMR style 'hard science' has poorly served the user community or academia as a contribution to segmentation knowledge. We need to recognise the uncertainties, the instabilities of the segments, while still investigating how that which remains is useful to managers in organisations, the implications for strategy, and so on. Researching how to manage these uncertainties is, we feel, better than pretending they don't exist.
The academic segmentation literature is now dominated by obscure number crunching that ignores the contradictions and uncertainties argued over by Wensley and Saunders and yet the user community is crying out for work that addresses these contradictions and can work with the imprecision and uncertainties of reality to create (imperfect but realistic) solutions to their real life challenges. Significant organisational issues arise in managing segments which has been addressed (see for example Hammond et al 2001; Dibb and Wensley 2002;Dibb and Simkin 2001;Hunt and Humby 2003) but this stream of work is not currently recognised by elite journals. This is a pity, because we'd argue this is where critical future work in segmentation should lie. Hunt and Humby's (2003) book 'Scoring Points' discusses in detail the story of UK supermarket chain Tesco's use of consumer loyalty card data to create a strategic, company changing segmentation approach that took enormous resources, five years of hard work, and was one of the most important factors leading to the enormous power Tesco now has in its markets. The case study reveals many of the management and commercial difficulties, and provides numerous insights that give hints on issues that are of interest to academics: such as the most effective variables to use; the most effective methodologies in practice; the role of data in strategy and tactics; segmentation implementation difficulties; the importance of context and situation in deciding how to go forward; how segmentation can influence market orientation, and so on. The value of 'Scoring Points' is that it provides a piece of context specific knowledge that improves understanding. It inculcates a multiplex of theory and data and resists answering the dilemmas with one model, one analysis or one theory. In short it acknowledges the complexities and situation specifics of marketing life.
This leads us to our second suggestion, that academic rigour is too often focussed on methodological robustness, while missing the fundamental foundation of the robustness of the research platform. The robustness of the research platform relates to the degree to which the research is connected to a reality that would be recognised within the wider marketing community. Where academic rigour is purely defined in terms of a rigid scientific model to the detriment of a consideration of what is important, useful and meaningful we would question the extent to which a piece of research truly makes a contribution to knowledge. In this section we have illustrated how 'hard science' can impact broadly on the output of an academic sub-discipline and render much of the research meaningless to anyone, but a small group of academics. In the next section we critique one paper in detail.

