Chapter 1 TRUSTED ADVISOR: HOW IT HELPS LAY THE FOUNDATIONS FOR INSIGHTS

Rajiv Grover, University of Georgia Marco Vriens, Microsoft Corporation

As alluded to in the Introduction section of the book, the trusted advisor status enables the market researcher to delineate the right problem definition, which, in turn, ensures that insights will be generated. In this chapter, we develop two processes for arriving at the right problem definition under the trusted advisor model of engagement. The first process applies to a class of problems which we term as decision-making problems. These normally stem from or surface with a symptom. The second process pertains to the market-learning class of problems that begin with a fuzzy, broad-based need for information. Since for both these processes, it is essential that the market researcher serve in the capacity of trusted advisor, we first establish the need for trust in the marketing research context, build a model of trust, and develop the attributes of a trusted advisor.

The Need for Trust in the Marketing Research Context

Research has shown that trust between two parties results in better quality of decisions and, therefore, in the performance of the dyadic unit. Trust engenders better usage of marketing research results and mutual satisfaction with the relationship between the researcher and the client. When trust exists between two parties, the following factors facilitate the resulting positive outcomes: (1) open communication, (2) lesser blame attribution and conflict, (3)

acceptance to influence and control by the other, and (4) exhibition of positive behaviors beyond the call of duty (Zand, 1972). Although it may be clear why trust is important between a client and an outside supplier, its significance in an in-house context might not be evident. Even when internal researchers conduct the research, their clients in the same company are exposed to risks. And when there is risk, there is a need for trust (see, e.g., Mayer, Davis, and Schoorman, 1995). In the context of marketing research, there are two types of risks that clients face, which might be termed as information usage risk, and personal exposure risk. Information usage risk is that which results from the use of information provided by the market researcher, and it exists because information is inherently never fully accurate or complete. In the context of market research, inaccuracies of information stem from research design as well as execution of the design. Every research design has some potential for bias, and further errors occur during execution of the research. For example, the design of a mail survey may be tainted by any one or more of the standard sources of bias: leading questions, questions that result in alternative interpretation, or questions that test the limits of the recallability of respondents. Even if these design biases are addressed, biases in execution, such as in sampling and or that resulting from nonresponse, might still prevail. In other words, the results of a particular market research project might be idiosyncratic in the sense that the same problem researched with another audience or method might yield different results. Besides inaccuracy, incompleteness of information generated by market research is quite inevitable. For making a decision, several aspects of the problem might need to be investigated, but not all aspects might be obvious and/or researchable due to the potential practical constraints. Thus the market researcher might not be able to take those aspects into consideration, ultimately leading to provision to the client of information that is incomplete.

The second type of risk, personal exposure risk, is the risk the client faces of being upstaged by the researcher, and it exists because the market researcher has an edge over the client in having more information about the project than the client. The market researcher can potentially leverage the imbalance of information to gain and demonstrate undue advantage over the client in situations where parties are present to discuss the results with others in the organization. This threat is one of the factors that prevents clients from offering market researchers a seat at the table. The above risks dictate the need for two types of trusts that the client must find in the researcher: professional and personal. The next section discusses these two types of trusts.

A Model of Trust
The construct of trust is among the most researched concepts as it has a useful role in many disciplines (e.g., organization behavior, sociology, psychology). Correspondingly, there are numerous definitions of the concept (see, e.g., Bigley and Pearce, 1988; Mayer, Davis, and Schoorman, 1995)1. From our perspective, we loosely define trust as the willingness of a client (trustor) to be vulnerable to the actions of a market researcher (trustee) based on the expectation that the researcher will perform a particular action important to the client irrespective of the clients ability to monitor or control the researcher. This definition weighs in on willingness, whereas other conceptualizations in academic research put forth the concept as a belief, a belief about the exchange partners trustworthiness (e.g., honesty, integrity, capability), or as an expectancy, an expectancy that a person can be relied on. From a practical point of view, these definitions have the same implications. Literature has also identified many bases of trust. The term basis denotes the origin or source of trust in a trustor. Two of these that have been considered as umbrella sources are

cognitive and affective bases (Jeffries and Reed, 2000).2 A cognitive basis of trust exists when trustors use rational reasons to trust the trustees. Such rational reasons relate to objective evaluations of the trustees expertise, dependability, or work quality. Conversely, an affective or emotional basis of trust stems from a gut feeling of the trustor who relies on traits such as likeability or similarity to judge the trustee. The two types of risks - information usage and personal exposure - and the two bases of trust - cognitive and affective, translate well into the two types of trusts we propose. The first of these professional trust, refers to a clients trust in the market researchers capability to provide as accurate and complete as possible information that is required for decision making. It stands to reason that such trust would be strongly based on the cognitive component. The second type of trust, personal trust, pertains to the trust a client has in the market researcher regarding the latter supporting and promoting the client or the clients cause in various forums. Research on homophily in the sociology area points to the affective basis as the source of personal trust. Figure 1.1 shows the factors that influence professional and personal trust. Professional trust a client has in a market researcher results from the expertise of the market researcher and the clients prior direct or indirect experiences with the market researcher. The market researchers expertise is judged by the client on the basis of the formers qualifications and experience. The other major component that leads to professional trust in the market researcher is the professionalism demonstrated by the market researcher. The indicators of professionalism include dependability, timeliness, congeniality, and tactfulness (Moorman, Deshpand, and Zaltman, 1993). The figure shows that formative indicators of personal trust in a market researcher are qualities such as likeability, similarity, sincerity, honesty, and integrity of the market researcher.3

