MARKETING MATERIAL

9098 – Marketing Practice 7 MARKETING RESEARCH 

Chapter 7

MARKETING RESEARCH

After the diversion at the end of the last chapter into the complexities of the complex (industrial) sale let us return to a simple focus on the uncomplicated consumer. In particular let us address our theme in this chapter: marketing research. How do you find out about the outside world in general and about that consumer in particular?

More of the material in this chapter will, indeed, be descriptive - and there will be rather fewer practical rules than in the other chapters - since the intention is that the reader understands what lies underneath the surface of sophisticated research programmes. The practice should normally be that the work itself should be subcontracted to the best available research agency.

First, then, let us look at how such a marketing research process generally works. To make the explanation simple I will concentrate on just one project: say a piece of conventional doorstep survey research. Much the same features can be found in all such research, in desk research for example, but in these cases it is complicated by the number of such individual pieces of work running at the same time, so it is more difficult to see where one begins and ends.

Such research projects have a characteristic shape:

In broad terms the horizontal dimension indicates the passing of time as the project progresses, and the vertical one the number of people involved at each stage.

HIDDEN TIME - the main message which emerges from the horizontal progression is that the more obvious elements where you might expect time to be taken, especially those involved in the seemingly complex and relatively lengthy process of data collection, typically only represent a minor part of the overall process. If the research is to be fully productive then the periods at the beginning when it is designed and at the end when it is used as the basis for action plans must be allowed to progress at their own pace. Rushed research is too often wasted research. Sometimes you do need answers very fast, but you must then recognise that, in such a situation, the questions better had be very simple.

THE BULGE IN THE MIDDLE - it is in the middle of the diamond where the main manpower resources are eaten up ( and the major costs incurred). But this is usually a matter of relatively menial legwork. Getting the first, design stage right can often reduce this bulge to a more manageable size. Thus, time spent earlier, on design, can often save money; and not infrequently time as well - since the later stages are better planned.

THE CUTTING EDGES - the diagram is well named the 'Research Diamond' since its most important features (albeit the least well recognised) are the cutting edges at the beginning and end. The most important input to research is the design objectives. If you have only a fuzzy idea what you want out of it at the beginning, you will get a fuzzy set of results at the end.

But by far the most important cutting edge, however, is that at the end: the ACTION which is generated as a result. Indeed, the focus on that action must begin with the objectives. If the research has no planned actions depending upon it you must question why it is being done. But, above all, you must act on what you eventually find.

The Japanese often seem less sophisticated in the research they carry out, but they always do something with the results. What is more they usually do something significant. Western marketers are all too prone to look at the many pages of results they receive, find them quite interesting and then file then; before continuing their business activities once again, untainted by what they have just read!

ENVIRONMENTAL ANALYSIS

As with all good rules there must be exceptions, and environmental analysis in general, or scanning - which starts the whole information gathering process - in particular,  certainly falls into this category, since it is not required to directly result in action. Environmental analysis is a very wide ranging activity. In its broadest sense it encompasses all those activities which the organisation uses, formally and informally, to keep abreast of those changes in the external environment which will affect its future. At its widest, as in scanning (the horizon), it can include all the actual (news and documentary) material to be seen on television or read in the newspapers and periodicals.

Certainly you should not rely on just one newspaper (even the Wall Street Journal or the Financial Times), which will almost inevitably be biased in one way or another, but taking a balanced range of these.

It is sometimes called Undirected Viewing[1]. If it is carried out formally, it potentially represents a major expenditure of resource. But it also offers the most important benefits in terms of long term survival (though it is typically neglected by many organisations).

The one further requirement is expertise in your subject. This will not help you if your viewing is not wide enough, but with such expertise you are more likely to recognise what items are most likely to bear upon the future of your organisation. There is no easy answer; but this 'attitude of mind', which I call 'Informed Viewing', is the best that is generally available.

To be most useful this file, say a number of drop-files in your filing cabinet, should be arranged by groups of the subjects which you are currently studying. In the first instance these groups will need to be very broad, because undirected viewing by definition covers a very wide scope. Slowly but surely, though, a few of them (less than one in a hundred perhaps) will start to develop a focus, as you gather more information and patterns start to emerge: this is the whole essence of scanning. Recommended practice might be to collect the clippings as you find them and then file them once every two weeks (but two weeks in arrears); as we will see, once a week is probably too often and once a month too infrequent (not least because you will forget to do it). When you do this filing you should ruthlessly discard, before you file them, those clippings which with the hindsight of the two or more weeks now seem to be a flash in the pan (hence the two weeks - less than this does not seem to add sufficient perspective). At the same time you can weed out those already in the file which now too seem to had outlasted their time; and, most importantly, see if any new trends are emerging.

One of the most interesting suggestions for handling this came from an organisation which asked all its employees (shop-floor workers as well as its managers) to clip any news item (found in the newspapers and magazines they regularly read) they felt might be relevant to the future of the organisation. All of these clippings, from the most sensational tabloid newspapers as much as from the serious press, were then 'scanned' by the environmental analysis group. When a pattern emerged, of a phenomenon being reported across a number of such sources, it was reasoned that these particular 'weak signals' possibly indicated an important underlying trend, and it was thereafter tracked in more detail. This seems to offer a particularly comprehensive approach to such coverage. It may be beyond the culture of most organisations, but it could be adapted to work across a smaller group (say those in the marketing department - but including a range of personnel, for example secretaries as well as senior managers).

This process will, in any case, benefit from the advent of increased computerised communications - where the extraction of multiple occurrences of data might be detected automatically by computer systems.