A critique of a recent 'top journal' paper
To illustrate our points, let's look in detail at a 'hard science' top journal paper. Bart et al's (2005) Journal of Marketing paper 'Are the drivers and role of on-line trust the same for all web site consumers?' reported the results of a large scale cross sectional study that linked web site characteristics (privacy, security, navigation, brand strength, order fulfilment, community features and absence of errors) with consumer characteristics (on line expertise, familiarity with the web site, and Internet buying/entertainment experience), trust and subsequent positive outcomes. A pilot qualitative study underpinned a survey to over 6000 respondents; with the results analysed using structured equation modelling. Throughout the paper the research is of huge scale (albeit cross-sectional), the techniques are rigorous and professional, modelling is of the highest order. The findings seem broadly reasonable in terms of common sense experience (although the segmentation analysis looks like an exercise in number crunching that achieves little).
If we look a bit harder at the questionnaire used, complex constructs such as trust are simply measured by the level of agreement with phrases such 'this site seems more trustworthy than others I have visited', 'this site represents a company that will deliver on promises made', and 'my overall confidence in the recommendations made in this site is…'. Within the local rules developed by academic marketers on measuring trust this is understandable. Trust has been endlessly debated, and for the authors of this paper to rehash these arguments would have been rather tedious. But the extent of the debate about the definition and measurement of trust reveals the difficulty of claiming that by calling something 'trust' in a piece of research that everyone is on board, or that you are close to the external truth of what 'trust' is. If ever there was something that may be better treated as a social construct, 'trust' may be it. Substitute words like uncertainty, dishonesty, confidence, reciprocity, friendship, relationship, faith, hope, reliability, and so on, could have been applied. If I talk to you about trust and you talk back to me, how do we know we are both talking about the same thing? The same issue of subjectivity applies strongly to questions in questionnaires. All too often, the person answering the question has read something different into it than was intended by the researchers. I could access your web site and I may trust you to give me a product that has value for money, but maybe I don't trust you to be honest. If you then research my 'trust' in the site, what exactly are you researching?
Bart et al may counter by arguing that all these issues average out and the model as a probabilistic entity still holds. Maybe -this isn't totally convincing -but let's assume this is so. It still begs the question: what is the point of giving the impression that this model accurately and precisely explains and predicts? Apparently 'shopping experience' is 0.12 correlated with 'trust'. If this is a science then 'shopping experience' is something out there that we all agree about, ditto 'trust', and we should be convinced that the relationship between them is only 12% correlated. An experienced web business manager may have difficulty swallowing these certainties. But they would certainly ask: what is the point of the precision of the numbers in the model? The impression is given of a highly tuned, finely balanced and hugely predictive system of web sites, people, and outcomes all dependent on privacy, trust, navigation and so on with executives able to tweak the controls: "Dan, HQ here, increase navigation by 12%, ease up on the privacy by 6% and this will improve sales by 3.8%". This is nonsense of the highest order and everyone who is still in touch with reality knows this is nonsense, including, we suspect, the authors. So why is it that these models are seen as the apex, the final triumph of marketing knowledge?
Perhaps we should re-iterate that in no way are we implying that the research was shoddy. Neither do we see this paper as any worse than others of this type. Indeed it is better than many -at least the authors' investigation covered different sectors and took account of these differences -many such studies do not. Our view is that the marketing literature is replete with studies like Bart et al's, which miss the chance to relax a little, and move away from the pretence that we are in a position to treat the study of marketing as a predictive science. Quantitative studies systematically give the illusion of comparability with each other, when in fact the significance and meaning of the numbers is constantly shifting. This brings us back to our two suggestions: on the one hand, of the need for methodological relaxation and on the other, for grounding research in a reality that would be recognised by the wider marketing community.

A better way -establishing a more robust research platform
What could Bart et al have done instead?
Their budget for this work involved sending questionnaires to over 90,000 consumers, so we can assume funding was at least reasonable. Perhaps they could have looked for qualitative understandings of the concept of trust on-line, how and why this differs from general shopping environments. A general quantification of some key relationships between variables would be useful, but precise modelling less so. Tracking variables over time, split by different on line sectors (in the very sensible way defined by Bart et al) would produce valuable data on the stability of any relationships. We'd then develop some tentative hypotheses from the work to date, treating our findings with appropriate scepticism, and take these hypotheses into interviews with experienced executives in the field. Then onto case study work, keep the tracking study going, tweaking it as we learn more. And so on. Such a process, as illustrated in Figure 4, would build knowledge in close proximity with reality rather than in isolation:

Figure 4 How Bart et al. could have conducted their research
The methods are less important than the iterative process that we outline. In this way knowledge is built up with repeated reference to the social context in which the phenomenon under study operates. Scientists understand that their theories, models and constructs are social constructions of reality rather than reality itself, so why can't we? While 'hard science' research may be appropriate for tackling a number of questions within marketing, the problem, as we see it, is that it tends to be applied on a carte blanche basis to most research problems reported in elite journals. We have six major issues with this:

Why we need to rethink the research approach in academic marketing
In contrast to chemistry, constructs in marketing are woolly and prone to disagreement. All too often in academic marketing we don't deal with agreed factual entities like product purchases. Quite often, what constitutes a fact is a lot less certain and rather woolly. There are lots of phenomena where facts are disputed, such as attitude data, survey responses, observations of social phenomena such as what took place at a meeting, or complex psychological constructs such as self image, proneness to nostalgia and so on, or variables that are difficult to define such as market orientation. These subjective and socially (dis)agreed variables may be less easy to treat scientifically than, say, simple purchase data. Unfortunately in the elite marketing literature no allowance is made for these differences. Statistical treatment of operationalised constructs give the impression of science, but there is a world of difference between a hard fact like a sale, and an uncertain construct such as self image. No amount of statistical wizardry makes up for this.