Qualifications Expertise

Experience Positive Prior Market Research Experience

Professional Trust Dependability, Tactfulness, Timeliness, Congeniality

Personal Trust Likeability, Similarity, Sincerity, Honesty, Integrity

Figure 1.1: Factors Forming Professional and Personal Trust

Constituent Attributes of a Trusted Advisor

From the above discussion it is clear that professional and personal trustworthiness are two key attributes a market researcher needs to demonstrate to be able to serve as and be accepted as a trusted advisor. Among all the formative characteristics of the two types of trust shown in Figure 1.1, in-depth marketing research knowledge is perhaps the most important. That knowledge by itself, however, is not enough to transform the market researcher into a trusted advisor. Beyond marketing research knowledge, a trusted advisor needs to also be equipped with business knowledge. By business knowledge, we refer to knowledge of subject areas such as new product development, branding strategies, consumer behavior, organization

buying behavior, pricing, distribution channels, promotions, marketing strategy, and so forth. The implication is that for researching and acquiring insights into an advertising issue, for example, a researcher must have a good understanding of the advertising discipline. Industry knowledge subsumes knowledge about the organizations products and strategies and about the products and strategies of competitors. Figure 1.2 summaries the key attributes that help a market researcher develop into a trusted advisor.

Business Knowledge (e.g., Marketing, Strategy) Trusted Advisor

Industry Knowledge

Professional and Personal Trustworthiness

Figure 1.2: Attributes of a Trusted Advisor

Having stated the case for the clients need to be able to trust the market researcher, we next propose that the market researcher also needs to be able to trust the clientthat is, trust the client to provide the right information to enable the formulation of the right problem definition. If the client does not reveal all the relevant information, the market researcher will be unable to provide a meaningful solution. Figure 1.3 graphs the total amount of mutual trust between a client and a market researcher. On the far left-hand side, the client uses the market researcher only for data

collection. The market researcher, perceived to have weak industry and business knowledge, is told what to collect, thus making the market researcher an order taker. At the extreme right-hand side, the client has a weak base of marketing research knowledge and is unable to gauge the contribution of marketing research in the yet unspecified needs for information. The researcher is not able to trust the client to articulate the issues completely. In such a situation, the marketing researcher will end up shooting in the dark. In the middle of the graph, both the market researcher and the client have the requisite amounts of knowledge of the industry, the business discipline, and marketing research with the researcher knowing more about the research function and the client more about his/her own function (see, Madhavan and Grover, 1998). The potential of mutual amount of trust is maximum in this situation. The market researcher can behave and serve as a trusted advisor and the client too accepts him in this role. When the market researcher is a trusted advisor, the client articulates the symptoms of the problem to the market researcher, who then interacts with the client to determine the problem and the research approach. In such a role, the market researcher evaluates the diagnosis, designs and executes the methodologies for gathering complete and accurate data, and interprets the implications of the findings for the client. This dynamic scenario is analogous to the patient physician interaction and relationship. The physician diagnoses the problem from the symptoms articulated by the patient and then prescribes a treatment regimen. Just as a physician would resist and refuse a request for a particular drug that is unwarranted, we argue that the market researcher should resist and refuse to take orders from the client regarding the type of data or methods of data collection. Just as a physician would be prescribing in the dark if the patient does not disclose, whether intentionally or unintentionally, all the symptoms and case history, we argue that a researcher would be shooting in the dark if the client is not able to provide or is not forthcoming with all the necessary information. Finally, just as a physician is proactive in

advising check-ups to the patient, we recommend that as a trusted advisor, a market researcher should be proactive in advising clients on when and what to research next. In short, market researcher should strive to develop a relationship with the client similar to that a physician has with a patient. From the above analogy, it also follows that many times the patient (client) might not be aware of what information would be relevant to the physician (researcher) and, hence, it behooves and falls upon the physician (researcher) to ask the right questions when interacting with the patient (client). In other words, client-researcher interaction becomes key in determining whether and how the root problem is identified. The processes for such interaction between client and researcher to arrive at the stage of problem definition are developed next.

