In the second of our continuing series delving into the ESOMAR archive of historic papers from our 65 year history we turn to the ESOMAR Congress 1975 held in Montreux. A year that saw the end of the Vietnam War, Muhammed Ali’s Thrilla in Manila and a young Bill Gates founding Microsoft. Leonard R. England’s paper on quality in research shows that although methodologies may have evolved, issues surrounding quality and the relationship between researcher and client are ever-present.

Two Way Quality

Leonard R. England 

One of the very few advantages of a major recession is the extent to which it drives industries to reconsider first principles.  In this respect, market research is no exception. Throughout the United Kingdom (and probably throughout Europe) 1975 has been a year in which attention has been switched, at least in part, from technique to the use of market research.

Published papers have stressed the need to understand and to apply findings whatever the methodology and, while this has been a constant cry from the earliest days, more people now seem to be listening. No matter how elegant the design, how ingenious the solution, research which cannot have practical applications will not get commissioned.  It was eleven years ago that a chairman of the British Market Research Society said that the only graph of ultimate interest to a researcher was a sales graph which started to climb soon after research had been completed.   The speaker was (and is) very concerned with quality and technical development; but he has always got his priorities right.

Now, despite this acceptance of action, there can be little doubt that a very great deal of research does not get used to the end which it should be, or to the extent that it should be. Nor can there be much doubt that the reason for this is that those who make the really vital decisions either are not informed of the relevant findings of research (or at least are not informed of them in a way which makes them appear relevant) or have doubts as to whether what they are told is accurate. Both these faults are due to bad communication. But whose fault is this?

At this point the in-fighting in market research breaks out. There is no need in this paper to repeat a series of very familiar arguments. They centre basically on the function of the market research company – the doer of the research rather than the commissioner. Nobody can deny that it must be the company employee ultimately who must ensure that senior management knows what is going on and what it all means. But some see any attempt by the market research company to provide more than computer sheets as unwarranted interference, while others argue that the outsider is in an unrivalled position to be impartial and to stand back, and also that experience over many fields of research and its relevance to marketing must imply the ability to contribute to an important degree.

It is difficult to see logical reasons why this argument should exist at all. Clearly, the emotional overtones involved are extensive. The market research company has many examples of good research getting no further than a junior brand manager frightened of its implications, while the client has equally specific case histories of results being dismissed simply because of absurd statements to management made by an outside researcher with no knowledge of internal problems.   Both are frightened of the power of the other. But the logic of the situation demands that, at the stage of completion of a survey more than at any other time, there is a desperate need for research agency and the research company executive to work together.

The principle should, in fact, be extended.  We tend to, for our own convenience, divide survey work into “syndicated” (where results are available to all who wish to buy) and “ad-hoc” or “tailor made”.   The latter is so often interpreted as applying to a single survey when it really should be considered in terms of a single problem.  No major marketing development is likely to be solved by a single ad hoc survey, however much syndicated research is also brought to bear.  But clients and researchers often drift apart between one project and the next, however closely they have worked on a particular issue.

If the theory is impeccable, what of the practice? This paper is, in fact, intended to be the case history, not of a survey, but a client-agency relationship. It aims to show its advantages to the quality of research – and quality in the production of research which is used rather than being beyond the criticism of fellow practitioners.

Some background facts are necessary. The client concerned is the paint manufacturer Berger, a UK subsidiary of the German firm of Hoechst. Although they are a major company, they are not the biggest in the UK, nor is their research budget disproportionate to their size. They use other companies as well as EG+A for other parts of their research, and they are very far from being the biggest clients  of the research company. There are, therefore, no special circumstances involved,  and neither client nor agency is in any way dependent on the other on the one hand for all research,  on the other for all profit.

The contact began in 1971 when EG+A received a formal brief together with a note to indicate that three other companies had been contacted and that those who did not get the job would still get a token fee. Now this is still a very unusual arrangement in the UK and one which is not universally approved of even by research companies themselves. But whatever its weaknesses it indicates a client who appreciates that quality of thinking and expression begin from the very first proposals and who does not treat them just as a simple costing exercise.

Since that time EG+A has carried out some 26 different projects for the same client. The largest cost £12,500, the smallest under £100. While they have covered a very wide range, they have concentrated on three main areas:

The painting process – this is far more than the profile of the paint user and the paint buyer.  It embraces many other issues such as the importance of decoration in the whole concept of care of the house.

New paints – technological developments have resulted in paints which are more efficient if they are rightly used. Will this happen, and are the qualities those which are wanted by users in the home?

Brand names – Berger have always had two brand names which the majority of users do not realise come from the same company. What is the effect of trying to join these together?

As well as covering a wide range of problems, the surveys have employed a wide range of methods:   group discussions, hall tests, product tests in the home, advertisement checks – above all, a very elaborate form of cluster analysis based on a large random sample.

