Editorial

Citation

Crosier, K. (2007), "Editorial", Marketing Intelligence & Planning, Vol. 25 No. 6. https://doi.org/10.1108/mip.2007.02025faa.001

Editorial

The selection of articles in this issue marks something of an editorial departure. Not a volte face, exactly, nor a Damascene conversion, nor yet an epiphany. But a definite, albeit hesitant, change of heart. The clue is that four papers report the application of model building and sophisticated statistical analysis to data collected via questionnaires in the form of answers on simple interval scales.

A long established, highly regarded restaurant in Glasgow, The Ubiquitous Chip, was so named as a sardonic hint to the minority of cognoscenti that it was not going to follow the practice of the ignorami of the times by offering “chips” – French fries – with every dish on the menu. In the same spirit of rebellion, I was on the verge of renaming our journal The Ubiquitous Cronbach Alpha.

You may think you recognise the avoidist mindset of a typical British qualitative researcher, with a love of focus groups and vague “content analysis” accompanied by dread of quantification beyond nominal data, ordinal scales, standard error and confidence limits. Well, perhaps. But I like to think that my concerns were in fact methodological. As the recipient of many, many submissions announcing that structural equation modelling has been applied yet again to numbers representing the answers of students in the researcher's own university to poorly designed questionnaires and seven-point Likert scales, I could not help worrying that the availability of software-based packages was leading young researchers into temptation, particularly if they were good at mathematics. One reviews the academic literature of a subject area that someone specialises in, spots a small gap in the studies published so far (possibly a mirage), collects the same old data from a different source, loads them into one's PC, and ... out pop the “answers”. Hey presto. No need to work out why specialists in Marketing Intelligence & Planning (MIP) might be glad to read them.

The give-away symptom of this kind of sloppiness is the way in which the numbers are presented in narrative text. Machine codes and computer jargon replace the language of mathematics. (Yes, it is a kind of language, for explaining things that words cannot.) There are no spaces before the equals sign and no zeros in front of the decimal point in numbers below one ... and so on.

And yet our reviewers saw merit in the current batch of papers and thought you could learn something useful from them, if the authors were willing to explain the managerial implications of their findings beyond a paragraph just before the one on methodological limitations and directions for further research. They were, and did, and I was persuaded.

Thus, I resolved to emerge from the primeval swamp, and duly read up on structural equation modelling. A revelation! The problem had been, of course, that “structural” means nothing to a non-specialist as an adjective qualifying “equation”. (Believe me; I asked several.) Furthermore, the “modelling” in question bears no evident relationship to the models of, say, Engel, Kollat and Blackwell, but seems in effect to be another statistical testing procedure.

Some of you will correct me if I am wrong, but I think I now understand that SEM is a potentially powerful multivariate analysis procedure that draws in specialised versions of several other statistical methods, particularly factor analysis in the generic sense of the term. It can be applied to causal modelling (aha!), sometimes called path analysis, to second-order and confirmatory factor analysis, to regression analysis, and to covariance structure (aha!) analysis. It therefore typically begins with a mathematical model or a path diagram. The software churns out specification of the model, estimation of parameters, and assessment of its “goodness of fit” to the observations. The researcher is then meant to modify the model accordingly, interpret the findings and communicate the implications to interested parties. That is the commonly unsatisfactory bit. The final stages of the process, replication and revalidation, are typically delegated to “future research”.

Even if you tell me that this insight is superficial or even wrong, it helped me to interpret the reports of more expert reviewers, and to make reasonably informed editorial judgments about manuscripts that often, frankly, erred on the side of obfuscation. So may I just make a gentle plea? If you submit a paper based on structural equation modelling, why not describe exactly what you and the software did, for the benefit of your inexpert readers, rather than incanting “SEM!” and daring us to ask the magician about his or her spells. What was the point, in the first place? Why did you do the tricks you did? How did you read the runes? What will we have learnt? What can we do with it?

Enough. Let us turn to the applications of quantitative analysis presented in five of the articles here, after a word about the one that in fact contains no arithmetic.