Total Mutual Trust between Client and Market Researcher

Market Researcher as Order Taker

Market Researcher as Trusted Advisor

Market Researcher Shooting in the Dark

Weak Market Researcher

Strong Client and Researcher

Weak Client

Client and Researcher Strengths

Figure 1.3: Total Mutual Trust and Strengths of the Client and the Researcher

Processes for Defining Market Research Problems

Marketing problems which can be addressed by marketing research can be broadly classified into two categories decision making problems and market learning problems. Decision making problems are those which when solved, help clients make a decision. Market

learning problems, on the other hand, are those which when addressed, shed light on the companys and the competitors products, existing and potential customers, environment, and so forth. The answers to such problems are not necessarily used for making imminent decisions. There is a distinct and separate process for arriving at the problem definition for each of these two types of problem. Both the processes require the trusted advisor status of the marketing researcher. The Process for Defining Decision Making Problems A characteristic of the decision making class of problems is that it is the symptom that first emerges. The symptom could be a decline in sales or market share, a decline in customer or distributor/dealer satisfaction, a change in the repeat buying rate, and so forth. A key point we make here is that the symptom is very different from the problem from which it arises and which has to be solved by marketing research. And it is a long and arduous path to decipher the underlying problem from a symptom. Pursuing the medical analogy again, coughing is a symptom and not a problem. The problem could be a viral infection or cancer of the lung and investigations must be carried out to differentiate between the two. Just as the treatment for these two conditions would be drastically different and a wrong treatment due to a misdiagnosis could potentially be life threatening, in the market research domain too, a wrong diagnosis of the problem could lead to futile research and, at times, research that might potentially lead the organization down the wrong path. It is imperative that the market researcher define the right problem. A problem that is rightly defined and identified for research has the following characteristics: (1) it is clearly articulated, (2) it is completely articulated, and (3) it is necessary to solve it for arriving at a decision. Clear articulation implies that everybody reading the problem definition statement takes away the same message. To accomplish this, the problem

definition statement should list and differentiate related problems that are not being targeted or addressed. Complete articulation results when the why (business/market objectives) and the how (brief methodology) are stated in addition to the what is being done. A problem is deemed as necessary for investigation and research if the clients of that project can verbalize the potential different findings and the resulting different decision outcomes. If each of the potential different answers would not lead to a different decision, the problem is not worth being addressed. By following the process outlined next, these requirements for the problem to be the right one to be resolved are met. Figure 1.4 shows the process for identifying and defining the right problem for research from a symptom. On the left-hand side of the figure are rectangular boxes that show the recommended, ideal progression of a market research project under the trusted advisor approach, starting with a symptom and culminating in research design - we label this recommended progression as the Deductive Logic Approach (DLA). In a situation where the market researcher is an order taker, the path from symptom to design is much shorter. Here it is not unlikely for the client to tell the market researcher to employ a particular methodology to collect some type of data or for the researcher to offer the client some methodologies based on the symptom. The resulting research tends to be underused and undervalued. Following the recommended DLA steps, on the other hand, ensures that the research results are eagerly awaited and accepted. The forthcoming sections will establish that there is improved usage of results from DLA as compared to the results from the order-taker approach. To illustrate DLA versus the order-taker approach, we use the example of a large manufacturer of servers, ABC, that is facing the symptom of declining sales of its low-end Server X. The Order-Taker Approach

Following the classic model of market research, ABC requested proposals from two market research suppliers that immediately obliged. Research Supplier I proposed to study the following aspects using the methodologies described below. It quoted a price in the $100,000 range, justified the costs well, adequately proved that it had all the assets and resources to execute the projects more than satisfactorily, and promised a completion time of four to six weeks. The strengths of Supplier I are also shown below. An objective analysis of the research proposal revealed a well-thought-out, tightly knit proposal.

Symptom Exploratory Research Problem Area Business and Industry Knowledge Business Options

Business Decisions Interaction Market Research/Client Importance of Problem

Information Needs

Problem Definition

Trust

Market Research Design and Execution

Figure 1.4: Getting to the right Problem Definition: The DLA Approach

Supplier I: Proposed Market Research Projects

ABCs image in the server market. Trends and happenings in the server market. Competitive stance of ABC X with competitors, in the eyes of end users (and other similar servers). Advertising and promotion effectiveness of ABC X. Distribution issues with ABC X. Price perceptions of ABC products in general and Server X in particular.

Proposed Methodologies

Brand Image and General Price Perceptions Personal interviews of 100 Fortune 500 CIOs (adjusted for industry). Telephone survey of 1000 small and medium-sized business decision makers (random sample adjusted for geography, industry, sophistication, and so on). Trends 50 personal interviews with experts (e.g., R&D, product managers from industry leaders such as Dell, Hewlett-Packard, Sun, Oracle, and Microsoft). Competitive Evaluation, Ad/Promotion Effectiveness, Price Comparison, Availability for X 1500 mail/e-mail surveys of decision makers across all potential users of Server Xtype products (random sample). Distribution Internal company records. Competitive intelligence.

Company Strengths

Strengths in high-tech market research. Warehouse of high-tech knowledge. 50 interviewers with strong technical knowledge. Large CATI facilities. Propriety Web-research software.