As hinted at above, there has been a considerable amount of academic discussion about 'weak signals'. These are small pieces of information which signal important changes which are as yet unrecognised, since their main impact has still to come. The comment, largely by academics, has concentrated upon retrospective analysis. Thus, for example, the 'signals' that Japanese manufacturers were coming to dominate certain industries (the motorbike industry, for instance) are, once you look for the related 'weak signals', obvious to see; and yet they were totally overlooked by the existing suppliers in these markets. This is an important message to convey to those conservative organisations which have long since settled into a comfortable rut; and may not be able to, let choose to, read these danger signals.

The problem, which most of these academics tend not to address, is that whilst it is easy to see these patterns with the benefit of hindsight it proves very difficult indeed to detect them in advance. They are, by definition, weak signals. There is no obvious evidence of their special importance and they are buried amidst large amounts of similar data; which acts as 'noise', drowning them out. As mentioned earlier, the only generally recommended answer is that the 'reader' should be an 'informed' observer - who knows rather better than most (because he or she is an expert in the 'industry', for instance) which of these 'weak signals' is most important. Indeed, the term 'weak signals' is a bad one - since it implies that (with its analogy of radio signals) there is some scientific device which can amplify these. There is no such device!

There is, in fact, a whole range of approaches to the general topic of environmental analysis;

As can be seen from the above diagram, the scanning approaches start with the very wide approach of undirected viewing - scanning - and progressively narrow until an increasingly tightly focused 'search' activity is employed. We will look at these later (informal search is typically associated with desk research and formal search with that and survey research). In the present context, however, we are concerned with 'conditioned viewing'. To a degree, at least as defined by Aguilar, this overlaps what I have described as 'informed viewing'; but in essence it is narrower than that. It takes the more clearly defined areas where relevant information might be most likely to be forthcoming (such as the industry sources, including journals and personal contacts) and concentrates on the types of information which are most likely to be most productive. In this way it narrows the scope of the research down by several orders of magnitude - so that it becomes a much more practical proposition for most organisations to resource. It is still, however, characterised by a lack of awareness as to exactly what to expect. The one major drawback is that it will not find those 'fractures' which originate outside of this field of interest. These are potentially the most dangerous (precisely because they are the most unexpected) but, as
some consolation, it is quite probable that even the wider (and much more expensive) 'undirected viewing' would also fail to identify many of these.

INTERNAL DATA

We now move on to discuss what specific marketing information is available for the marketer; and how it can be retrieved and processed. The first, and often the most important, elements of this relate to the information held within the organisation. In recent years, and in particular since the widening availability of computerised databases, the whole process of collecting and distributing internal held management information (including that relating to marketing) has become systematised. The system, which handles these processes in a controlled and co-ordinated fashion has come to be called the 'Management Intelligence System (MIS). Such a system, though, can be much simpler than this - without the implementation of any computerised systems at all - and is often most effectively used in these simplest of forms. In this context, it is just the systematic collection and organisation of the 'data' which is relevant to the needs of the marketer.

Even so, in most organisations the key data on its performance is likely to be already available on its computer databases. This is because the related paperwork, typically deriving from order processing and invoicing, is now usually handled by computer. This should include accurate sales data; split by product and by region. In this electronic age it should even be possible to obtain up-to-date information via your own personal computer.

If the basic data is suitably organised, on a computer database, it may be possible to access it from terminals. The abstracted data can then be processed from a variety of perspectives. This means that ad-hoc reports or enquiries may be easily prepared. Regrettably, though, many of the key measures may not have been recorded. The data collected by the average system is driven by accounting needs. It records, even if the system is near perfect, only those transactions which result in the organisation actually completing a sale. In any case, such accounting systems are driven by accounting requirements, and in particular by accounting periods; and will often reflect an unbalanced picture until the month-end procedures have been completed.

More information is not necessarily a boon, however, if there is so much that it cannot be used effectively. There are a number of possible answers to the potential torrent of data

The simplest starting point is the ABC Analysis we have already looked at. It is very easy, in this context, for all it means is that reports are sorted with the most 'important' customers (or products or whatever is the subject of the report) at the beginning. Typically this will be in terms of volume (or value) of sales; so that the customers are ranked in order of their sales off-take - with the highest volume (and hence most 'important') customers at the top of the list and the many low volume customers at the bottom (since it matters less if they are not taken into account in decisions). As the 80:20 Rule, which you also met earlier, says the top 20% of customers on such a list are likely to account for 80% of total sales; so this approach can, in effect, be used to reduce the data to be examined by a factor of five.

A more sophisticated approach, if you are primarily interested in using the data for control is: 

This means that only those items where there are 'variances' need be reviewed.

The performance data described above has the great advantage, in terms of analysis, of being numeric. This makes abstraction and manipulation much easier; where arithmetical operations are commonplace. Much of the remaining data within an organisation is, however, available only in written form; as memos or reports, of which perhaps the most useful to marketing are the sales reports.

The reliance on words rather than figures may seem to make the manager's job easier; for many, if not most, managers are more at ease with words than numbers. Precisely because this verbal data seems so approachable, there is an unfortunate tendency to immediately accept it at face value. The reader's critical faculties are suspended, particularly if the message reinforces his or her own prejudices. In addition, the 'data' is often difficult to analyse; where writers of such reports tend to use the same words to mean different things, and the importance they attach to events more often reflects their own enthusiasm rather than any absolute measure. Collating a number of such reports and distilling these into an overall impression therefore becomes a matter of judgement rather than a simple analysis; and all too often is used to create evidence to bolster the manager's own preconceived ideas.

Often, access to the key data is limited to a few people; where the circulation of a memo is rarely to more than half a dozen people - especially from sales personnel in the field (where there are likely to be no carbon copies). The traditional system then requires that the recipient (say, the regional sales manager) recognise the importance of any data and then abstract this to incorporate in his or her own reports to higher management. The message thus travels hierarchically through the organisation - being filtered and distorted at each stage. This inevitably incurs delays (where reporting periods typically increase in length the higher in the organisation the message reaches). More important, it demands that a number of intermediaries recognise the significance of the data. If just one of them ignores it (because he or she does not see its relevance, or even because he doesn't want it to go higher) that data is lost to those in the chain above.