Chemists deal with absolutes, marketers usually deal with probabilities
Probabilities are strange things to treat scientifically. On balance, observations and findings suggest more women enjoy clothes shopping as an experience than do men. Or people tend to spend less time buying a chocolate bar than buying a car. But these aren't hard certainties. Both facts and theories are woolly and vague compared to physical science. But marketing science systematically gives the impression of precision, with probabilities expressed to decimal places, adding to the illusion of precision. Perhaps, we need a more honest approach that admits the limitations of marketing theory and research. One implication of this might be less guruism and more of a critical/reflective approach to problems from both the academic and practitioner sides of the community. Academic marketers do not have all the answers and practitioners should not expect them to. Rather, both parties have complementary experience and skills that could be more effectively utilised if brought together in tackling marketing issues.
Sodium explodes in hot water all round the world. But people behave differently from one place to the next. Marketers face uncertainty over external validation of their results across different contexts. Loyalty cards may be of minimal value to consumers in Europe but does this translate across all markets? The external validity of much of our elite journal work is pitifully low, but this is never discussed. Again this relates to the point made above of the need for a critical/reflective approach from both academic and practitioner sides in different contextual situations.

Chemists can isolate the problem. Marketers can't.
Marketing academic models have a tendency to work a bit better at the gross aggregate level but break down when we get to any sort of useful level, for example a firm and its customers. The failure of Tom Peters' Excellence models; the difficulty in persuasively demonstrating the link between market orientation and profitability in different contexts; the relative lack of any successful link between marketing planning and a firm's success. All these things point to systems, structures, variables that are so vast and complex (almost weather like), that they are pretty much impossible to predict. Technically, the modelling techniques can't handle the complexity. But even if they could, the hundreds of independent variables that you'd have to put in mean that any one variable would have minimal impact. (But even the weather is easier to try and model than human, firm, or market behaviour. At least the weather doesn't wake up one day, spot the trends it has set, and deliberately set out against them). Ironically, marketers can't isolate the system, but do isolate their theories As we discussed earlier in science theories do not exist in isolation, but as part of an integrated 'web of belief' (Williams 2000). But in marketing, because we lack agreement, there is not a proper web of belief. And this means when a theory is developed in a paper there is a sense in which it is an isolated event, linked to other 'top journal' literature, but completely lacking cohesion with a more general view.
There is a growing recognition in contemporary science of the importance of Mode 2 knowledge production (Gibbons et al. 1994). Mode 2 knowledge is not just produced in universities and research institutions; rather it may emerge from a range of organisations and institutions. Most significantly it is impossible to detach Mode 2 knowledge from the context of practice, recognising that new knowledge is produced during the process of implementation and therefore needs to be studied in this context. Within this the social context in which knowledge is produced becomes crucial and requires that universities recognise that their scientific and social roles overlap to a large extent (Nowotny et al. 2001). If ever a discipline could be seen to require a Mode 2 approach to research, it would seem to be marketing with its origins in practice.

The formula driven approach of scientific marketing kills thinking by the researcher that in turn kills insight and understanding
Our view is that a narrow research formula is increasingly deployed in academic marketing as a substitute for thinking rather than an aid to it. This is reflected in the format of many top journal articles in marketing. The language and the stilted process seem to disallow freedom to muse and ponder. The formulae drive the researcher to an expected outcome in which the answer must be… a model with antecedents and consequences. The very distance this leads us from our lived experience inevitably leads to a dumbing of the intellect -there is nothing for us to iterate our models against, no room for interpretation, for a search for understanding. Bringing insight and understanding requires a more open approach in which research is debated and contested outside of the confines a handful of academics.