Supplier II: Research Supplier II contended that the answer to the firms declining sales was in the behavioral data. It had an extensive database of sales of personal computers, servers, and mainframes,

including information on buyers, sellers, dates of transaction, places of transaction, places of installation, intended usage, and so forth. Supplier II provided elaborate models of how the data would be analyzed to determine the root cause of the declining sales. The supplier had a good reputation in the marketplace, but ABC had a hard time convincing itself whether the cost of approximately $75,000 was justified; after all, it is difficult to put a price on quality intangibles. Nonetheless, it was clear that the models would unearth the causal independent variables for the decline in sales. An analysis of the two proposals reveals that both delved straight into research designs from a description of the symptom. Although it is likely that both approaches would reveal some useful data, there remains significant danger of one or more of the following disadvantages of this symptom-to-research-design approach to conducting market research: (1) In the first research design, volumes of data will be collected. It is quite likely that the client will be given a massive report. Synthesizing insights from such large reports is a formidable task which generally then is not undertaken. The client will be left inundated with data/information with not much to confidently act upon. (2) The first proposal is susceptible to both response and nonresponse biases. Perhaps the best way to estimate the extent of these biases is for the reader to put himself or herself into the shoes of the CIO, R&D managers, product managers, and decision makers to determine their enthusiasm to participate in such a survey. (3) If the research reveals counterintuitive results, it runs the risk of not being used. (4) Compared with the proposed DLA progression, both proposals are significantly more expensive and time consuming. (5) The behavioral data proposal is limited in the insights it provides because it does not reveal what it would take for the potential customers to accept ABCs Server X. The results from this proposal will hence have limited actionability. (6) Finally, the data provided by both

proposals may not be on target to help in decision making. These disadvantages become clearer when the DLA approach is understood, which we describe next.

The DLA Approach

From Symptom to Problem Area. The DLA approach recommends considerable interaction, with the client and execution of exploratory research at the point at which the symptom is discovered, to understand the problem area. In the preceding example, interaction with the client and exploratory research with salespeople and channel members revealed that a competitors (DEF) Server X was eating into the sales of ABCs Server X. Perhaps more frustrating was that DEFs Server X was less compelling than ABCs, as surmised by discussions with ABCs product managers. Further exploratory research with the channel members revealed that DEFs Server Z was superior to ABCs Server Z and that channel members were selling DEFs Server X along with DEFs Server Z as a package. The superiority of DEFs Server Z was significant enough for customers to overlook the faults of DEFs Server X. From Problem Area to Business Options. With the problem area identified, serious interaction and brainstorming with the client must begin in order to identify what business options are available. For the sake of this example, some of the business options are as follows: (1) ABCs X could be sold along with DEFs Z, (2) ABCs Z could be significantly improved, and (3) ABC could sell X directly to customers. From Business Options to Business Decision. Business options should be thoroughly analyzed within the company in terms of internal plausibility. Exploratory research should be conducted mostly internally (and, perhaps, sometimes externally) to prioritize the business options. The idea is to prioritize the options with the hope that one is more attractive in terms of business criteria (e.g., feasibility, sales, profitability) than the others; it is not to appraise each option completely. Two or more options might appear equally attractive from a business

perspective after preliminary research. In that case, correspondingly two or more market research programs must be conducted. Table 1 shows the high-level information needed and the potential sources for evaluating each of the three options.

Information Needs 1. Any hardware and software compatibility issues? 2. Dealers reactions/ concerns? 3. Financial impact on ABCs Zs sales? 1. Efforts and resources involved in improving ABCs Z? 2. Financial impact of improved ABCs Z? 3. Impact on ABCs product line?

Sources 1. Internal R&D 2. Dealers 3. Salespeople

Option 1: Sell ABCs X with DEFs Z Option 2: Upgrade ABCs Z

1. R&D and product managers

Option 3: 1. What is the valueSell X directly to added by dealers? customer 2. Degree of customization of X, if needed? 3. Time required for customization? 4. Customer buying process?

1. Dealers 2. Internal sources 3. Salespeople

Table 1.1 Options, Information Needs, and Sources For the sake of simplicity, let us assume that from a strategic point of view the first two options are not as attractive as the third option. The first option is not attractive because ABC does not want to install its X with its archrivals Z because of competitive secrets that may be

compromised when striving for compatibility. Similarly, the second option is not attractive because, strategically, Z should not be upgraded because of its potential to eat into the market share of the highly profitable next level of server. With the help of exploratory research and interaction with the client, the trusted advisor should now determine the nuances of selling directly to customers. If, for example, an issue of customization is determined to be relevant, the trusted advisor should determine ABCs capabilities to do such customization and the approximate time required for customization. If, from a business point of view, this option seems to be feasible, the business option should be reframed as a business decision. In our example, it could be stated as follows: Should ABC sell low-end Servers X directly to customers? This now becomes a decision that the client must make, and because it is still a question, the client needs more information to make the decision. From the Business Decision to Information Needs. The trusted advisor should now question the client about the information it needs in order to make a go/no-go decision. Such questions force the client to think ahead in terms of what it will do with the information, and they also make it salient to the client that such information is currently not available. The interaction at this stage is critical to the usage of researcha problem that has been plaguing market research for years. This interaction will also convince the user that insights have been delivered when the results are delivered. As discussed in the Introduction, one of the key to results being considered insights is in the user wanting it. This stage does exactly that; it makes the client want the results. In summary, given that the client is faced with a business decision, that the lack of information has been made salient, and that alternative courses of action have been laid out based on the information, most likely the client will not only use the market research but also eagerly await the results.