On the other hand, the increasing use of 'electronic mail' should have a dramatic effect on the availability of such information. Using this, it is almost as easy to send a memo to a hundred recipients as to one. Indeed if, as is the case with most such systems, 'standard' distribution lists are available, it is even easier. In this way, the data now becomes available to everyone immediately and it may be distributed to ten times as many managers; providing them with data they previously might have missed and yet needed for their work. More likely, but equally important, it provides them with a better perspective of what is happening throughout the organisation.

Perhaps the main shortcoming of verbal material is its retrieval. If not filed in the waste-paper-basket, memos and reports are consigned to the vagaries of the manager's or department's filing systems These usually rely on categories which have long since outlived their creators, and are meaningless to their current users - so that the most useful file in many offices is the 'day file', in which copies of everything being sent out are filed in date order (which means that the area of search can at least be narrowed to a range of dates, if not to the subject).

There now are specially developed computer programs for storing and retrieving vast quantities of verbal information.  

EXTERNAL DATA

We now move on to look at external data. This is, as the name implies, data which does not come from within the organisation, and for which the main source is usually that offered by desk research. This is based on published data (in its widest sense), often referred to as secondary data (because it has been generated in response to someone else's questions). Once the data has been located, its handling follows the same processes as for internal data; described in the previous sections. It is the 'finding' that is different, and the key to sound desk research.

Some useful sources of data which might be considered in this search (by managers in general, not just those involved in marketing) are;

LIBRARIES

DIRECTORIES

NATIONAL and LOCAL AGENCIES

DATABASES

TRADE ASSOCIATIONS

EXHIBITIONS AND CONFERENCES

NEWS MEDIA

The widest ranging source of published data (on everything from details of ancient civilisations through to the latest stock market prices) is usually a library, typically a public library. Many libraries have a wide selection of non-fiction books which will provide background reading on most subjects; and should not be ignored as a source of data. The reference libraries which are usually part of the central library will hold even more. More important, though, is that these libraries have access to national libraries. As a result, if you can find sufficient information to specify the book (usually Author, Title, Publisher and Date of Publication - though often just the Author and Title may suffice) it can usually be retrieved from this source.

On the other hand, much of the published data is located in journals, often specialist periodicals, and the best source may then be one of the more specialised libraries, such as those run by trade associations.

The most important directories will also be available in your central library; but, again, the more specialised ones may only be found in those of trade associations. National and local government departments and agencies are often major providers of data; especially to support specific 'initiatives - but they may still be useful for other purposes.

A growing amount of information is being made available (mainly by commercial information providers, but also by government bodies) on computerised databases. This data, again, covers almost every subject; with vast quantities of information on technical subjects, such as patents, as well as the marketing information with which this book is primarily concerned. This data can be particularly easy to access; though the cost of doing so may be high.

One of the best sources of data, not least the 'informal' data acquired during conversations at meetings, is that of trade associations. There is usually a fee for membership; but this is frequently very good value, in terms of what may be learned from these sources.

Once more, though, the most important source of external data is likely to come from face-to-face contact. In the case of members of the sales force this will largely be part of business-as-usual, in the form of sales calls. Elsewhere, such opportunities have to be created. In most organisations perhaps the most fruitful sources of such new material are those likely to be found at exhibitions and conferences.

Finally, there is are a range of consultants who will take the whole information gathering process off your hands; for a fee (usually quite a large one - so that this is, understandably, the most expensive solution).

As a footnote, the most prevalent, but unrecognised, source of external data for all managers (and the one which covers, 'scans', the widest perspective) is that of the news media (especially the morning newspaper and the television news and current affairs programmes). The amount of information these provide is probably vastly greater than that received from any other source; albeit that the coverage is so much wider. The choice of news media read, in particular of newspapers, thus becomes important. The quality papers are, obviously, likely to be of more value than the 'tabloids', but ideally a number of newspapers should be read (even if only occasionally - and then only an additional Sunday newspaper) to judge the bias each almost inevitably imparts to even the simplest news item.

SURVEY RESEARCH

The most generally recognised aspect of marketing research is that conducted on consumers/customers (usually on a sample of these; and by professional interviewers). The stereotype is the market research interviewer standing on the street corner, accosting passers by or walking the streets clipboard in hand knocking at closed doors. As we have already seen, this probably represents only a very small part of the data available to any organisation. It is, though, particularly important data (since it often provides the only true 'listening' part of the dialogue with the consumer). In the main it is a process undertaken by consumer goods companies.

The starting point for obtaining such survey research is the specialist organisations which offers market research services; for it is a very large, and unusually sophisticated, organisation which will have the resources to handle all aspects of its own research.

The suppliers who usually offer the easiest and quickest solution are the providers of syndicated services. They typically have on-going standard research programmes the results of which they sell to a number of clients. Shared cost is one advantage of such an approach, but another is that these services are usually are sold on the basis of quality, rather than (as much of other research) only on the basis of price. In addition, some researchers will sell 'space' (or more accurately interviewer time) on the back of their omnibus surveys; so that you can ask one or two simple questions - which will be added to the end of the main survey.

As already suggested, apart from the ease and speed of obtaining the research information, the great advantage of all these shared approaches is usually that the cost to any one client can be that much lower. It is, thus, quite cheap (a few thousand pounds) to ask one or two key (but simple) questions of a large sample.