There is, however, a real danger that the very fact of itemising surveys in this way implies that each was considered on its own.  It is true, of course, that most of the work on new products was carried out for the laboratories and while it was continuous over almost the whole period in which we have been working, the test of each new formula was, to some extent, a study on its own. On the other hand, almost all the rest of the work has been able to stand on the shoulders of what has gone before.

The most obvious example of this was the two major random surveys carried out in 1971 and 1973. The first of these was aimed at providing a background pattern of facts and data about painting and decorating and while the information was of such value that much of it was later brought up to date, in another it failed completely to distinguish reasons why people bought one brand of paint rather than another. Extensive reanalysis plus brainstorming sessions (in which the advertising agency J Walter Thompson also played a major part) seemed to make it clear that the premises of the first survey were too narrowly based. We should not have been studying painting in a vacuum but relating it to the whole problem of what the house meant to the occupier and what the occupier saw as the importance of his “spare time” – did he like painting the house or do it because he had to? The follow-up survey filled in the gaps most satisfactorily. It would not have been possible had it not been the second stage in a continuous series of surveys.

In a similar way, our work on brand names and on the way in which these are printed on the paint containers springs from a whole series of surveys  on what people want on the labels (instructions,  reassurance and so on)  and on the relation of the name Berger to the brands Magicote and Brolac.  Each survey starts with a great deal of knowledge based on previous research.

Thus, the first signs of a two-way quality relationship is that it should imply a continuous process, seeing each survey as a development of what has gone before rather than an individual effort.

This is so obvious that perhaps it should not need saying. But in the UK at least one of the most depressing signs of lack of quality is the extent to which so many of those who commission market research feel that every single new survey, whether or not it is an extension of a previous project, needs  a “fresh mind”.

Recently EG+A received from a client three surveys dated 1971, 1972 and 1973, each done by a very reputable organisation, each a perfectly competent study.  But clearly more recent survey companies had never been shown the previous work and, for all practical purposes, had repeated the same job as had already been done. The waste of research time within the company and the loss of technical know-how from the research company in these circumstances defies belief. The lack of this continuous contact has another very unfortunate implication.  There may be many problems in the relationship of an advertising agency and its client, but it does not stop-start between each campaign. The people in both companies get to know each other and therefore are able to be frank with each other.  One of the delights of EG+A’s relationship with Berger is that, when we are badly briefed or asked to do the impossible we complain bitterly rather than making polite noises:   and when we do a bad job in forgetting the vital question or going unprepared to a rehearsal of an important presentation we are rightly chastised. But things are put right and the sun shines again.   We can argue, disagree, lose our tempers without believing that by so doing we lose the client. We can become passionate about getting the best answer, whoever produces it.  And this too makes for quality research, This understanding of each other implies quality in a different context. To one who was working in market research in the very early days, one of the great disappointments of the contemporary scene is that fear of being wrong has driven out the cheap-and-cheerful survey when this is all that is required, Very often, of course, it is not all that is required, But equally often, provided that all sides know the limitations, it can be of the greatest use.

For known clients EG+A often carries out a mini-survey, the “mini” implying not a short questionnaire but a small sample. For if there are a whole range of problems involved about which little is known, it seems best to get some feel from the public rather than relying on the inevitably partial views of the client. And a half an hour interview with fifty or a hundred people can elicit factual data in much the same way as a few group discussions can provide qualitative material.

But we never recommend this technique to new clients, however much we feel it could be of value, for the simple reason that we have seen too often how data which is meant to do no more than provide the guidelines, within which a technically sound survey can be devised, comes to be regarded as an end in itself.

This is only one example of the short-cuts which can be made which save both time and money once a two-way understanding has been built up. Formal reports can be done away with or at least reduced to no more than an outline; surveys can be halted or changed in midstream because the evidence already emerging is enough to allow movement to the next stage:  the same survey framework can be used to cope with two problems.

These and half a dozen other simplifications which we have used could be faulted by the youngest research trainee and, on their own, they would be indefensible. But taken as part of an ongoing survey plan, they have a major part to play a part which is only possible in the context of two-way trust and understanding.

At quite the opposite end of the scale, such a trust builds up acceptance of the argument that, on other occasions, no short cuts are possible. The main cluster analysis survey referred to earlier at first sight seemed, to top management, highly expensive particularly as there was a risk that it would prove no more likely to solve the central problem than our earlier efforts.  And yet we stuck to our guns, insisting on a random sample, insisting on very elaborate analysis and on a very long  (and expensive)  “wind-down” time to allow results to be interpreted and then to be put in marketing language. Now this survey and all others form a client data bank of which use can be made not merely in the form of the published computer sheets but of special extra analyses. We do not have the experience (which is not unknown) of fieldwork being commissioned to help solve a problem the answer to which already lay in punched cards of an existing survey were they only to be analysed in a different way. Indeed, for one recent Berger survey an additional objective was built into the second leg of a two-part survey: the purpose of the survey was recognition of new containers, but the by-product was fresh data on brand duplication and rejection. Also springing from this same survey – and only possible because of the constant dialogue between research executive and research company – came a simple method of testing rapidly, as occasion required,  developments on the merging of brand and manufacturer names. The survey was intended solely to be a single study of a specific problem. But it is likely now to serve as the baseline for a series, perhaps stretching over a number of years.