From Information Needs to Market Research Problem Definitions. Again, for the sake of simplicity, let us assume that customer issues are the key determinants of the go/no-go decision that was stated as Should ABC sell low-end Servers X directly to customers?. That is, the degree of customization within a time frame that is acceptable to the customer is not a problem, and the distributor issues are marginal in selling directly to customers. Thus, in our example, the problem definition could be stated as follows: To determine customer concerns and needs when buying Server X directly from ABC to be delivered in T days. From Problem Definition to Research Design. The last must-do in-house stage in the DLA process is the research design. How does the market researcher go about determining whether customers would be willing to accept this proposition? Several methodologies can be considered immediately to answer this question. For example, a tightly drawn-up survey, focus groups, and individual interviews of customers could be conducted. The focus groups could be used to brainstorm potential objections that ABC could then address. The interviews could analyze the different customer segments decision-making processes, and surveys could size the attractive segments and their purchase intentions. In addition, the DLA process recommends reliance on triangulation, i.e., interviewing different audiences by perhaps using different methods. In this example, dealers, salespeople and so forth could be interviewed individually or collectively to learn their assessment of how customers may react to this proposition. If all roads lead to the same conclusion, the confidence in the conclusion is much greater. From Research Design to Execution. Having selected the most appropriate methodologies, the trusted advisor now seeks out the best vendors to deliver the results. Quality, time, and cost are the criteria used to choose the vendor.

The Process for Defining Market Learning Problems

The preceding section described the DLA process of arriving at the right problem definition from a symptom for the class of decision making problems. Many marketing research problems, however, do not present as a symptom nor are they investigated for the purpose of immediate decision making. They emerge from fuzzy information needs and are researched to enhance the conceptual understanding of the marketplace. Such broad and fuzzy problems include examples such as the need for an organization to understand how its different customer segments evaluate the quality of its product-line and organizations interest in learning customer attitudes toward its products in different usage situations. If structured right, the problem definition of market learning problems ends up being stated as a conceptual framework of important variables and their relationships. This conceptual model itself proves quite insightful for many users of marketing research. For others the conceptual model may be tested for confirmation and/or for estimation of parameters of the model. The process we propose be used for getting to the conceptual models which define market learning problems is based on inductive reasoning. We term this process Grounded Insights Discovery (GID). GID is different from traditional and popular research methodologies which are mostly based on deductive reasoning. Research in a deductive model proceeds as theory hypotheses data testing. The theory is assumed to be correct, and the hypotheses that emerge from the theory are tested with data. On the other hand, in an inductive or grounded model, the research process is conducted as data pattern hypotheses/theory/insights. In other words, there is no assumption about the theory the theory is developed rather as a part of the process. Many scholars believe that deductive reasoning does not augment their knowledge of the world, but inductive reasoning does. Since the GID methodology is based on the less

popular inductive approach, acceptance by users of market research of any conceptual model developed by this methodology requires a trusted advisor status of the marketing researcher. The process of Grounded Insights Discovery is based on seven pillars: (1) hub-andspoke structure of the research team, (2) multimethodmultisource researching for the complete problem space, (3) purposive sampling, (4) domain experts/peers as data collectors/researchers, (5) natural interviewing style (6) simultaneous collection and analysis of data, and (7) the three Cs of the researchers stance - clean slate, clear mindset, and complete problem resolution. We discuss each of the seven pillars next. Hub-and-Spoke Structure of the Research Team. A critical component of GID is the huband-spoke operational structure of the research team in which one head researcher collates all the information from other researchers. The head researcher may be the trusted advisor on the project or may report to the trusted advisor. This person has greater expertise than the individual researchers in the research category that the team is engaged in (e.g., quality assessment, segmentation, ad research, satisfaction), the research techniques the team is using (e.g., coding of qualitative data), and the domain the team is researching (e.g., automobiles). Because of a higher level of expertise, the head researcher can better connect the dots, and ascertain the bigger picture or themes from the different pieces of information coming from individual researchers (Frase-Blunt 2001). The hub-and-spoke structure is different from the default brainstorming model in which all the researchers come together to jointly draw conclusions and make hypotheses. Figure 1.5 illustrates these two models. The inefficiencies of brainstorming sessions have been increasingly mentioned in the literature (Sutton and Hargadon, 1996). For example, participants in brainstorming sessions can feel intimidated in the presence of other participants and may thus leave their ideas unvoiced. Further, discussions can proceed far beyond their usefulness as there

is no moderator. In contrast, GIDs hub-and-spoke model does not suffer from these issues because head researcher who also serves as a moderator.