Retail audits, such as the A C Nielsen store audits, are the most sophisticated of such syndicated operations. The concept, though, is simple. An 'auditor' regularly visits each retail outlet on the panel and by means of a physical stock check on the lines being surveyed, combined with the information on deliveries, determines the 'consumer sales'. Such retail audits are generally believed to offer the best results in terms of accuracy; of the volumes of consumer sales and, in particular, of the value of such sales, for brand share calculations; as well as for the all-important figures of prices and distribution levels. This information is, needless to say, invaluable to any FMCG company wishing to control its sales though retail outlets.

Custom research is, however, the staple diet of the market research industry. The research organisation is commissioned by a client to undertake a specific piece of research. The research company then accepts responsibility for all aspects of the research, and will typically plan and design the research itself. But it may then appoint a sub-contractors to do the detailed fieldwork and analyses. The client company does not see these sub-contractors, and does not need to; all it needs to see is the outcome (and to be happy with the validity of this).

Because it gives a good flavour of what is involved in this very important process, I will briefly describe the six stages which are most generally followed in conducting survey research:

DEFINING THE OBJECTIVES

As was indicated by the 'research diamond' the most important stage of almost all market research, and the one at which the research is most likely to be misdirected, is defining the objectives. Only the client can know what he or she wants the research to investigate; though an expert from the market research organisation is usually involved before the research itself takes place - to translate the client's ideas into a framework which will be most suitable for that research. Even so, the objectives need to be clear, and clearly stated (so that the researcher understands them), and unambiguous. On the other hand most market research fails because it is merely asked to confirm the existing theories of the commissioning organisation.  

It is at this stage that the initiative passes to the research agency to undertake the detailed planning; though the more sophisticated client will still want to remain involved.

PLANNING THE RESEARCH

Before the main, quantifiable survey research can start, however, it is almost invariably necessary to undertake some qualitative research: to find out exactly what are the questions to be quantified. Only if you ask the right questions will you get the right answers, and in practice it proves to be remarkably difficult to decide what those questions might be. It is not enough to think that you know them, for if you did you probably would not need the research - and the great benefit of marketing research is that it explores what you do not know. Too much research is expensively undertaken to justify an existing prejudice which the manager already holds. Even if the prejudice is genuinely justifiable and the resulting answers are right, which they rarely are, such research must inevitably be suspect (the worst insult you can attach to any research).

So, the first step is to find out in as open-ended a way as possible what the customer thinks the questions should be. This poses almost as big a problem as the manager's prejudice. Customers, unfortunately, do not spend their time thinking through exactly what are their complex buying motivations. That is the brand manager's job. All the customers want to do is get on with the enjoyable occupation of buying things. It requires considerable skill and sophistication, therefore, to unlock the their minds and start to understand what really motivates them. This is, generally speaking, the province of the expert psychologist not the friendly door-step interviewer. Accordingly, qualitative approaches to research revolve around the very sophisticated skills of such psychologists.

There are number of such approaches in use, but the one which is used in the great majority of cases is group research, where the psychologist or specially trained interviewer leads a discussion amongst a group of customers.  As a result of the format these are often called group discussions or focus group research.  

The interviewer ('group leader' or 'moderator'), who has to be skilled in the technique and often is a trained psychologist, carefully leads the discussion; ensuring that all the group members are able to put forward their views. The interviewer's role is essentially a passive one, where his or her prime concern is to foster group interaction (and to avoid any one individual dominating the group).

The essence of such group discussions is that the participants can develop their own ideas in an unstructured fashion; interacting with, and stimulating, others. Indeed it is based upon the somewhat surprising observation, derived from clinical psychology, that individuals who share a problem are more willing to talk about it amid the security of others sharing the same problem. As a result it can be very powerful in revealing hidden attitudes.

The whole session is typically captured on a tape recorder for later analysis in depth. This approach, thus, allows insights which may be hidden from the pre-conceived questions posed in questionnaire surveys. In particular it allows the customers' own language to emerge so that the subsequent survey questions can be phrased in ways which are meaningful to them.

It is an excellent method for generating hypotheses when little is known, and is thus a particularly productive approach to piloting the first stage of larger research projects. It is, though, increasingly being used as a cheaper and faster alternative for those organisations which cannot afford the full-scale research - and, in line with the move to low cost research, even for those which can. This is arguably better than nothing, but if such use is planned the client should beware of attributing too much significance to it. The sample sizes are usually far too small to allow any statistical conclusions to be drawn. Unfortunately, some of the less reputable agencies commissioning this research may still try to attribute something approaching statistical significance (or at least an unjustifiable degree of importance) to the results. In this case, unfortunately, it must be you, the client, who has to recognise the very real limitations of the information which can be derived from them.

One especially powerful method of starting the process of unearthing the customer's deeper motivations is:

The precise aim of this technique, often called Kelly Grids after the name of the inventor, is to discover what are the key dimensions of the respondents attitudes towards the matter in hand (usually a product or brand - typically as part of a positioning exercise).  Thus, in an individual interview, each respondent is presented with a list of items. These items (around 15 - 20 being the recommended number) are normally presented in simple word form, for example the names of products or brands, or statements about them. Three of the items ('triads' - in the jargon) chosen at random from the list are presented to the respondent at a time. The respondent is asked to choose the two most alike. He or she is then asked to say why these two are similar, and are different to the third. The remaining items on the list are then sorted equally between these two extremes. The process is repeated with three further stimuli, again selected at random, and the respondent is asked to give another way in which the selected pair are the same, and different to the third. The whole process continues until the respondent cannot find a further new reason for the similarity/difference (which typically occurs after ten or so triads, depending upon the complexity of the subject being studied).

A number of such interviews (from 10 - 50) are conducted and the output analysed - usually these days by computer - to see which factors can be clustered. This stage of the process (as well as the interview itself) requires considerable skill - if the information is to have any worth. The output, though, is a set of dimensions by which the customers themselves would describe the situation (usually something like a product/brand position). Its great virtue is that it forces the market researcher subsequently, in the quantitative survey research which is the second stage to all these pilot investigations, to research dimensions which are meaningful to the consumer rather than to the supplier.