Geoff Simmons of the University of Ulster offers us almost ten thousand words of valuable explanation and discussion of how branding, a key mainstream marketing strategy, can best be adapted to the environment of the internet. His comprehensive literature review transfers principles and practice from many domains well beyond the usual sphere of marketing competence. Some of what he has to say will be familiar to all of us, but its value lies in the joining up of the dots and the sheer scope of the discussion. Interactivity in cyberspace involves familiar vocabulary and everyday hardware, but remains a bit of a mystery to those who are not actually doing it.

From the Murdoch Business School in Western Australia, Bill Chitty, Steve Ward and Christina Chua apply confirmatory factor analysis to data collected from questionnaires completed by almost 300 current customers of “backpacker hostels” in their vicinity. My first reaction was that this was a very small market segment; my second, noting the startling number of directly related items in the very comprehensive list of references, that writing about backpacking must be a cottage industry Down Under; my third, after reading the whole paper, that the authors had chosen an excellent case in point as the basis on which to model the antecedents of customer loyalty, and test the predictive ability of its components. Their key finding is that brand image predicts satisfaction and perceptions of value predict loyalty, though they modestly underclaim the generalisability of their model. To give praise where it is due, theirs is a shining example of how to explain the methodological basis of one subset of structural equation modelling.

Moving north to the Department of Management and Marketing at Hong Kong Polytechnic University, we find Y.H. Wong, Humphry Hong and Wing-ki Chow reporting a study that applied factor analysis to data collected by street interviews with just over 200 customers of financial services providers. Though Hong Kong is a major world financial centre, they conceded that the attitude and behaviour of its citizens as consumers of banking and insurance are not directly comparable with those of their counterparts in the equivalent centres in the West. With that proviso, they found significant links among the quality of customer-provider relationships, information sharing between the parties, customers' intentions to re-buy, and their willingness to recommend “their” service provider to others.

Back in Western Australia, Donna Gill and “Ram” Ramaseshan of the Department of Marketing at the Curtin University of Technology report their study of long-term repeat business, this time focusing on the determinants of wine importers' decisions to keep giving their business to the exporters who are already their suppliers. On the basis of regression analysis of data collected by self-completion questionnaire from just over 150 importers (in the UK, as it happens), they conclude that the supplier's performance with respect to relationship commitment, payment facilities and product quality has a positive impact on repurchase intentions, but price and brand recognition do not. The links to Wong, Hung and Chow are obvious. The dominant focus on lessons for exporters could of course be turned through 1808 by intelligence-gatherers and planners at the importer end of the telescope.

Now, to Europe. Spiros Gounaris, George Panigyrakis and Kalliopi Chatzipanagiotou of the Department of Business Administration at the Athens University of Economics and Business report their development of a new framework for evaluating the effectiveness of marketing information systems from a critical review of the relevant literature. They fed the responses of just over 250 marketing managers of five-star hotels in Greece into exploratory and confirmatory factor analysis, and found that the effectiveness of their systems depended on both internal and external components: the extent to which a user improved functional effectiveness and corporate climate on the one hand and its adaptability to market conditions and customer intelligence on the other. Their proposed measuring instrument integrates those components into a single test of effectiveness.

Finally, we move to another continent and an altogether different academic discipline. (It pleases me to be able to say that, to an international readership and as a committed enthusiast for knowledge transfer.) S. M. Musyoka, S. M. Mutyauvyu, J. B. K Kiema, F. N. Karanja and D. N. Siriba are all from the Department of Geospatial and Space Technology at University of Nairobi, in Kenya. Their article applies geographical information systems (GIS) technology and virtual mapping to the devising of logistical and transport solutions for a company distributing soft drinks from the bottling plant to retailers in one district of the city. The outcome shows vividly, despite its highly specialised foundation, how the analytical and visualization capabilities of GIS can enhance the communication and utility of intelligence for marketing planning, as compared with its normal presentation as text and tables in narrative reports. The wider relevance of this case study for specialists in MIP requires no further comment from me.

Keith Crosier