Data Collectors/ Researchers

Head Researcher

Hub-and-Spoke Model

Brainstorming Model

Figure 1.5: Hub-and-Spoke Model Versus Brainstorming Model

MultimethodMultisource Researching for the Complete Problem Space. For a broadbased problem, insights result from addressing the complete problem space. However, traditional research paradigms are not designed to address the complete problem space. Rather, they tend to artificially constrain the problem that is being addressed so that problem suits the methodology or they allow the methodology to determine what part of the fuzzy problem is researched. Findings, hence, are limited to what the selected methodology can uncover. In traditional research approaches, the emphasis is on making the methodology valid for the artificially limited problem space, but in GID, the emphasis is on covering the entire problem space by using as many sources of information and methods as required. Even if multiple

methods are used in traditional research programs, gaps in knowledge still remain. The difference between an approach of a traditional research program to solving a problem and the GID approach is best illustrated with the help of the diagram in Figure 1.6. In Figure 1.6, Panels A and B, the oval spaces represent the information needs of a broad-based business problem. Panel A shows the coverage by a typical research program in which exploratory research covers some space and quantitative research confirms some of those findings. Some of the space remains uncovered. Panel B shows that, in contrast, GID expressly attempts to cover all the space, employing any and all methods and sources of information that may be required.

Confirmatory Research

Gaps Ovals Represent Information Needs/ Problem Space

Expert Interviews

Observations

In-company Sources Focus Groups Focus Groups In-company Sources Secondary Research Purposive Sampling to Cover Gaps Secondary Research

Panel A: Traditional Research

Panel B: GID

Figure 1.6: Traditional Research Program Versus GID Optimization and efficiency result from GID because in the GID methodology, the sample sizes or the number of information sources used (e.g., 500 interviews or 5 focus groups)

is not predetermined. When new information is not forthcoming from a particular method or source, data gathering from that method or source is abandoned. The trusted advisor determines the methods and sources required to cover all the space through the technique of Purposive Sampling. Purposive Sampling. In purposive sampling, a respondent or a source of information is selected for interview or study on the basis of the information the respondent or the source is capable of providing. While Chapter 4 discusses several types of purposive samples, there are a few which specially pertain to GID (see also Morse, 1991). The most relevant type of sampling for GID is Theoretical Sampling where selection of a source of information or type of respondent is based upon the stage of the research and upon the types and extent of insights that have been generated until that point (Coyne, 1997, pp. 624, 628; Eisenhardt, 1989; Sandelowski, HolditchDavis, Harris, 1992; Schatzman and Strauss, 1973, p. 39). In other words, the subsequent method or source to be investigated at any stage in a particular research project is determined by what needs to be assessed from that point onwards. In addition to filling gaps in knowledge, purposive sampling should be conducted to also confirm or disconfirm findings generated at each stage. Sources of information can be identified via the snowballing technique in which existing sources help identify other sources. Not only does GID focus on increasing the diversity of the sample, but it also ensures that adequate samples of like sources are used to replicate findings. If any interesting findings result from one source, the researcher actively obtains other similar sources to determine if the findings can be confirmed. Needless to say, at some point replication stops if the topics being covered saturate or reach redundancy.

Domain Experts/Peers as Data Collectors/Researchers. GID advocates that data collectors and researchers have expertise in the domain of relevance to the research they are undertaking and in case of interviews, if necessary, they even be peers of the respondents. This requirement is especially applicable to research involving technical or professional fields. Peers, by definition, are domain experts, whereas domain experts are not necessarily peers. An example of data collection by domain experts is secondary research in a medical domain or interviews of physicians carried out by a nurse or a biology student, individuals who understand the field and the topic under investigation. By the same token, data collection by peers would mean that physician interviews be conducted by a physician. Where interviews are concerned, Grounded Insights Discovery argues that if a respondent is unable to identify with the interviewer, the information generated from the interviews will be superficial and of poor quality. Professional groups often operate within certain technical and cultural boundaries which serve as barriers to outsiders. Peer interviewers, and to a lesser extent domain experts, in addition to having no barriers to overcome, are able to speak the language of the respondents and relate to them. They are more knowledgeable about the type of questions to ask for eliciting a fuller discussion and more intricate details. The sense of shared experience and shared identity, that peers and domain experts are likely to generate in respondents, makes it more likely that their research efforts will lead to deeper and more meaningful insights. The benefits of peer interviewing has been proven in many research fields, including feminism (Bhopal, 1995), social science (Alderson, 2001; Egharevba, 2001), anthropology (Messerschmidt, 1981), and human resources (Frase-Blunt, 2001). GID also recommends that peers and domain experts who serve as data collectors/ interviewers/researchers, besides having expertise in their domain, gain the know-how of