It has to be admitted, though, that using Repertory Grids is relatively expensive (at least in terms of pilot research) and, possibly as a result, they are rarely used in practice. This is a pity, since they offer one of the most valuable starting points for sound marketing research.

COLLECTING THE DATA

The most widely used formal marketing research is, or at least should be, survey research. Typically, this may be designed to find out, descriptively, what are the participants habits, attitudes, wants etc. It is simply based on asking the participant, the respondent, a number of questions.

Thus, the classic device used on such surveys is the questionnaire: a pre-printed form on which the interviewer, or the respondent, fills in the answers to a series of such questions.

If a key question is not asked it will not be answered. Even then, many words used by researchers and their clients, including those used in their daily language, may be strange to the respondents they are testing; particularly where these are less well educated respondents. In addition, if the form of questioning is too complex (or, on the other hand, too vague) it, too, may elicit confused answers. Above all, however, the most basic fault of much research is that, as a result of bad design, it plays back the answers that the researcher expects (or even wants) to hear.

The next stage of the research is when the army of interviewers descends on the unsuspecting public. There are, however, a number of possible methods of contacting respondents. Mail is the cheapest solution and large overall samples can be used, allowing investigation of small markets groups - especially in industrial markets - still within acceptable statistical levels. But, in many respects, it is the least satisfactory one. The questions which can be asked are necessarily simpler and the questionnaire shorter; and it must be particularly well designed - to keep the respondent interested and motivated to reply. More fundamentally, the response rates may be so low that their statistical validity may be questioned; since it is arguable that the majority, the non-respondents, might behave differently to those who have responded. It is worth noting, however, that response rates in fact may not be significantly worse than those for the typical face to face interview. The only difference is that in the latter case the 'non-responses' are not know (or at least are not recorded). Some - very carefully managed - mail surveys may, therefore, offer a viable alternative to face to face approaches.

Alternatively, the interviewer might use a telephone to contact respondents. It is a very fast survey technique; so that results can be available in a matter of hours, and hence it is often now used for those opinion polls where time is of the essence. It is also relatively cheap (and, hence, often affordable even in industrial markets). On the other hand, the interview can only last a short time and the types of questions are limited. Even so, it is probably under-used as a technique; since it is relatively cheap and quite flexible in use.

Personal interviewing is the traditional (face to face) approach to marketing research, and it still is the most versatile. The interviewer is in full control of the interview, and can take account of the body language as well as the words. It is the most expensive however, and is dependent on the reliability of the interviewer; and on his or her skill in the case of some of the more sophisticated techniques. This means that the quality of the supervision provided by the field research agency is critical; which may be a problem where so many organisations now place the emphasis on cost-cutting.

SAMPLES

The basic principle of sampling (derived from statistical theory) is that you can obtain a representative picture of a whole 'population' (the term used by theorists to describe the total group of people, or objects, being investigated) by looking at a small 'sample'. The number is usually, in this context, only a few hundred - but correctly handled it way still may give an accurate picture of an overall population of millions. This applies to testing grain sold by the farmer just as much as it does, here, to market research. It is, needless to say, a very cost effective way of obtaining information.

Samples are a somewhat academic subject. They are, though, important in terms of understanding the accuracy (in terms of just how representative of the overall population they may be) which may be placed on the results which emerge; and (as they represent a large part of the costs involved) offer a good indication of the quality of the work being carried out.

There are two main ways such a sample may be chosen. The classically correct method is to select a sample at random. The list of the total 'population' to be sampled is chosen. Usually, for consumer research, it is the electoral register. This list is then used as the basis of selecting the sample; most rigorously by using tables of random numbers, but most simply by selecting every n'th name.

The great advantage of random samples is that they are statistically predictable. Apart from any questions over how comprehensive the original lists are, they cannot be skewed (that is biased). The major disadvantage is that they are usually more expensive (and, in any case, the necessary lists may not be available). Accordingly they may be less frequently used for commercial work (except that based upon mail questionnaires).

Quota samples, on the other hand, aim to achieve much the same effect  by asking the interviewers to recruit respondents to match an agreed quota of sub-samples; which is supposed to guarantee that the overall sample is an approximately representative cross-section of the 'population' as a whole. The interviewer, by means of knocking on doors or standing in a busy street, is required for example to select certain numbers of respondents to match specified age and social categories.

This clearly may be subject to 'skew', selecting only the more accessible; those who make a habit of visiting their local high street for example, and excluding the more elusive elements of the population. It is also difficult to apply rigorous statistical tests to the data. There is, in addition, a hidden problem of non-responses. The levels of these may be as high as in mail surveys; but, as they are not recorded, they are out of sight (and hence ignored)! Quota samples are, though, significantly cheaper than using random samples; so they are the approach most frequently chosen for commercial research. It has to be said that, despite its apparent theoretical shortcomings, they often work well; and have done so, with documented results, for several decades.

As quality does, though, depend very directly upon the quality of the interviewer, and in particular on the quality of supervision, it is the approach most likely to suffer from the shaving of quality to achieve cost savings. At the worst level, badly controlled, it may all too easily degenerate into 'convenience sampling'; which is a polite phrase for interviewing whoever comes easiest to hand - and is not a genuine form of sampling by any standard!

There are some nice statistical calculations which will show you how many interviews you will need to conduct (in theory on the basis of a random sample, but the same figures are often applied to quota samples).  

Sample sizes as low as 300-400 are often found to be acceptable. The point to notice is that you rarely need more than 1,000 and, much more important, you can equally rarely expect to get accurate results with less than 200 respondents.