A summary of our position: the militant middle ground
If approaches based on a narrow highly positivistic interpretation of what is 'good' research imply a precision that we think is false, then social constructivist positions can be just as damaging to the academic-reality gap. To take the position that all reality is a social construction is, to us, merely academic posturing. Do those who adopt this stance in university debates or in articles then take this into their lives by, say, standing on a motorway and testing whether the cars coming towards them are social constructions? Those who argue that nothing is predictable and all forecasts in social science are a waste of time may also be striking an academic pose that they don't take into their everyday lives. When they drive in and get stuck in traffic jams day after day do they adjust their behaviour by forecasting that there may be a jam tomorrow? One suspects they do. Those who reject science wholesale as they tap out the words on their laptops strike an equally odd position. The debate about the use of scientific methods in academic marketing research is not always played out with critical neutrality. It is worth reflecting that many critics enter social science because they are critical of the social order and part of that order is science. In other words as Williams (2000) suggests, many social scientists arguably begin as rebels, and their critiques of the use of science in marketing are perhaps a little too aggressive as a result. We don't believe pure social constructionism is the answer, and we wonder if those who advocate this philosophy carry that stance through to their everyday lives. Fascinating though the entrenched science versus social construction debate is to academics, their everyday living habits suggests they see probably most things in life as 'external' and 'real', but with a realisation that agreeing on meanings, understandings and explanations of these is problematic.
Our observations and analysis of the problems of the overemphasis on a narrow interpretation of what constitutes acceptable research have led us to advocate a critical realist stance (Pawson and Tilly, 1997) for academic marketing research. This asserts that the (marketing) world is external but it can only be known in terms of incomplete descriptions and discourses. This acknowledgement allows critical realists to relax away from the constriction of pretending that we are dealing with a hard science. So, critical realism is comfortable with an interpretivist epistemology because it shares the view that the context of social phenomena needs to be understood and that only first hand engagement with participants reveals the subjective meanings and motivations that constitute actions. But it demands explanations as well as understandings -sitting somewhere between positivism and interpretivism. It acknowledges the world exists regardless of what we think of it, but that our knowledge of it is historically and culturally shaped. Thus, there can be differences between reality and our descriptions of reality. Another layer of relaxation derives from taking a socio-psychological rather than economic perspective where appropriate. The former lend themselves sympathetically to interpretive studies where an allowance is made for the difficulties of agreeing social phenomena, such as how and why consumers interact with each other.
We are also influenced by Carson et al (2001) and Wilk and Mick (2001) who suggest a tolerant pluralism, including the use of different forms of enquiry sequentially, and the use of multiple perspectives simultaneously. The call here is for different research perspectives to work together to demonstrate their value, rather than fight to demonstrate their superiority. Science drives the researcher towards atomism: breaking down phenomena into constituent parts. In contrast pluralistic approaches encourage holistic approaches that try to build up a 'big picture', acknowledging the crucial importance of the situation or context that the consumer, organisation, or market is in. Certainly, if practitioners usage is important then we need to reflect their world: users of the work in practice will have to deal with uncertainties, imperfect information, half finished, messy problems, and politics that colours the objectivity of what to do, all the time (Carson et al 2001).
Our position advocates the importance of interacting with the world of practice in setting research agendas and in designing, carrying out and interpreting our research.
In this, we are not as naïve as to think that practitioners are queuing up out there in order to participate in our research. However, in our experience there are a substantial minority of practitioners who are open to approaches from academics and recognise the benefits of critical reflection. These people need to be nurtured. But more than this we need to demonstrate the value of collaboration to a wider audience of practitioners by considering research agendas that do reflect the needs of their world. We also need to consider how we communicate with practitioners. What kind of impact would marketing academics have on practice if we put a quarter of the effort that we put into publishing in academic journals into publishing in the business and management press? Furthermore we could be more proactive in encouraging and developing more knowledge networks, joint forums and conferences and other forms of collaboration.