research techniques and knowledge of the research category in which they are conducting research. The requirements described above for researchers engaged in the GID process are quite different from those necessary in researchers involved in traditional marketing research paradigms. The latter essentially focus on data gathering skills whereas the former focus more on the background and domain expertise of the researchers. Grounded Insights Discovery propounds that data gathering skills and knowledge of the research category are obtained more easily than expertise of a particular domain. Natural Interviewing Style. For discovering insights about the marketplace that are valuable to an organization, it is often necessary to interview selected individuals for a glimpse into their deeper thoughts and feelings. One-on-one interviews serve this purpose better than group interviews. When one-on-one interviews are conducted in a natural conversation style, they further increase the prospect of meaningful result. The natural conversational style of interviewing sometimes uses free association in an unstructured format (Zaltman 2003). It is superior to an interview conducted with a structured discussion guide of questions that has the danger of precipitating a boring and awkward conversation in which the interviewees might feel they are undergoing a test, leading them to give guarded and superficial answers. A one-on-one interview fosters trust between the interviewer and the respondent (Rubin and Rubin 1995).4 A natural conversation style of such as an interview furthers the probability and level of that trust. Support for the natural interviewing conversation style in qualitative research is found in (1) anthropology (Messerschmidt, 1981); (2) nursing research (Archibald, 2003); (3) social research on anxiety, stress, and coping (Ibanez et al., 2004); and (4) psychology (Mills and Daniluk, 2002).

Simultaneous Collection and Analysis of Data. GID involves the simultaneous collection and analysis of data, a laborious activity with multiple tasks. These include: cognizance of research category, note-taking, constant comparing of new to existing data, memo-writing, sorting and grouping notes and memos, coding and collation of data collected by head researcher, and arrival at problem definition. It is necessary that researchers have training in these tasks. In the hub-and-spoke model this training requirement is lower for individual researchers as the head researcher with expertise in all these tasks is available for real-time consulting. The individual tasks of collection and analysis of data follows. Cognizance of the research category of their project is important for researchers. Research category refers to a problem area and the primary purpose of the research project, e.g., segmenting the market, assessing product quality, understanding and measuring customer satisfaction and conducting a trade-off analysis. Cognizance of the research category is important for marketing researchers so that they can consciously stay within the bounds of the research category. In employing multimethod-multisource research for defining a market learning problem, the potential amount of raw data gathered can be enormous. Staying within the bounds of the research category ensures that no unnecessary information is collected and also limits the amount of data a researcher must sift through. The benefits of adherence to the research category are especially applicable in case of one-on-one interviews conducted in natural conversation style. Researchers record their comments and findings during interviews or other data gathering processes by note taking. The purpose of note taking is to document, on an on-going basis, the key emerging issues relevant to all the research aspects, such as the product-market, the research category, and the source/respondent. It is preferable that note taking occur as the data are

gathered. This allows researchers to engage in constant comparison, the first activity in which data collectors begin to behave like researchers. Concomitant to data gathering, they compare the newly emerging data with data that have been collected previously. The later might have been gathered from different sources or from the same source at earlier times. Researchers also write memos to themselves to record their findings. Memos are at a higher level of abstraction than notes referred to above. However, memos could also be simply researchers to-do list or tofind-out list. After each round of data collection, researchers meet with the head researcher and communicate their notes and memos which are sorted and grouped. The head researcher then conducts coding and collation of data which involves extraction of new constructs, uncovering important variables, and deducing new relationships between different variables from the collected notes and memos. Examples of these new constructs and variables are represented by new segments, new attributes/benefits that are being used to assess product/quality/satisfaction, new processes for decision making, new segment descriptors, and so forth. If coding reveals evidence to disconfirm prior beliefs, the researching team hypothesizes reasons for this and proceeds to gather new data to resolve the anomaly. As described in the section on purposive sampling, the steps of the research cycle are repeated till all aspects of the problem space are addressed and all gaps filled for full and complete problem definition. The Three Cs of the Researchers Stance - Clean Slate, Clear Mindset, and Complete Problem Resolution. The outcome of the research project depends upon the motivation and intent with which the researcher or the researching team carries out the research. In this respect, the success of the GID process and the quality of resulting insights are tied to the three Cs stance

of the researcher clean slate beginning and clear mindset for a complete problem resolution. The first C of the researchers stance refers to clean slate and implies that researchers begin considering the research issues and tackling the research task with no preconceived notions of the possible solutions or insights that might arise from the project. The second C stands for clear mindset and signifies that the researcher stay unbiased through the project, as qualitative research especially lends itself to bias. Researchers must ensure that both confirmatory bias and oversight bias are kept to a minimum. Confirmatory bias relates to the tendency of the researcher to latch on to an observation that confirms a prior belief or expectation of the researcher and, at the same time to ignore or fail to perceive an unexpected observation. The potential of oversight bias stems from the use of experts who, because of their familiarity with the domain and their thought processes being set in a certain way, might miss certain observations. The last C refers to complete problem resolution within the boundaries and space that are defined by the client/researcher for the problem and the research category. It suggests continuous sampling be conducted to cover all gaps and rule out any disconfirming evidence. Cutting corners in sampling and data gathering can result in erroneous definition of the market learning problem and can generate a faulty conceptual model with false insights.