Marketing research in the industrial goods area is typically less involved with survey research. On the one hand, the output of statistics about the 'average customer' may be less useful - where each customer's needs often have to be considered separately (the value of their business justifies this, and the contact with the salesperson makes it a possibility). On the other hand, the difficulty and cost of conducting such survey research on industrial customers is much higher. For one thing, the 'lists' (which define the population from which the sample is drawn) are often not available, or are inaccurate and incomplete. The result is that desk research is even more prevalent, and even more important, than in consumer marketing research; and this is often conducted by experts rather than by the individual manager. Much of the survey work which is done tends to revolve around in-depth (unstructured) interview with relatively few respondents. In any case, the total population, the 'universe', may be just a few hundred organisations (especially if consideration is governed by the value of purchases). The interviews are usually conducted, by experts, on senior managers (and the views of the organisation sought rather than those of the individual).

ANALYSING THE RESULTS

The statistical data collected by surveys can be analysed in a wide variety of ways (as can that which is already available in the results emerging from desk research; and, in particular, those from performance data). Increasingly these analyses use the availability of massive computing power to cut through the superficial results to try and see what lies beneath. The mathematics of these various techniques are beyond the scope of this book; and the practical skill needed is that of finding the best expert to implement them - and knowing how much reliance to put on his or her judgement.

To give a feel for what is involved, however, some examples of the techniques now used include multiple regression analysis which is a complex and sophisticated statistical (computerised) method which attempts to determine the structure of relationships between factors where there are more than two of these factors involved. It is concerned with establishing what the contribution of each of these factors, the 'independent variables' (for example price and advertising), make to the overall results or 'dependent variables' (for example sales or customer attitudes). Factor analysis is another sophisticated technique which is used to group together 'related' variables (by the detection of related patterns in the data, usually concerned with buying behaviour). These may, superficially appear to be independent but in fact can be shown (at least, statistically, in terms of the dependent outcomes) to be highly correlated. That is, there is an underlying relationship which means that they behave in much the same way and have a similar impact on the final results. It is primarily a tool used to reduce a large number of possible variables to a fewer, aggregated or summarised, number - which can be more easily handled. On the other hand, using cluster analysis factors can be found which strongly differentiate, for example, certain customer groups from others; so that the 'cluster' is isolated from other clusters, whilst being internally alike. It is a particularly important technique in the case of segmentation; where the aim is to split customers (and hence markets) into clearly differentiated groups. As such it is one of the most powerful analytical techniques available to the marketer.

REPORTING THE FINDINGS

The final stage is to report the results to all who need to know them. This process may require more effort, and be more important, than the simple clerical task it superficially seems to be. For one thing, it requires some understanding of to whom the results may be useful; and, indeed, important results may have relevance to managers throughout the organisation. Equally important, the language of the report may need to be translated for these different audiences. Very few of these managers will understand the terminology of market research (and, rather more important, the limitations on what conclusions may be drawn from the figures). The results will, therefore, have to be conveyed in terms which are meaningful to these managers.

This does, however, pose some problems. It is inevitably a process of simplification; and may accordingly result in the loss of some meaning (and some of that which remains will be subtly changed). The favourite approach (at least in presentation to top management) seems to be that the dry statistics (which have probably already been considerably simplified) are illustrated by verbatim quotes from individual respondents. The particular, and very real, danger here is that the senior management, being thus confronted (and unversed in market research skills), will remember the most impactful comments (particularly the ones that reinforce their existing prejudices) rather than the boring statistics.

Many managers find themselves on the receiving end of research reports, often those deriving from marketing research. Almost as many managers, on the other hand, are poorly trained in the skills needed to make sense of such reports. As a result, they tend to read the related conclusions uncritically, accepting (or sometimes rejecting) them at their face value; usually on the basis of what they think of the researcher presenting them, or whether these results confirm their own existing prejudices.

What, then, should you do?

To a certain extent the answer to this question must depend upon the specific circumstances; what is contained in the report, and what you will want from it. Even so, there are a number of initial guide-lines you might wish to bear in mind;

RELEVANCE - before you even look at the first page of the report, you should ask yourself whether the subject is relevant to your specific needs. Can you afford the time to study it. Fortunately, this can usually be deduced from a quick skim through the summary; coupled with an understanding of where the report has come from and why it was produced.

RELIABILITY - this is perhaps the most important question, but the one which is least often asked. How reliable are the results reported? What weight can be put on them and on the judgement of the researchers; and, probably even more important, on the experts who are likely to be recommending some form of action to be taken on the basis of the findings. Be aware that personal reliability, in a social context, does not necessarily underwrite technical reliability, in a business context (and, indeed, usually tends to blinker those involved to even the most dubious shortcomings of the material).

A more rigorous approach would be to examine the methodology (the questionnaire and sample design, say) since this is likely to give the best indication of the 'quality' of the work.

ACCURACY - having established that the material is both relevant and reliable, the next step is to decide what the accuracy might be, The researcher should tell you this, but all too often this is a technicality which is buried deep in the technical appendices which never reach the general reader. As I have already indicated, in marketing research the answer can normally be deduced from the sample size. If the sample size is over 500 the results are likely to be accurate to within 2 - 3% (always assuming the research has been well run). If the sample is over 1,000 it may be within 1%. Below 100 though, as many of the more dubious pieces of 'quantitative' research may be, any statistical accuracy may be almost non-existent!

BIAS - most research reports contain bias, conscious or unconscious. It is very difficult for even the most professional researcher to remove all of his or her biases; and you would be wise to assume that the material still contains such elements of distortion. This bias may not be without value. The best research starts with a strong thesis as to what is likely to be found. Whilst this will inevitably colour the results it also ensures that the research is focused and provides meaningful insights; just so long as you recognise what that focus (that bias) is and you discount its effect on the results. To do this you need to know what assumptions (biases) are implicit in the work. Once you have written down what these are (or at least what you might reasonably believe these are), and hence have made them explicit, you will be much better placed to understand the real implications of the figures.