Conclusion: never believe without doubt
Probabilistic science in marketing can work but has quite severe limits. Entities like consumer behaviour create philosophy problems -ontologically we cannot agree on only one version of social reality, and practical problems -our methods aren't good enough to map one-one with external reality. Sometimes at the extreme, researchers will report causal laws for human behaviour. This kind of determinism becomes ludicrous if taken literally: again, free will prevents scientific style 'laws' being anything other than watered down to probabilistic predictions.
In a recent (2004) editorial to his journal, the editor of MIS Quarterly Ron Weber expressed surprise at the way positivism was being characterized by interpretivists, and made the following observations about positivists in social science: • "Of course the world is external to us. But how well we can perceive reality, how well we interpret reality, and what actions we take in light of our perceptions and interpretations are other matters. In this regard, I suspect it is easier to obtain agreement about certain kinds of phenomena (e.g., what happens if we step off the ledge on the third floor of a building) versus other kinds of phenomena (e.g., what happens when several individuals interact with each other, or what some person believes when she or he observes some event)".
• "I know many researchers who claim to be positivists. As best I can tell, all recognize the inherent limitations of the knowledge they seek to build. They understand fully that their culture, experience, history, and so on impact the research work they undertake and thus the results of their work" • "Positivists accept they can only know reality through the artefacts they have created--theories, frameworks, constructs, and so on" • "Few positivists, if any, would subscribe to the correspondence theory of truth (a statement made by a researcher is true when it has a one-to-one mapping to the reality that exists beyond the human mind)" • "As a positivist I reject the notion that I try to measure reality in my research. I have no way of knowing reality, so how can I know whether my measure of reality, whatever reality might be, is valid?" • "The differences between positivism and interpretivism, if indeed any exist, are shallow rather than deep".
This kind of relaxed positivism is exactly what we have tried to illustrate here as an exemplar for academic marketers. In an everyday sense managers, executives, consumers, people in their ordinary lives make decisions that essentially assume that a) the world is 'out there', but that b) different points of view exist about what is going on and how it works. In everyday life people who hold no doubts that they are wrong are generally ostracized. The same may be true in 'everyday academia' but for some reason this judgement is suspended in journal articles. As a result a false precision is communicated. This hard science applied to marketing has been rejected by the outside world in favour of a mix: models that offer stability and a platform for knowledge are accepted, but are mixed with more human, softer, stories and case studies.
Marketing science academics need to adjust to the imperfections of reality. To some, interpretivist-critical realist methods create knowledge that is messy and less pleasing than a nice clean scientific model with an elegant underlying theory. In chemistry the Periodic Table has great beauty with its reach and awe-inspiring qualities of prediction. Alas, in business and management we have no equivalent, and the desire to find such a thing is apparent in the positivistic struggle of the 'elite' literature.
As a result we seem to have lost our collective ability to think. Our hard science stance removes our room for manoeuvre. A golfer who worries about his putting needs to relax. Relaxing allows the golfer to get away from logical analytical mode and into an intuitive, kinaesthetic feel for the problem. The analogy doesn't stretch too far, but a whiff of truth remains -relaxation about method allows us to concentrate on content instead of techniques.
We should not eschew models, but we should make explicit our everyday, intuitive understanding of their limitations. Models are often just a start point for problem solving, not an end point. This gap between the model and the final solution is a fruitful space for academic research. It is also an opportunity for dialogue with the world of practice that can enrich the quality of our findings. The gap may be filled with case study work, qualitative studies, ethnographies or indeed any technique that helps us understand the difference between a model and the local reality. The issue is not that this doesn't happen, but that this kind of work is not valued by upper echelon journals, is under-valued by RAE assessors, and hence is frowned upon by deans with targets to reach.
Adopting the stance that we don't have to pander to outside interests has got us into trouble. If our work engaged with reality, and the scientific marketing approach was getting somewhere, then the ivory tower stance is justifiable. The trouble is that our lack of agreed definitions, our pretence that woolly constructs like trust or satisfaction are absolutes, our ability to ignore hundreds of extraneous variables in our modelling, while suspending any disbelief, is collectively breathtaking in its pseudo-scientific ignorance. Would a better approach to be to get off our hard science high horse, roll up our sleeves and engage with the 'messy' reality of market places? By seeing the world as it actually is rather than as we want it to be we might have a better chance of understanding and explaining it.