Conclusions
In this chapter we provided some ideas on how marketing research can get back on track for providing insights to users of marketing research. We focused on the need and importance of right problem definition in this context, and saw that the right problem definition might itself prove insightful. We identified two types of problems in marketing research, decision making problems and market learning problems, and suggested individual approaches for defining these

two separate types of problems. We also discussed that for implementation of these approaches, the marketing research function be reformed to incorporate the role of trusted advisors in an integral way.

REFERENCES Alderson, P. (2001). Research by children. International Journal of Social Research in Methodology, 4(April), 139-153. Archibald, G. (2003). Patients experience of hip fracture. Journal of Advanced Nursing, 44(November), 385-392. Bhopal, K. (1995). Women and feminism as subjects of black study: The difficulties and dilemmas of carrying out research. Journal of Gender Studies, 4(2), 153-168. Bigley, G. A., & Pearce, J. L. (1998). Straining for shared meaning in organization science: Problems of trust and distrust. Academy of Management Review, 23(3), 405-421. Coyne, I. T. (1997). Sampling in qualitative research. purposeful and theoretical sampling: Merging or clear boundaries? Journal of Advanced Nursing, 26(September), 623-630. Doney, P. M., & Cannon, J. P. (1997). An examination of the nature of trust in buyer-seller relationships. Journal of Marketing, 61, 35-51. Egharevba, I. (2001). Researching an-other minority ethnic community: Reflections of a black female researcher on the intersections of race, gender, and other power positions on the research process. International Journal of Social Research in Methodology, 4(July), 225-241. Eisenhardt, K. (1989). Building theories from case study research. Academy of Management Review, 14(4), 532-550. Frase-Blunt, M. (2001). Peering into an interview. HR Magazine, (December). Ibanez, G., Buck, C., Khatchikian, N., & Norris, F. (2004). Qualitative analysis of coping strategies among mexican disaster survivors. Anxiety, Stress, and Coping, 17(March), 69-85.

Jeffries, F. L., & Reed, R. (2000). Trust and adaptation in relational contracting. Academy of Management Review, 25(4), 873-882. Madhavan, R. and Grover, R. (1998). From embedded knowledge to embodied knowledge: new product development as knowledge management. Journal of Marketing, 62 (October), 1-12. Mayer, R. C., Davis, J. H., & Schoorman, D. F. (1995). An integrative model of organizational trust. Academy of Management Review, 20(3), 709-734. Messerschmidt, D. (1981). Anthropologists at home in America: Methods and issues in the study of ones own society. New York: Cambridge University Press. Mills, L., & Daniluk, J. (2002). Her body speaks: The experience of dance therapy for women survivors of child sexual abuse. Journal of Counseling & Development, 80(Winter), 77-85. Moorman, C., Deshpand, R., & Zaltman, G. (1993). Factors affecting trust in market research relationships. Journal of Marketing, 57, 81-101. Morse, J. (1991). Strategies for sampling. In J. Morse (Ed), Qualitative nursing research: A contemporary dialogue (pp. 127-145). Newbury Park, CA: Sage. Rubin, H. & Rubin, I. (1995). Qualitative interviewing: The art of hearing data. Thousand Oaks, CA: Sage. Sandelowski, M., Holditch-Davis, D., & Harris, B. (1992). Using qualitative and quantitative methods: The transition to parenthood of infertile couples. In J. Gilgun, K. Daly, & G. Handel (Eds.), Qualitative methods in family research (pp. 301-323). Newbury Park, CA: Sage. Schatzman, L., & Strauss, A. (1973). Field research: Strategies for a natural sociology. Englewood Cliffs, NJ: Prentice Hall.

Sutton, R., & Hargadon, A. (1996). Brainstorming groups in context: Effectiveness in a product design firm. Administrative Science Quarterly, 41, 685-718. Zaltman, G. (2003). How customers think: Essentials insights into the mind of the market. Boston: Harvard Business School Press. Zand, D. E. (1972). Trust and managerial problem solving. Administrative Science Quarterly, 17, 229-239.

There is sufficient evidence that trust between organizations is different from individual-level trust (Doney and Cannon 1997). Organizational-level trust is addressed in the next chapter. 2 The other bases of trust that have been discussed in the literature are value, calculus, knowledge, capability, and benevolence. However, these conceptualizations seem to define the term bases more as reasons for trusting than as the source within the trustor.
1 3

Expertise, professionalism, and personal trust could be conceptualized as three concentric circles, with the innermost circle representing expertise, the next professionalism, and the outermost personal. Such a conceptualization is useful in terms of the skills that need to be developed as a market researcher progresses through its life cycle. As a fresh graduate, a market researcher is armed with technical expertise (a necessary condition for him or her as a market researcher). Soon he or she must master the requirements for professionalism in this discipline, and finally he or she must win over personal trust of one or more clients.

Focus groups are over-used by market researchers. They are effective when serendipity, creativity, and idea development are important. These factors are not of importance when conceptual understanding of the marketplace is desired. In fact, these factors may lead to erroneous conclusions because individuals in a group setting can succumb to group pressure and create fictitious stories based on others comments.
4