We have so far considered “two-way” quality in terms of one person in the client company – the research executive.   But we have already indicated our awareness of how important it is, if research is to be accepted, that this research executive is accepted by top management.  And in this instance both “sides” believe that credibility is increased if the research company acts as part of a research team, integrated rather than ostracised.

This does not, in this instance, mean  (although there is much to be said for the approach)  that EG+A sits in at client policy meetings when research implications are being translated into action points.   But it does mean firstly that presentations of survey results are so organised as to involve a very wide spectrum of staff from the client, and secondly that they are carefully rehearsed as a joint exercise between company staff, advertising agents and research company.   Each take their share, the whole implying a study of not merely what has been found out but what can be done about it. Indeed, one such presentation was exported to an international conference in Australia.

But presentations are, by their nature, ephemeral. The main findings may be remembered but little else. And although all main surveys are likely to have a report with both summary and full text, it is not realistic to assume careful readership by all those who someday might want to know something about some of the findings: nor do either of us believe in human nature sufficiently to argue that anybody in this dilemma will immediately ring up the research executive to ask him what data is available.

What has therefore been provided – and we believe that this is an unusual development even of a “quality relationship” – has been a beginner’s guide to research findings.  All consumer research for the last five or so years has been summarised into one document divided into four or five simple sections – the home, the decorating process, the paint and so on. Within each section there are further divisions mainly of the who, when, where, how type, and one double page  – never more – is allocated to each sub-division. On the left hand side is a summary of the main findings as they apply to this sub-division often culled from a number of different surveys; on the right hand side are very simple tables with certain essential data and also a cross-reference to the relevant surveys which are kept by the research department. The whole is carefully cross-indexed and is indeed partly illustrated by cartoons, all to encourage use rather than to discourage it. Each year it is updated, and wide distribution implies that most marketing and sales executives can no longer have an excuse that research findings are difficult to find.

These are some examples of what we imply by two-way quality – an ongoing relationship between client and agency allowing constant interchange of ideas and the building of survey data by devising each new project in the light of what has gone before.

Now clearly there is nothing very new or unusual in this. There can be few agencies in any part of Europe who do not have at least one client with whom they work in this way, and it would be odd if, in consequence, they could not give examples of other ways in which surveys are improved by such working together. But it is equally likely that, in all agencies, this two-way quality exists for only a minority of surveys. There could be a number of reasons for this:


  • The agency does not bother with client continuity and constantly changes executives.


  • The Client regards his market research department as part of executive training from which the efficient quickly pass on.


  • The agency believes its responsibility stops if it effectively does what it is told.


  • The client expects his instructions to be carried out to the letter without argument.


  •  The agency feels that a special technique developed for one survey is worth selling for all sorts of other ones.


  • The client believes in shopping around to get new ideas and to keep the agency on its toes.

These views are undoubtedly very common.  One wonders whether either “side” – and the importance of not seeing the relationship as “sides” is inherent in all that has been written here stops  to realise how much they are missing by such an approach; for they imply a contempt for the fact that research implies thinking as well as doing,  creativity as well as computer, a lack of belief that a research study which is acted on is more profitable than one which is not, not merely in terms of job satisfaction but ultimately also in hard cash.

The client is, of course, the buyer and it is in his hands, up to a point,  to control the situation by an insistence of his selected agency maintaining, as far as is possible, an attitude to survey development similar to that described here.  But the market research executive himself has no control over his own management in the sense that he cannot prevent his own promotion if market research is seen as only one junior step on the ladder. Nor can he prevent instructions always to obtain competitive quotes or frequently to take a “fresh look” at a problem.

The author of this paper has discussed at length the vicious circle in which market research now finds itself. It must convince top management of its worth in problem solving: but how can it as long as market researchers are seen as inferior and their advice is ignored? Surely it is in everybody’s interest to press for the type of relationship described here? And surely the agency can help the client’s own research department by stressing in every report and in every presentation the continuity of research development or any specific project by doing everything in its power to strengthen the department’s own concern for recognition while not, at the same time, squeezing each new assignment into the most convenient from for staff-loading and machine-processing.

Once successful companies can be shown to owe at least some part of their success to market research, the path will be easier. Two-way quality should help this process.

First published and presented by Leonard R. England of England Grosse & Associates at ESOMAR Congress 1975, Montreux