SCOPE - the final question, before starting on the main body of the report, is what range of information might it provide (often researchers only report the results that interest them - where you may be able to ask further questions of the material about topics which interest you). This scope can be found most easily by looking at the questionnaire, to determine exactly what questions were asked (the results of which will be held on a database somewhere, even if they are not reported upon, and can thus be subsequently analysed in whatever form you wish).

It is only at this stage that you should start to read the main body of the report. The reason for this circuitous approach is that without it the writer of the report probably has gained your uncritical attention, and the opportunity to manipulate your judgement, as soon as you start to read his or her introduction. A good writer can permanently warp your judgement long before you get to any of the details listed above!

The subsequent stages are;

SUMMARY - the first element to be read should now be the summary. This will provide the context for understanding the detailed results. It should, though, be read in that spirit; not as a list of proven facts, and certainly not (if the research is important) instead of the detailed material.

DETAILED RESULTS - these should then be examined; preferably by looking at the original analyses (tables) as well as the written interpretation - since any interpretation distances you from the facts. They should, once more, be examined critically. Do you agree with what the researcher has deduced from them? There is no reason that any researcher should be more capable than you - in terms of this analysis. He or she, it is true, should be more experienced in the marketing research techniques involved; but you should be more experienced in the field being researched.

'Challenge' the results! Most times you will find you agree with the researcher. Occasionally, however, you will find new information which is worthwhile; and in all cases you will better understand what the research is about.

THINK OUT YOUR OWN SUMMARY - only then can you think through (and put down on paper) what the key results are, in terms of what affects your own work. It is a long process; but if the research is worthwhile then the effort you give to it should match its importance - and, hopefully, your subsequent decisions will be correspondingly better informed.

PERSONAL RESEARCH

So far, in the main, we have looked at the formal research systems; in this context even desk research can be seen as formal (even if it may seem to you to be very disorganised). On the other hand, informal contacts (typically based on the spoken word as opposed to the written word) are the staple diet of management. But, if it is difficult enough to abstract useful data from written paperwork, it is not usually even considered worthwhile mentioning oral data. Yet this probably represents the most important source of data available to any manager; particularly to one involved in the sales and marketing functions. In any case, every meeting, be it a formally organised meeting of a group or an informal one between two individuals, is potentially rich with useful data. To take advantage of this a number of personal techniques need to be employed, which are normally not mentioned in the context of marketing research;

PERSONAL RESEARCH TECHNIQUES

As with formal research the key to personal research is questioning. Yet again, you only get the right answers if you ask the right questions.

QUESTIONING

This might seem easy to achieve when you are asking the questions for yourself, but most people simply are not very effective at asking such questions. They ask very specific (closed) questions which tend to narrow discussion; and, in particular, tend to confine the discussion to the areas set by their own personal preferences or prejudices. Much more useful are those (open) questions which allow the person being questioned to adopt a wider view. These questions (of which the simplest - Why? How? What? - are the most powerful) encourage the speaker to say what he or she considers is most important about the topic. The listener can then gain the most benefit from the speakers knowledge and expertise. Later in the conversation directive and then closed questions can be used to steer the conversation to the topics of greatest interest to the listener.

They also seem to be the most difficult for a manager to ask; perhaps because they are not so obviously leading directly to the answer that is wanted. But they are the key to unlocking the tongue of the person facing you. If the conversation proceeds with very short replies (and particularly just yes or no), it is likely that you are not using enough open questions; and may be missing the real issues. The more open the question the better. The most powerful question is quite simply 'Why?', often closely followed by 'How?'. In practice, open questions come naturally if the questioner is genuinely interested in finding out what makes the prospect's business tick.

A particular technique, used for example by skilled researchers, is 'laddering'. 

Unfortunately, in most normal discussions (as opposed to its use in research), it is a very aggressive technique and must accordingly be used with great care.

A slightly less stressful and equally successful, if little reported, approach is 'rambling'. Eden, Jones & Sims[2] report that 'Probably the most obvious method for getting to know about the view a person has of the problem is to give him the time and space to 'ramble' around his subject.' This can be an enormous strain on the listener, for it is difficult not to interrupt. But, in the same way as interrogation methods often depend upon allowing the prisoner to ramble so that they inevitably give away 'too-much' of themselves, it seems possible that:

Even if managers do ask the correct open questions, they often undermine the progress by stopping the speaker in mid flow. The natural accompaniment to an open question is silence. It is, though, a surprisingly aggressive technique, and you should not make it too obvious - it is best just to look very thoughtful. The person you are questioning will eventually feel obliged to talk, and usually what he or she then says is especially enlightening (since he or she too will have had time to consider).

As indicated above, closed questions, typically requiring the answer "yes" or "no", have (justifiably) received a bad press. But it is still necessary to use them quite extensively to clarify points. The problem only comes when they are used instead of open questions. In many situations by far the most important closed questions (and arguably the most important questions of all) are those where you check for agreement. As the discussion progresses, it is imperative that you establish whether or not you are taking the other person with you; or, as is all too often the case, is he or she politely acting out the role of audience to your orator?

LISTENING

Just as important a skill as questioning is listening. Many managers are too busy trying to put their own view across to hear what is being said in reply; and thus they miss much of the key data in such conversations.

or the listener will soon become a sleeper.

It is conventionally reckoned that a good questioner should spend two thirds of the time listening and only one third talking. What is important, though, is how you use that time. It is the quality of the listening (which has much to do with how you analyse what you hear) that is as important as the quantity

UNDERSTANDING

Thus, hearing, and even listening, is not enough. The key to questioning by managers is understanding. This is a process to which the main contribution must, of course, be made by what the person being questioned says; though it should be noted that this may include what he or she said in a number of previous meetings as well as in the current one. But it will also include all the other evidence you have unearthed. Put it all together and, hopefully, you will be able to complete the jigsaw.

Understanding of informal communications is, therefore, a cumulative process; that may span several discussions. It is a an important skill for managers; yet it is largely ignored by management educators.

RECORDING AND ORGANISING THE DATA - even before the advent of the Filofax and its electronic equivalents, managers used to carry note-books in which to write notes on their various meetings and conversations. This is an excellent habit, and forms the basis of the process needed to convert these conversations into a form which is then retrievable. Ideally, such rough notes should then be converted to formal (if brief) reports which then go into the main filing system. As with the clippings file, however, it might be productive to file them every two weeks (once again two weeks in arrears) so that you can begin the process of editing them; and tracking new developments.

PRIORITISE AND TRACK CHANGE - finally, as with all good marketing research, you should do something with these gems you have been collecting. The first step is (as it was with the clippings file) to group the ideas. Thereafter, as this internal data will already be quite focused, you will be able to prioritise it so that you can give the most important topics the attention they deserve. Most important, though, query anything which emerges as different to what you might have expected. Why has it changed? It is all too easy to ignore such internal 'weak' signals, but much of management is about dealing with change; and the earlier you recognise that changes are emerging the better.

You will, of course, realise that the latter paragraphs on taking notes and organising them is a counsel of perfection! I have rarely met managers who do this; though those who have done so have been amongst the most effective I have known. Most managers carry this data around in their head. They must, therefore, find some other process of carrying out the above instructions. If their head then starts to hurt then perhaps they too should invest in a notebook!

To complement these informal meetings and conversations which fortuitously provide you with useful 'research' data, you can also use other informal sources. Not least of these is the grapevine; the collection of rumours an gossip which permeates most organisations. It frequently acts as a form of hidden neural network for the managers as much as the employees. It is often better informed and more accurate than the formal information channels; so the tighter you can tap into it the better. It is often paralleled by a network of 'fixers', managers who, perhaps because of their length of service or just their travels throughout the organisations, seem to have contacts everywhere that matter. They know what is happening and what is to come, and in their spare time resolve the cross departmental problems which others don't even know exist. The thoughts of these individuals should also be grist to your information mill.

The penultimate technique in this section has the simplicity that is the hallmark of the Japanese, and it is their favourite approach:

It has none of the statistical validity which survey research enjoys, and even desk research can often lay claim to. Yet better than anything else it conveys the flavour, the essence, of what is being studied. If you want to understand Toyota you can spend months of desk research reading the hundreds of papers which have been written about its efficiency, or you can spend half a day watching the confident grace with which the workers on its production lines assemble cars; as, decades before, Mr Toyoda himself learned his lessons by similarly spending time, three months in his case, watching American workers going through their, less graceful routines.

The essence is experience; assimilating what is happening - what is really important to the product or service.

This, then introduces the final technique, which (for those managers which can use it) brings together all the research data in the most practical form possible. For many, if not most managers, the marketing research data remains as so much impersonal data lying on hundreds of pages of tables in dusty files. Just a few managers, however, bring it alive by assimilating it into their everyday view of their business life. 

Needless to say, the technique of walkabout is the most useful of all in this process, because it gives the best 'feel' for what the key elements are.

It is rather like the actor who uses Stanislavsky's Method to bring the character he is to play inside himself. He does not play the part, he lives it. Thus, the synthesising manager should live the part of the customer.

The great benefit of this is that the manager does not have to search through the vast collections of data to know what the customer's reaction would be to any of the several dozen decisions which may be made in a day - that would be unproductively time consuming, and is precisely why the research results gather dust. Instead, he or she can draw upon their inner model to instinctively 'feel' what the customer's response will be.

This is a difficult process, especially when there are a number of different customer groups to assimilate in this way but, like the actor, the manager can - after considerable effort - usually achieve success.

Perhaps, though, the analogy should not been Stanislavsky but Brecht, His acting technique, of alienation, used the former's approach as the starting point; but then forced the actor to extract the essence of the character - for this simple essence was more powerful in communicating with the audience. The most skilled marketer will, similarly, extract the essence of the customer; for the most powerful marketing, too, uses the simplest messages.

FRACTURES AND MARKETING RESEARCH

One final footnote concerns areas where marketing research may not be particularly helpful. When there is a major discontinuity in the overall environment, described by Gareth Morgan[3] as a 'fracture', this changes all the factors to such an extent that market research may be largely useless.

In this situation most marketing research is meaningless since it essentially measures the historical position, unearthing data on what has gone before. The discontinuity means that the future is decoupled from the past - it means the future will be different. Even consumer research will be largely valueless when this happens, since the consumers asked their opinions will not know enough of the new developments to answer the questions accurately, but will base their answers (incorrectly) on their existing perspective - and this will, again, not offer the researcher a valid view of the future.

 John Stopford[4] makes the point that the really significant new products have not emerged from incremental (and marketing researched) changes; but have been genuine innovations. Perhaps more important, for the readers of this book, is his associated comment that the organisations which have survived such fractures (that is major changes in their environment) have been the ones which have thrown out their standard operating procedures manuals - and the sooner they did this the better they fared. This may say relatively little about marketing research, but it says a great deal about the value of outdated theory as compared with practical flexibility in coping with change!

Fortunately, for most managers, such innovations are very rare[5]; though they have recently preoccupied some theorists. Notwithstanding the comments of these theoreticians, some of whom have even suggested that the possibility of these discontinuities undermines the value of marketing research as a whole, obtaining the best possible information about the outside world is still the key to the great part of successful marketing. The more you understand about the environment in general, and customers in particular, the more effective your marketing is likely to be.

[back]     [home]

hits